JMM 2015 Student Poster Session Abstract Book

Document technical information

Format pdf
Size 624.6 kB
First found Nov 13, 2015

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

John Van Alstine
John Van Alstine

wikipedia, lookup

Stephen Smale
Stephen Smale

wikipedia, lookup

Georg Cantor
Georg Cantor

wikipedia, lookup

Linus Pauling
Linus Pauling

wikipedia, lookup

Daniel Daly
Daniel Daly

wikipedia, lookup

Daniel Adams-Ray
Daniel Adams-Ray

wikipedia, lookup

Robert Bell (geologist)
Robert Bell (geologist)

wikipedia, lookup

Paul Ekman
Paul Ekman

wikipedia, lookup

Benjamin Allen (politician)
Benjamin Allen (politician)

wikipedia, lookup

Organizations

Places

Transcript

Abstracts for the MAA
Undergraduate Poster Session
San Antonio, TX
January 12, 2015
Organized by
Joyati Debnath
Winona State University
CELEBRATING A CENTURY OF ADVANCING MATHEMATICS
Organized by the MAA
Committee on Undergraduate Student Activities and Chapters
and
CUPM Subcommittee on Research by Undergraduates
Dear Students, Advisors, Judges and Colleagues,
If you look around today you will see about 273 posters and 492 presenters, record numbers, once again. It
is so rewarding to see this session, which offers such a great opportunity for interaction between students
and professional mathematicians, continue to grow.
The judges you see here today are professional mathematicians from institutions around the world. They
are advisors, colleagues, new PhD.s, and administrators. We have acknowledged many of them in this booklet; however, many judges here volunteered on site. Their support is vital to the success of the session and
we thank them.
We are supported financially by the National Science Foundation.
Our online submission system and technical support is key to managing the ever-growing number of
poster entries we receive. Thanks to MAA staff, especially Julia Dills and Maia Henley for their work
setting up and managing the system this year. Preparation of the abstract book is a time-consuming task.
Thanks to Beverly Ruedi for doing the final production work on the abstract book.
There are many details of the poster session that begin with putting out the advertisement in FOCUS in
February, ensuring students have travel money, and organizing tables in the room we are in today that are
attributed to Gerard Venema (MAA Associate Secretary), Linda Braddy (MAA), and Donna Salter (AMS).
Zsuzsanna Szaniszlo (Valparaiso University) rallied volunteers to judge the session and coordinated the
judge assignments. Angel R. Pineda (California State University, Fullerton), James P. Solazzo (Coastal Carolina University), and Dora Ahmadi (Moorehead State University) organized an orientation for the judges
and authored the judging form. Thanks to all the students, judges, volunteers, and sponsors. I hope you have
a wonderful experience at this years poster session!
Joyati Debnath
Winona State University
1
Judges and Judge Affiliations
1. Saroj Aryal, Montana State University Billings
2. Liljana Babinkostova, Boise State University
3. Brad Bailey, University of North Georgia
4. Jeremiah Bartz, Francis Marion University
5. Joanna Bieri, University of Redlands
6. Russell Blyth, Saint Louis University
7. Nathanial Burch, Gonzaga University
8. Kristin Camenga, Houghton College
9. Eric Canning, Morningside College
10. James Carlson, Creighton University
11. Mehmet Celik, University of North Texas at Dallas
12. Leah Childers, Benedictine College
13. David Clark, Grand Valley State University
14. Daniel Daly, Southeast Missouri State University
15. Doreen De Leon, California State University, Fresno
16. Louis Deaett, Quinnipiac University
17. John Diamantopoulos, Northeastern State University
18. Elizabeth Donovan, Bristol Community College
19. Eric Eager, University of Wisconsin- La Crosse
20. Joe Eichholz, Rose-Hulman Institute of Technology
21. Eleanor Farrington, Massachusetts Maritime Academy
22. Brian Fisher, Lubbock Christian University
23. Tim Flowers, Indiana University of Pennsylvania
24. Sharon Frechette, College of the Holy Cross
25. Andrew Gainer-Dewar, Hobart and William Smith Colleges
26. Whitney George, University of Wisconsin-La Crosse
27. Nicholas Gorgievski, Nichols College
28. Matthew Haines, Augsburg College
29. Spencer Hamblen, McDaniel College
30. Jessie Hamm, Winthrop University
31. Jason Hardin, Worcester State University
32. James Hartman, The College of Wooster
33. Thomas Hoft, University of St. Thomas
34. Heidi Hulsizer, Hampden-Sydney College
35. Aminul Huq, University of Minnesota Rochester
36. Kevin Iga, Pepperdine University
37. Lynne Ipina, University of Wyoming
38. Chester Ismay, Ripon College
2
Judges and Judge Affiliations
39. Michael Jackson, Grove City College
40. Cory Johnson, California State University, San Bernardino
41. Jesse Johnson, Westfield State University
42. Mitch Keller, Washington And Lee University
43. Joel Kilty, Centre College
44. Joseph Kirtland, Marist College
45. Alan Koch, Agnes Scott College
46. Liz Lane-Harvard, University of Central Oklahoma
47. Donald Larson, Penn State University, Altoona
48. Najeem Lateef, University of South Africa
49. Jessie Lenarz, St. Catherine University
50. Ying Li, Saint Francis University
51. Benjamin Linowitz, University of Michigan
52. Shannon Lockard, Bridgewater State University
53. Brian Loft, Sam Houston State University
54. Timothy Lucas, Pepperdine University
55. Thomas Madsen, Youngstown State University
56. Marco V. Martinez, North Central College
57. Kristen Mazur, Elon University
58. Cayla McBee, Providence College
59. Laura McSweeney, Fairfield University
60. Herbert Medina, Loyola Marymount University
61. Kristi Meyer, Wisconsin Lutheran College
62. David Milan, The University of Texas at Tyler
63. Kimberly Muller, Lake Superior State University
64. Jeanine Myers, University of Arkansas-Fort Smith
65. Julius Myers, University of the Ozarks
66. Christopher ONeill, Texas A&M University
67. Stephen Pankavich, Colorado School of Mines
68. James Peirce, University of Wisconsin - La Crosse
69. Nathan Pennington, Creighton University
70. Angel Pineda, California State University, Fullerton
71. Vadim Ponomarenko, San Diego State University
72. Megan Powell, University of St. Francis
73. Kenneth Price, University of Wisconsin Oshkosh
74. Nathan Reff, SUNY Brockport
75. Sandra RIchardson, Virginia State University
76. Margaret Robinson, Mount Holyoke College
77. Eric Ruggieri, College of the Holy Cross
78. Amanda Schaeffer Fry, Metro State University of Denver
3
4
Judges and Judge Affiliations
79. Marion Scheepers, Boise State University
80. Corey Shanbrom, California State University, Sacramento
81. Derek Smith, Lafayette College
82. Mutiara Sondjaja, New York University
83. Kristen Stagg, The University of Texas at Tyler
84. Catherine Stenson, Juniata College
85. Jessica Stewart, Goucher College
86. Cynthia Taylor, Millersville University of Pennsylvania
87. Paul Tokorcheck, Iowa State University
88. Anthony Tongen, James Madison University
89. David Torain, Hampton University
90. Violeta Vasilevska, Utah Valley University
91. Oscar Vega, California State University, Fresno
92. John Villalpando, California Lutheran University
93. Xiaodi Wang, Western Connecticut State University
94. Xiaodi Wang, Western Connecticute State University
95. Nathan Warnberg, University of Wisconsin-La Crosse
96. Marion Weedermann, Dominican University
97. Tessa Weinstein, Morningside College
98. Steven Wilkinson, Northern Kentucky University
99. Cassie Williams, James Madison University
100. G. Brock Williams, TTU Dept of Mathematics and Statistics
101. Janine Wittwer, Westminster College, UT
102. Elizabeth Wolf, Saint Mary’s College
103. Mohammed Yahdi, Ursinus College
104. Haley Yaple, Carthage College
105. Ellen Ziliak, Benedictine University
Titles, Authors, Advisors and Abstracts
1. Tight Frame Structure and Scalability
Rachel Domagalski Central Michigan University
Hong Suh Pomona College
Xingyu Zhang Pennsylvania State University
Advisor(s): Yeonhyang Kim, Central Michigan University
In Rn , a frame is defined to be a spanning set. A collection F D ffi gkiD1 Rn is a -tight frame if there exists
Pk
2
0 such that for every f 2 Rn , kf k2 D
i D1 jhf; fi ij . We examine the structure of frames through factor
posets and scalability. A factor poset for a frame F D ffi gkiD1 is the set P D fJ f1; : : : ; kg W ffj gj 2J is a
tight frameg, partially ordered by set inclusion, ; 2 P . This definition leads to the question: given a poset P , when
is P a factor poset? We call this problem the inverse factor poset problem (IFPP). The IFPP was solved in R2 in
2013. In our goal of solving the IFPP in Rn , we discovered combinatorial properties of tight frames and explored
constructions of frames from posets. Next, we examine the scalability of frames. For a frame F D ffi gkiD1 , a scaling
p
is a vector w D .w.1/; : : : ; w.k// 2 Rk0 such that f w.i /fi gkiD1 is a 1-tight frame in Rn . We establish results on the
structure of the scalability polytope and its connection to the factor poset. This research completed at Central Michigan
University’s 2014 REU.
2. Classifying Seven-Dimensional Solvable Lie Algebras with Six-Dimensional Abelian Niradical
Kyle Ferguson Grand Valley State University
Advisor(s): Firas Hindeleh, Grand Valley State University
This poster is the first in a series that examine seven-dimensional solvable Lie Algebras with a six-dimensional niradical. Low dimensional solvable Lie Algebra classification started back in 1963 by Mubarakzyanov. They were
completely classified up to dimension six. A general theorem asserts that if g is a solvable Lie Algebra of dimension
n, then the dimension of the nilradical is at least n2 . For the seven-dimensional algebras, the nilradical’s dimension
could be 4; 5; 6 or 7. The four and seven dimensional nilradical cases were classified. We examine the six-dimensional
niradical case. In this project we focus on the class where the nilradical is six-dimensional abelian, i.e isomorphic to
R6 .
3. Isomorphy classes of trivolutions of SL 2 .K/
Fernando Betancourt Velez University of Puerto Rico
Ontario Stotts Wake Forest University
Joyce Yang Harvey Mudd College
Advisor(s): Aloysius Helminck, North Carolina State University
Involutions of real matrix groups have been used to construct real symmetric spaces for more than a 100 years. Using
automorphisms of order n of matrix groups defined over an arbitrary field k one can construct spaces similar in nature.
These are called generalized symmetric spaces. This area has a strong connection to physics. As a first step in the study
of these generalized symmetric spaces we consider trivolutions of SL2 .K/. Trivolutions are group automorphisms of
order 3. SL2 .K/ is the group of 2-by-2 matrices with determinant 1. In this paper we will give a characterization for
the isomorphy classes of trivolutions of SL2 .K/ with K any field of characteristic not 2 or 3. This work is analogous
to work by Helminck and Wu on automorphisms of order 2.
4. Common Algebraic Errors: Interesting Math Problems
Luis Ortiz Universidad Metropolitana
Advisor(s): Luis F. De La Torre, Universidad Metropolitana
The algebra is a branch of mathematics where students have difficulties. In this research, a list of common algebraic
errors were selected and studied. These errors become interesting mathematical problems when are treated as multivariate equations. It means, each error is related to a mathematical problem associated, where a solution set is required
to be found. These solutions were initially found in the set of integers (Z) and then in the field of real numbers (R).
Most of these associated problems were solved using only algebra; in some cases it was necessary to use calculus,
5
6
Titles, Authors, Advisors and Abstracts
number theory and numerical analysis. In this research 20 problems associated with the most common algebraic errors
were solved. Some of these problems have unique solutions while some other of these problems have infinite solutions.
Algebraic errors are well discussed in the literature. However, to visualize these errors as mathematical problems are
an important tool to help our students to establish the differences and do not make these kind of mistakes.
5. On the Cohen-Macaulayness of Sn -Invariant Subspace Arrangements
Aaron Brookner Massachusetts Institute of Technology
Advisor(s): Pavel Etingof, Massachusetts Institute of Technology
In this paper, we define a family of algebraic varieties, X C n , where is an integer partition of n and X is a
union of subspaces of C n . We are specifically interested in which X satisfy the Cohen-Macaulay (CM) property,
which roughly states that the singularity of X at the origin is “not too bad.” We give examples, prove several results,
and make several conjectures about for which these X satisfy this property, both facilitated by computer calculations
done in the Macaulay2 program. The motivation for this work comes from representation theory, which provides the
only other known proofs that these X satisfy the CM property. Throughout this presentation, we highlight that X is
not CM “by chance.” That is, X seems to be CM only when there is a reason from representation theory and integrable
systems for it to be.
6. Reduced n Irreducible n Factor Graphs
Michelle Rosado University of Puerto Rico in Mayaguez
Advisor(s): Reyes M. Ortiz, University of Puerto Rico in Mayaguez
A nonzero nonunit integer x has a n -factorization if x D ˙x1 x2 xk ; where for each i ¤ j; n j xi xj : We also
say that x is a n -product of the xi and each xi is a n -factor of x: We called x a n -irreducible, if k D 1: In 2011,
Ortiz and Lucena defined a n -irreducible n -factor graph of x; as a graph in which the vertices represents the n irreducible n -factor (up to associates) and two vertex are connected if there is an n -factorization into n -irreducible
where the n -irreducible n -factor associated with such vertex appear in such n -factorization. We defined, the reduced
n -irreducible n -factor graph, in a similar manner the only difference from the n -irreducible n -factor graph is that
associates of a reduced n -irreducible n -factor cannot be represented by the same vertex. We present preliminary
results of a characterization for such graphs when n D 0; 1; 2; 3: They are distinct with respect to the ones Lucena and
Ortiz characterized in 2010.
7. Methods of Relating Types of Canonical Algebraic Curvature Tensors
Elise McMahon Ave Maria University
Advisor(s): Corey Dunn, California State University San Bernardino
S
ƒ
We relate canonical algebraic curvature tensors that are built from a self-adjoint (RA
) or skew adjoint (RA
) linear map
A. By Nash’s imbedding theorem, an algebraic curvature tensor built from a self-adjoint operator is realizable as the
curvature tensor of an embedded hypersurface in Euclidean space. We develop an identity to relate the skew-adjoint
canonical algebraic curvature tensor to the self-adjoint canonical tensors, which will allow us to employ previous
ƒ
methods to solve new problems. We compute the structure group of RA
, and develop methods for determining the
linear independence of sets which contain both builds of algebraic curvature tensors. We consider cases where the
operators are arranged in chain complexes and we find this case to be highly restrictive. Moreover, if one of the
operators has a nontrivial kernel, we develop a method for reducing the bound on the least number of canonical
algebraic curvature tensors that it takes to write an algebraic curvature tensor.
8. Leamer Monoids and the Huneke-Wiegand Conjecture
Sara Stover Mercer University
Dalton Worsnup Arizona State University
Advisor(s): Roberto Pelayo, University of Hawaii- Hilo
The Huneke-Wiegand Conjecture has earned much attention in commutative algebra; once a theorem was developed
connecting the Huneke-Wiegand Conjecture to numerical monoids, the study of arithmetical sequences in numerical
monoids was sparked. In our research, we restricted ourselves to characterizing numerical monoids with consecutive
generators (intervals). In particular, we focus on determining when a numerical monoid €=hm; m C 1; : : : m C ki
Titles, Authors, Advisors and Abstracts
7
has an irreducible sequence of the form fx; x C s; x C 2sg for s 2 Nn€. We discovered and proved a theorem
for finding an irreducible sequence of this form when € has two generators, thus satisfying the Huneke-Wiegand
Conjecture for monomial ideals generated by two elements. This lays the groundwork for finding conditions when
fx0 ; x0 C s; x0 C 2sg € for any number of generators.
9. A dynamic algorithm for compution of !-primality in numerical monoids
Thomas Barron University of Kentucky
Advisor(s): Christopher O’Neill, and Roberto Pelayo, Texas A&M, and University of Hawaii - Hilo
In a numerical monoid S (a cofinite additive submonoid of N), the ! invariant assigns to each element n of S a
positive integer !.n/ which measures how far n is from being prime. Existing algorithms for computing !.n/ for
a given n become prohibitively slow as n grows large. By using an extension of ! from S to the whole of Z, we
develop a dynamic algorithm (one which re-uses prior computations to speed up new queries) for computing !.n/
which provides drastic improvements in computation times. We also give an improved description of the asymptotic
behavior of !, which in combination with our algorithm, allows for very efficient computation of the complete behavior
of ! for a given numerical monoid. Our algorithm has been included in the NumericalSgps package in GAP, a popular
computational algebra system.
10. Hall-Littlewood Polynomials and a Twisted Six-Vertex Model
Roger Van Peski Princeton University
Vineet Gupta Stanford University
Uma Roy Boston University - PROMYS
Advisor(s): Daniel Bump, Stanford University
The six-vertex or ice-type model of statistical mechanics, originating with Linus Pauling’s attempt to calculate the
residual entropy of ice, has seen much work since its inception from both physical and pure combinatorial viewpoints.
In particular, Tokuyama’s deformation of the Weyl character formula, which expresses the Schur polynomial (of Cartan
type A) in terms of combinatorial data from Gelfand-Tsetlin patterns, has been rephrased in terms of partition functions
of six-vertex systems. In recent work, the presenters generalized Tokuyama’s original result to the Hall-Littlewood
polynomials, suggesting that there should be a analogous formula expressing the Hall-Littlewood polynomial as a
six-vertex partition function divided by the deformed Weyl denominator. Here we provide such a generalization,
expressing the type A Hall-Littlewood polynomial in terms of a partition function of a column-twisted six-vertex
model. We also present progress on a generalization of our formula to the type C Hall-Littlewood polynomials.
11. Characters of GL2 .K / with K a Finite Field
Aura Salazar Cardona University of Texas at Brownsville
Advisor(s): Ju-Lee Kim, Massachusetts Institute of Technology
Characters of group representations are fundamental tools for performing calculations in representation theory. The
entries of the character table of a group G are the values of each irreducible character on the conjugacy classes of
G. From the character table of a group we can obtain important group theoretical information, most notably we can
deduce all the irreducible representations of this group. In his paper “The Characters of the Finite General Linear
Groups,” J.A. Green provides a method to compute the character table for GLn .Fq /, the group of n n non-singular
matrices with entries coming from the field Fq . However, Green’s methods rely heavily on finite group theory, thus
they can hardly be used to find the irreducible representations of GLn .F / with F a local field. Our goal is to provide a
method to construct the character table of GL2 .K/ without explicitly using the finiteness of K, allowing future work
to expand this method to the case of GLn .K/ where K is a local field.
12. Random Groups at Density 1/2
Shelby Kilmer Bucknell University
Advisor(s): Moon Duchin, Tufts University
Random groups in the density model (with density 0 < d < 1) have presentations with a random set of .2m 1/dk
relators of length k on m generators. The classic theorem in the density model states that for d > 1=2, random groups
are asymptotically almost surely trivial or isomorphic to Z=2Z, while for d < 1=2, random groups are asymptotically
8
Titles, Authors, Advisors and Abstracts
almost surely infinite hyperbolic. This summer our research group studied random groups at d D 1=2 and found both
infinite hyperbolic groups and trivial groups are generic, depending on how we tuned certain parameters. This is the
report of a research project completed at Tufts University, Summer 2014, as a part of an NSF research cluster.
13. A Combinatorial Model for Rational Base Representations of Natural Numbers
Hailey Olafson Pacific Lutheran University
James Van Alstine Pacific Lutheran University
Advisor(s): Tom Edgar, Pacific Lutheran University
We investigate representation of natural numbers in terms of rational bases and some of the associated properties.
We then define a collection of combinatorial data associated with a rational number and study the properties of these
sequences. Finally, we use this combinatorial data to construct an infinite, labeled, rooted tree that describes many of
the interesting features of the relevant rational base representations of natural numbers, and we explore some of the
combinatorics of these trees.
14. Representing Finite Fields using Fibonacci Cycles mod p
Jeremy Porche Angelo State University
Shelby Robertson Angelo State University
Caitlyn Conaway Angelo State University
Advisor(s): John Smith, Angelo State University
Finite fields can be represented in various ways. Generally, they are most easily understood when represented as
vectors. This representation poses serious computational issues when considering the multiplicative operation. Several
methods have been devised to deal with this problem. Holger Schellwat presented a simple slide rule and William
Wardlaw simply circumvented the problem of dealing with vectors, by modeling finite fields using square matrices.
This research project used the work done by Wardlaw to represent finite fields using Fibonacci cycles mod p. This
new approach also yields a vector representation where the multiplication is somewhat transparent, and may even be
extended to obtain a simplification of the vector and slide rule problem.
15. Special Values in Leamer Monoids
Jackson Rebrovich Angelo State University
Nico Rojina University of North Carolina at Chapel Hill
Advisor(s): Roberto Pelayo, University of Hawaii at Hilo
A numerical monoid € generated by an arithmetic sequence with step size identical to the distance between the generators is called an arithmetical Leamer monoid. Much is known about the factorization theory for arithmetical Leamer
monoids. For every numerical monoid of embedding dimension two it is generated by the arithmetic sequence fa,
a+dg. However, much is still unknown about Leamer monoids when the step size does not equal the distance between
the generators. The ultimate goal is to prove that the Huneke-Wiegand conjecture holds for embedding dimension 2
numerical monoids for any s 62 €. One approach is to attempt finding a closed form for x0 . In this paper,we found a
formula to calculate the last step size before x0 D xf consecutively. In addition, we found a formula to find x0 and
xf , for the last time x0 ¤ xf for that particular step size. As a result, we found a closed form for x0 , which involves
Apery sets and modular arithmetic.
16. Toward a Generalization of the Newton{Girard Formulae
Emily Scott William Jewell College
Advisor(s): Azadeh, Samuel, Rafizadeh, Chamberlin
A symmetric polynomial is a polynomial in n variables such that any permutation of the variables leaves the polynomial unchanged. An example of a symmetric polynomial in two variables is x1 C x2 . This poster will present
our progress towards proving a conjectured generalization of the Newton–Girard formulae, which relate two types of
symmetric polynomials: the elementary symmetric polynomials and the power sum symmetric polynomials. These
formulae have applications in Galois theory, invariant theory, group theory, combinatorics, and general relativity. The
conjectured generalization we are trying to prove relates two other types of symmetric polynomials: the monomial
symmetric polynomials (which are a generalization of the elementary symmetric polynomials) and the power sum
symmetric polynomials. Our progress thus far has been to prove the conjecture in n variables in several cases.
Titles, Authors, Advisors and Abstracts
9
17. On the Kronecker Product of a Hook and a Box
William Hallahan College of the Holy Cross
Advisor(s): Cristina Ballantine, College of the Holy Cross
The algebra of symmetric functions is the collection of functions such that taking any permutation of their arguments
does not change the function. This collection is endowed with the operations of addition and multiplication. One
basis of this algebra consists of the Schur functions, which correspond to partitions. The Kronecker Product of Schur
functions is an operation which produces a symmetric function. There is no simple combinatorial rule for writing
the Kronecker Product in the basis of the Schur functions. Jonah Blasiak gives a combinatorial interpretation for the
coefficient of a given Schur function in the decomposition of the Kronecker product of two other Schur functions when
one of the corresponding Young diagrams is a specific shape, a hook. We are attempting to find a simpler rule for the
special case when the other shape is a box. We prove a simpler rule for coefficients of some of the Schur functions
appearing in the decomposition. We also give some stability results for the coefficients in the Kronecker product of
Schur functions indexed by a hook and a box.
18. Permanents of Random .0; 1/ | Square Matrices over ZP
Alexander Wilson The University of Texas at Tyler
Advisor(s): Kristen Stagg, The University of Texas at Tyler
Calculating permanents of n n matrices is impossible to compute in polynomial time for large values of n. Our aim
is to examine the permanents of random .0; 1/ - square matrices over the field ZP . We utilized Burnsides’ Theorem
2
in an effort to condense the 2n cases. We begin with 2 2 and 3 3 matrices and applied this technique to larger
matrices. We also hope to comment on the open problem of finding the probability that an n n random matrix has
permanent 0 over the field Zp as n approaches infinity. It is hypothesized that this probability is is 1=p.
19. On the divisibility and valuations of the Franel numbers
Samantha VanSchalkwyk Mount Holyoke College
Adela Yang Bowdoin College
Abraham Schulte Northwestern University
Advisor(s): Herbert Medina, Victor Moll, Loyola Marymount University, Tulane University
The Franel numbers are the sum of cubes of binomial coefficients. The set of primes may be partitioned with respect
to the p-adic valuations of Franel numbers as follows: those whose valuation is 0 for all Franel numbers, those whose
valuation is equal to the number of occurrences of a particular digit in the base-p representation of the index of the
Franel numbers, and those which fall into neither category. One of the goals of this project is to classify as many
primes as possible into these three types. Furthermore, the 2-adic valuations of the Franel numbers have interesting
properties that are explored in more detail.
20. Exploration into Birkhoff's Algebraic Structure Problem
Jessica Young Westfield State University
Advisor(s): Maureen Bardwell, Westfield State University
In a book originally published in 1940, Garrett Birkhoff posed the problem of classifying all partially ordered sets
whose group of order automorphisms .f; ; / is a lattice ordered permutation group. As part of my project, I examined
original French manuscripts in abstract algebra and obtained a copy of the annals of the 1900 World Congress of
Mathematicians where background for Birkhoff’s book was originally given. Although many have attempted to solve
this problem, there is currently no general solution. My project outlines the problem and explores specific examples
of partially ordered sets with their corresponding automorphism groups. I use these examples as building blocks for a
general example of my own which I classify and which can be used to generate a multitude of other examples.
21. Distance Between Finite Groups: Formulae, Bounds, and the Triangle Inequality
Kim Pham University of California Irvine
Weiwei Li University of California Irvine
Advisor(s): Alessandra Pantano, University of California Irvine
The distance, d, between two finite cyclic groups Zm and Zn , is a function that describes how close these groups are to
being isomorphic. The smaller the distance, the closer these groups are to being isomorphic. This notion of closeness
has been made precise by a concept known as matching pairs. We prove Mark Lewers’ hypothesized upper bound for
10
Titles, Authors, Advisors and Abstracts
the distance. That is, If m jn, then the following holds: If m is odd, then d.Zm ; Zn / n2
2
.m2 /
.
2
.m2 C1/
.
2
If m is even,
then d.Zm ; Zn / n
We discover a necessary condition to obtain the distance, and give explicit formulae for
the distance between finite cyclic groups of small order. Our results also suggest that this distance function has all the
properties of a metric except the Triangle Inequality. However, we are able to show that the Triangle Inequality holds
for certain cases. We have discovered multiple lower bounds for the distance between the groups. However, our lower
bounds are not sufficient enough to verify the other cases of the Triangle Inequality. So, we plan to look for better
bounds for the distance in order to verify the Triangle Inequality and better approximate the distance.
22. The catenary degree of elements in numerical monoids of embedding dimension 3
Gautam Webb Colorado College
Reuben Tate University of Hawaii at Hilo
Advisor(s): Vadim Ponomarenko, San Diego State University
The catenary degree is an invariant that arises in the study of non-unique factorization theory. Given an element in
a numerical monoid, its catenary degree is an integer that gives a measure of the overall spread of that element’s
factorizations. Define the catenary set to be the set of possible catenary degrees of elements in a numerical monoid.
For numerical monoids of embedding dimension 2, the catenary set is well understood. We investigate the catenary
sets of numerical monoids of embedding dimension 3 and we present some results characterizing these sets for certain
monoids. In particular, we show that catenary sets can be arbitrarily large and we give conditions that determine when
catenary sets have exactly two elements. We also give a proof of a conjecture regarding the minimum nonzero catenary
degree of elements in a numerical monoid.
23. Towards the classification of unital 7-dimensional commutative algebras
Alexandria Yu MIT Primes and University School of Nashville
Advisor(s): Sherry Gong, MIT
We study the problem of determining how many unital commutative algebras exist in a given dimension and to find all
of these algebras.The motivation for this classification problem comes from number theory and algebraic geometry. For
dimension less than or equal to 6, Poonen has completely classified all unital commutative algebras up to isomorphism.
For dimension greater than or equal to 7, the situation is much more complicated due to the fact that there are infinitely
many algebras up to isomorphism. An algebra is called local if there exists a unique maximal ideal M. Local algebras
are basic building blocks for general algebras as any finite dimensional unital commutative algebra is isomorphic to a
direct sum of finite dimensional unital commutative local algebras. In this work, we classify all unital 7-dimensional
commutative local algebras up to isomorphism with the exception of the special case k.1/ D 3 and k.2/ D 3, where,
for each positive integer i , M.i / is the subalgebra generated by products of i elements in the maximal ideal M and
k.i / is the dimension of the quotient algebra M.i /=M.i C 1/. When k.2/ D 1, we classify all finite-dimensional unital
commutative local algebras up to isomorphism.
24. Omega Primality and Length Sets of Arithmetic Congruence Monoids
Bruce Zheng Texas A&M University
Advisor(s): Scott Chapman, Sam Houston State University
For positive integers a,b where a b and a2 a mod b, we define the Arithmetic Congruence Monoid (ACM)
M.a; b/ D fx a mod b W x 2 Ng. M.a; b/ is an commutative, cancellative, atomic monoid. For an ACM M and
nonunit x, we define that !M .x/ is the smallest positive
Q integer n such that whenever x j a1 at for atoms ai , there
exists a T f1; : : : ; tg with jT j n such that x j k2T ak . The omega function of M may be viewed as a measure
of how close x is to being prime. We give a formula for the omega function of any ACM. For any nonunit x of an
ACM M , we may describe the length set L.x/ of x, which is the lengths of all possible atomic factorizations of x. We
present some new results on the structure of the lengths sets of an ACM M.a; b/ when gcd.a; b/ is prime.
25. n-dimensional semi-hypercubes and the algebras associated with their Hasse graphs
Mitchell Lemons University of Wisconsin-Eau Claire
Austin Riedl University of Wisconsin-Eau Claire
Advisor(s): Colleen Duffy, University of Wisconsin-Eau Claire
The primary goal of our project is to determine the structure of a graded algebra, A.€n /, that is associated to the Hasse
graph, €n , of an n-dimensional semi-hypercube. We can consider a unit n-cube with one vertex at the origin, keep
Titles, Authors, Advisors and Abstracts
11
only those vertices with an even number of 1’s in its coordinate positions, and form new simplex and semi-hypercube
facets. The symmetry group of the semi-hypercube, and thus the automorphism group of the graph, is isomorphic to
Sn Ë Zn2 1 . Each symmetry of the n-dimensional semi-hypercube can be thought of as acting on the coordinates of
the vertices. For each symmetry, we consider the Hasse subgraph consisting of fixed k-faces of the semi-hypercube
under the action. From each Hasse subgraph, we determine the graded dimension of the subalgebra of A.€n / by
counting the directed paths between each pair of levels in the graph. We have determined the generating functions that
describe A.€n / under the action of elements from Sn Ë Œ0n2 1, and have a conjecture on the generating functions for
the remaining symmetries. This defines the complete structure of A.€n /.
26. Of Maximal Bullets of the Generators of a 3-Generated Numerical Monoid where !.n 1 /!.n3 /
Loren Anderson North Dakota State University
Advisor(s): Scott T. Chapman, Sam Houston State University
We examine the omega function for a 3-generated numerical monoid in a few general cases. For an element x in a
numerical monoid, !.x/ measures how far x is from being prime in the monoid. Unless the values of the generators
are known, it is difficult to compute !.x/ because the factorization of x is not necessarily unique. One way to ease
the calculation of !.x/ is to identify a maxi mal bul let of x. Here, we identify the possible sets of maximal bullets
of the generators of a 3-generated numerical monoid where !.n1 /!.n3 /. Some sets are unique to the !.n1 /!.n3 /
case, while others are not. Throughout, we determine if knowing the factorizations of maximal bullets is sufficient to
establish whether !.n2 / D !.n3 / or !.n2 /!.n3 / both when !.n1 /!.n3 / is known and unknown. Additionally, we
place bounds on gcd.n2 ; n3 /.
27. Some Sums on Additive Abelian Groups
Abraham Bekele University of Colorado Denver
Benjamin Wright Providence College
Katie Rosenberg Austin Peay State University
Advisor(s): Baginski, Paul, Fairfield University
The well studied classical Davenport constant of an additive finite abelian group, G, is the least natural number, `, such
that every sequence of length ` in G is guaranteed to have a nontrivial zero sum subsequence. The Davenport constant
arises naturally when studying nonuniqueness of factorization in algebraic number rings. The plus-minus weighted
Davenport constant is a variant of the classical Davenport constant, which allows weighting the elements of a sequence
P
by 1 or 1. A sequence S D s1 ; s2; : : : ; s` in G has a nontrivial plus-minus weighted zero sum if `iD1 ai si D 0
for some choice of ai 2 f 1; 0; 1g, with not all ai zero. The plus-minus weighted Davenport constant of G, D˙ .G/,
is defined as the least integer ` such that every sequence of length ` has a nontrivial plus-minus weighted zero sum.
Like the classical Davenport constant, the plus-minus Davenport constant arises in factorization problems, but it is also
used to construct dissociated sets in additive combinatorics. With this poster, we will introduce the general bounds on
the plus-minus weighted Davenport constant, describe the scientific process of computing Davenport constants, and
present the results of the Fairfield University REU.
28. Deformations of 5-Dimensional Associative Non Nilpotent Complex Algebras
Thao Tran University of Wisconsin-Eau Claire
Austin Riedl University of Wisconsin-Eau Claire
Hengzhou Liu University of Wisconsin-Eau Claire
Zach Forster University of Wisconsin-Eau Claire
Advisor(s): Michael Penkava, University of Wisconsin-Eau Claire
We have been studying the deformation theory of non nilpotent complex 5-dimensional associative algebras. By computing the miniversal deformation of an algebra, we can determine precisely to which algebras it deforms. This information tells how the moduli space of such algebras is glued together. In this poster we will explain some of the
methods we have used to carry out the computations. There are 285 isomorphism classes of algebras, including 16
1-parameter families. Each of these families is parametrized by a 1-dimensional projective orbifold, which is given by
CP 1 , modulo an action by the permutation group †2 . We have computed versal deformations for all of these algebras,
and have completed the study of these deformations for almost all of the algebras. We will present some highlights of
our results.
12
Titles, Authors, Advisors and Abstracts
29. Groups with Context-free Co-word Problem and Embeddings into Thompson's Group V
Heather Mathews Miami University
Rose Berns-Zieve Hamilton College
Dana Fry Mount Holyoke College
Johnny Gillings Morehouse College
Advisor(s): Reza Akhtar, Miami University
A conjecture of Lehnert states that a group has context-free co-word problem if and only if it is a finitely generated
subgroup of Thompson’s group V. Given a finite group G and a homomorphism W G ! G, one may define a
group V .G; / which generalizes Thompson’s group V . We prove that V .G; / has context-free co-word problem
by constructing a pushdown automaton and showing that the co-word problem of V .G; / is the cyclic shift of the
language accepted by the automaton. Since there does not seem to be an obvious embedding of V .G; / into V , our
construction is a candidate for a counterexample to Lehnert’s conjecture.
30. The Toppling Polynomial of a Graph
Megan Chambers Youngstown State University
Advisor(s): Luis Garcia-Puente, Sam Houston State University
Let G be a directed graph with a global sink. A sandpile c is a vector of nonnegative integers indexed by the nonsink vertices of G, where c.v/ represents the number of grains of sand at vertex v. A sandpile c is stable if, for all
vertices v, c.v) is less than the number of edges incident to v; otherwise, c is unstable. In the latter case, the unstable
sandpile c may be stabilized by a sequence of vertex topplings where an unstable vertex v topples sending one grain
of sand through each edge incident to v. The process of adding sand and toppling creates a Markov chain. It is of great
interest to study properties of the recurrent elements in this Markov chain. A principal avalanche is the sequence of
topplings that occurs when one grain of sand is added to a particular vertex on a recurrent sandpile of a graph. In 2004,
Cori, Dartois, and Rossin introduced the avalanche polynomial, which encodes the sizes of all principal avalanches
of a graph. We introduce the toppling polynomial, which encodes the specific toppling sequences of the principal
avalanches. This polynomial encodes more information than the avalanche polynomial. We provide general formulas
for the toppling polynomials of trees, cycles, wheels, and complete graphs.
31. Generalization of AES-Based Ciphers
Jeff Schreiner-McGraw Willamette University
Kevin Stubbs University of Maryland
Advisor(s): Liljana Babinkostova, Boise State
For the last three decades hash functions have been an essential element of the cryptography that is used for securing
computers and electronic communications. Their use is vital and the need to understand them in depth is crucial. The
Advanced Encryption Standard (AES), adopted by the US National Institute of Standards and Technology in 2001,
is the current symmetric key encryption standard and has widespread use in industry and government, from internet banking and wireless communications to data storage on hard drives. Several hash functions have been designed
using modified versions of AES internally. Our research centered on determining the algebraic structure of the sets
of encryption functions of AES-based ciphers used in these hash functions. We focused specifically on investigating
certain algebraic properties which have been proven to imply significant security weaknesses. Moreover, we define
almost translation based ciphers over an arbitrary finite field and study the permutation groups generated by the round
functions of such ciphers. We show that under certain cryptographic conditions this group is primitive, which generalizes some previous results. The class of almost translation based ciphers includes the ciphers of several well-known
hash functions.
32. Invariants of Generalized Arithmetic Numerical Semigroups
Sarah Hillier Lawrence University
Casey Fu Massachusetts Institute of Technology
Josiah Banks Youngstown State University
Julia Getsos Fordham University
Advisor(s): Vadim Ponomarenko, San Diego State University
A generalized arithmetic numerical semigroup is a numerical semigroup of the form S D ha; ah C d; : : : ; ah C xd i
where the gcd.a; d / D 1 and ax. Much is known for the arithmetic numerical semigroup (the case when h D 1), since
Titles, Authors, Advisors and Abstracts
13
for those semigroups there is a complete characterization of the length sets. Therefore, we explore various invariants
of the generalized arithmetic numerical semigroup. Specifically, we give an exact description for the following three
invariants: Specialized Elasticity, a measure of the maximum length of factorization of an element given its shortest
length of factorization, Omega-Primality, a measure of how far an element is from being prime, and Delta Sets, a
measure of how far apart factorizations are in length.
33. Computing the Hilbert series of invariant polynomials
Lauren Cowie Rhodes College
Advisor(s): Christopher Seaton, Rhodes College
Let G be a compact subgroup of GL.C n /, the group of invertible n n matrices with complex entries, and then
G permutes the polynomials in the ring CŒx1 ; : : : ; xn . Under this group action, there is an invariant polynomial
ring, the set of polynomials that are mapped to themselves under the actions. It can be quite simple to produce a
polynomial that is invariant under a group action, and knowing all of the invariants provides a lot of information about
the action; however, it can be computationally challenging to “count” the number of invariants of each degree. One
technique is to compute the Hilbert series, the power series whose nth coefficient is the dimension of invariants of
degree n. This talk will discuss computations of the Hilbert series in the case that G is isomorphic to a circle. In this
case, we can describe G in terms of a weight vector, an element of Zn , and the computation will be in terms of this
weight vector.
34. Error-Correction of Linear Codes via Colon Ideals
Benjamin Anzis University of Idaho
Advisor(s): Stefan Tohaneanu, University of Idaho
In this presentation, we show how errors in a received message can be thought of as codewords of minimum weight
in a new linear code. We then use techniques such as finding the primary decomposition of the saturation of a
homogeneous ideal associated with a code to find codewords of minimum weight and hence perform error-correction.
Finally, we show the equivalence, in this case, of saturating to coloning by the power of a variable, and how, in
the case of a unique error, finding the primary decomposition of the associated homogeneous ideal becomes
superfluous.
35. An Outside Analysis of the Mandelbrot Set
Dante Gates Rowan University
Advisor(s): Hieu D. Nguyen, Rowan University
Due to its fractal nature, much about the Mandelbrot set remains to be understood. While a series formula has been
proven to calculate the area of the Mandelbrot set, to date the exact value of this area remains unknown. The challenge
lies in computing the series coefficients which are recursively defined by a two dimensional sequence. We present
new approximations concerning the 2-adic valuation of the series coefficients. Moreover we use these coefficients to
generate high resolution plots of the Mandelbrot set to give an outside perspective of its fractal boundary.
36. A comparison of several algorithms of multidimensional continued fractions
Baudry Metangmo Montana State University Billings
Advisor(s): Saroj Aryal, Montana State University
Finding the most accurate algorithm to express a multivariate function as a generalized continued fraction stands as
an important problem until today. Generalization of a regular continued fraction algorithm to higher dimensions is not
trivial and has been studied for decades already. There are several algorithms available for expressing functions in two
variables as continued fractions, but their accuracy largely depends on a particular set of conditions. In this work, we
compare some of these algorithms, both analytically and numerically, with an attempt to find the one that works best
in most general situations. We primarily study the errors in the approximants of the continued fraction expansions of
some two-variable functions. We further explore convergence properties of the approximants.
14
Titles, Authors, Advisors and Abstracts
37. Infinite Products Arising in Paperfolding
Hadrian Quan University of California, Santa Cruz
Fernando Roman Kansas State University
Michole Washington Georgia Institute of Technology
Advisor(s): Victor Moll, Tulane University
The work here develops methods to evaluate certain infinite products in closed-form. These are finite products of values
of the gamma function. Presented here are infinite products of rational functions R(n) raised to the power of some
sequence Mn . The sequences satisfy certain regularity conditions as either a `-periodic or k-automatic. Of particular
interest is the Regular Paperfolding sequence considered by J. P. Allouche. We include a closed form expression for
Infinite Products to the power of an `-periodic sequence, and for a class of k-automatic sequences.
38. A Morphological Classification of the Arithmetic Sum of Two-Parameter Affine Cantor Sets with
C1=3
Sharat Chandra University of California, Irvine
Advisor(s): Anton Gorodetski, University of California, Irvine
It has been shown the sum of two affine Cantor sets either contains an interval or does not. Further it has been shown
that given parameters .; / 2 Œ0; 1 Œ0; 1 corresponding to middle- Cantor set, C , and middle- Cantor set, C ,
the arithmetic sum of C C C is a Cantor set if .; / is close to the origin, while the sum contains an interval if
.; / is far away from the origin. However, there exists a mysterious region in the parameter space, R where the
morphology was unknown. Boris Solomyak showed that for almost every point in R, the sum of the corresponding
Cantor sets contained an interval. Generalizing to more complex Cantor sets, and extending Solomyak’s argument, we
consider two-parameter affine Cantor sets C.˛;ˇ/ : Cantor sets where the thicknesses of the two generating sets are ˛
and ˇ respectively. We analyze the morphology of the sum of the standard 1=3 Cantor set with C.˛;ˇ / and find in the
˛ ˇ parameter space, there again exists a mysterious region where the morphology is unclear. We show for a.e point
in the parameter space, the sum does contain an interval.
39. Emergent Gastrointestinal Diagnoses and Procedures: A Nationwide Longitudinal Study of
Emergency Department and Inpatient Stays
Kimberly Siegler University of Milwaukee-Wisconsin
Advisor(s): Peter Tonellato, University of Wisconsin-Milwaukee
We performed a meta-data analysis of emergent gastrointestinal (GI) diseases and GI procedures from 2006-2010.
GI diseases contribute significantly to mortality, morbidity and healthcare spending in the United States. Anecdotal
evidence suggests that the 2008 recession affected emergent GI diseases, GI procedures and their associated costs.
We targeted 7 GI diseases and 4 GI procedures for analysis. We used the two largest publically available patient
record databases, totaling over 150 million patient records. Data was normalized to 2006 population and adjusted for
inflation. Using linear regression and ANOVA, our results show a change in the frequency of emergent GI diseases,
the frequency of selected GI procedures and related healthcare costs (p0.03) during the height of the recession. These
results suggest the recession may have influenced individuals decision in emergent healthcare use.
40. Low regularity local solutions to the generalized Magneto-Hydrodynamics Equations
Benjamin Schmachtenberger Creighton University
Lucia Magos Creighton University
August Hagen Creighton University
Advisor(s): Nathan Pennington, Creighton University
The incompressible magnetohydronynamics (MHD) system governs the dynamics of the velocity and magnetic fields
in electrically conducting fluids such as plasmas and liquid metal. The generalized magnetohydrodynamic (GMHD)
system differs from the standard MHD system by replacing the viscosity terms . 4/ by the generalized operator
. 4/˛ . Besides their mathematical interest, this allows the GMHD system to model more complex interactions than
the MHD system. In this paper, we prove the local existence of a unique short-time solution to the GMHD system with
initial data in non-L2.Rn / based spaces. These solutions are the first step toward establishing unique global solutions
in Lp .Rn / based spaces for p2.
Titles, Authors, Advisors and Abstracts
15
41. Towards a Stability Theory for Feynman's Operational Calculus in the Purely Discrete Setting | A
Direct Approach
William Graft Creighton University
Joshua Trebbien Creighton University
Advisor(s): Lance Nielsen, Creighton University
In this poster, we illustrate an abstract formulation of Feynman’s operational calculus which was originated by B.
Jefferies and G. W. Johnson. This formulation of the operational calculus allows Feynman’s heuristic ‘rules’ for
the formation of functions of several noncommuting operators to be applied in a mathematically rigorous fashion.
Consider a nicely behaved function, f of bounded linear operators A1 ; : : : ; An on a Banach space X. The operator
f1 ;:::;n .A1 ; : : : ; An / is called the disentangling
of the function f . Given continuous Borel probability measures
˚
1
1 ; : : : ; n on Œ0; T , select sequences j;k kD1 , j D 1; : : : ; n, of purely discrete finitely supported probability measures on Œ0; T  which converge weakly to j . Using Feynman’s operational calculus for purely discrete measures,
compute the disentangled operator f1;k ;:::;n;k .A1 ; : : : ; An /: We are working to establish, in a direct way,
lim f1;k ;:::;n;k .A1 ; : : : ; An / D f1 ;:::;n .A1 ; : : : ; An /:
k!1
Knowing how to compute such limits directly should lead to insights into the operational calculus and related evolution
equations.
42. Dynamics of Folds with Boundary Dependent Ratios
Nicolas Hyde St. Edwards
Advisor(s): Michael Saclolo, St. Edwards
We examine the behavior of a linear system of difference equations generated from the process of “folding” between
two smooth curves. Take an arbitrary transversal of these two curves. Then use the following rule: to create a new
transversal, take the angle between the most recently constructed transversal and the corresponding curve and let the
new transversal be alternately folded by two separate ratios, depending on which curve the transversal originates.
We then examine the sequence of angles that these transversals create and discover what they converge to. We use
analytical and as well as linear algebraic techniques to analyze the system and to generate the solutions.
43. Ergodicity and Conservativity of Products of Infinite Transformations and their Inverses
Isaac Loh Williams College
Advisor(s): Cesar Silva, Williams College
Weak mixing for finite measure-preserving transformations has many interesting equivalent characterizations, among
which is that T T is ergodic. This in particular implies that T T 1 must be ergodic. It has been known for some
time that many of these characterizations do not remain equivalent in the infinite measure-preserving case. In this
presentation we will construct infinite measure-preserving rank one transformations such that T T is ergodic but
T T 1 is not ergodic, and other related examples. This partially answers a question of Bergelson. The methods
are combinatorial and probabilistic and use the notion of descendants in the constructions of the transformations. In
parallel, we prove that for all rank-one transformations T , the product T T 1 is always conservative. We also show
that the situation is different in countable state Markov shifts.
44. A New Class of Basis Polynomials Derived from the Generalized Prouhet-Thue-Morse Sequence
Jennifer Crawford Rowan University
Advisor(s): Hieu D. Nguyen, Rowan University
The famous Prouhet-Thue-Morse (PTM) binary sequence f0; 1; 1; 0; 1; 0; 0; 1; : : : g has important applications in coding theory, number theory, and combinatorics. Its product-generating function is well-known and has been used to
identify sets with equal sums of like powers. We consider the product-generating function of the PTM sequence generalized to base p in terms of a new class of basis polynomials. New explicit formulas are given for these polynomials
in terms of eigenvectors and eigenvalues derived from their recurrence matrices.
16
Titles, Authors, Advisors and Abstracts
45. An exploratory investigation using electroencephalography and machine learning techniques for
fine motor classification in the EggLink brain-computer interface.
Chrono Nu Westminster College
Advisor(s): Richard Wellman, Westminster College
The EggLink is an electroencephalography-based brain-computer interface that will operate common household computers based on numerous data transformation techniques and machine learning classifications of fine motor signals
generated in the brain. At this stage, the project involves exploring how subtle differences in measurable neurological signals can be revealed mathematically for the prototype system that classifies fine motor activity (the flicking of
each of 10 fingers) to a high degree of accuracy. The performance of the system will be discussed in terms of overall
accuracy of true positive predictions made by machine learning ensembles.
46. Risk Taking in Epidemiology
Andrew Latham Western Carolina University
Aaron Moose Western Carolina University
Solvei Stenslie Concordia College
Advisor(s): Daniel Maxin, Valparaiso University
We study several epidemic models that incorporate risk taking behavior as a response to an effective treatment or
vaccine. We assume that knowledge about the number of recovered individuals has an effect in increasing the contact
rate between susceptible and infectious individuals. We show that a relatively mild risk-taking behavior response
changes the fate of an epidemic from disease clearance to disease persistence. Moreover, under certain conditions
on the parameters, increasing the efficiency of treatment or vaccine has a counter-intuitive and unwanted effect of
increasing the reproductive number suggesting a wider range of diseases may become endemic due to risk taking
alone. These results indicate that the manner in which treatment/vaccine effectiveness is advertised can have a dramatic
influence on how the epidemic evolves.
47. Investigating the Dependence of Transmission Rate to Water Temperature in a Host-Parasite
System
Mary O'Driscoll University of Wisconsin - La Crosse
Advisor(s): James Peirce, University of Wisconsin - La Crosse
Every year, thousands of waterfowl around the upper Mississippi River are dying from parasites transmitted to them
through an invasive species of snail. The parasite species exhibit temperature-dependent transmission patterns with no
transmission occurring when temperatures either fall below or exceed certain thresholds. The transmission window
overlaps the waterfowl’s seasonal migrations. The purpose of this study was to determine how the temperature of
the water affects the transmission rate of the parasite. First, we moved the range and the location of the temperature
window for parasite transmission. Second, we created an algorithm that defined the transmission parameters based
upon the length of time temperature spends above or below a known threshold. In both studies, the size of the infected
host populations depended on the temperature window and accrued time. Results reinforce biological observations and
emphasize why it is important to study transmission at various temperatures in the laboratory.
48. Computational Modeling of the Thyroid Hormones Homeostasis and its Manipulation by Chemicals
Catherine Read UNC Asheville
Kai Bartlette Providence College
Ross Donatelli Florida Institute of Technology
Advisor(s): Hisham El-Masri, Environmental Protection Agency
To describe the interaction between propylthiouracil (PTU) and thyroid hormones, a thyroid hormone production
model was created and connected to a physiologically-based pharmacokinetics (PBPK) model of all organs. As a treatment for hyperthyroidism, propylthioracil inhibits hormone production in the thyroid gland. Using a basic schematic
of thyroid hormone creation to model the previous phenomena, equations were derived, adapted, and parameters were
optimized from Ekerot et al. dog data. A physiologically-based pharmacokinetic model was then created to provide a
basis for concentration equations. After solving the differential equations for each organ, parameters were optimized
Titles, Authors, Advisors and Abstracts
17
to Cooper et al. experimental rat data. A dynamic model for T4 and T3 production required linking the thyroid in the
PBPK model to the original thyroid kinetics model.
49. The Effective Conditions of an Open Limestone Channel in Treating AMD
David Wolfe Saint Francis University
Advisor(s): Ying Li, Saint Francis University
Acid mine drainage (AMD) is the outflow of acidic water from metal mines or coal mines. When exposed to air and
water, metal sulfides from the deposits of the mines are oxidized and produce acid, metal ions and sulfate, which lower
the pH value of the water. An open limestone channel (OLC) is a passive and low cost way to neutralize AMD. A
mathematical model has been created to numerically determine the change in pH of the water and the concentrations
of species from the dissolution of calcium on the surface of the limestone into the acidic water. The model is used to
predict the conditions in which a OLC would be an effective solution for AMD. Effective ranges will be determined
for the concentrations of calcium and iron, as well as the temperature and velocity of the water.
50. Polynomial expansion method for the numerical solution of the Lenard-Balescu equation
Andrew Belt University of Tennessee, Knoxville
Nathan Ng University of Maryland
Advisor(s): Susana Serna, Universitat Autnoma de Barcelona
We explore an approach to solving the kinetic equation for plasmas by expanding the solution in orthogonal polynomials to obtain a system of coupled ordinary differential equations (ODEs) of the expansion coefficients. Here we
consider a one-species plasma described by the quantum Lenard-Balescu equation. We first obtain the system of ODEs
for a simplified version of the quantum Lenard-Balescu equation by excluding the dielectric function, which accounts
for electrostatic screening. Taking the expansion coefficients of common distributions as initial conditions, we are able
to numerically solve this system using certain computational techniques. For all tested initial conditions, the distribution function evolves to the expected equilibrium state on physical time scales, conserving both particle number and
energy throughout the process. From the simplified quantum Lenard-Balescu equation, we take the classical limit to
obtain the well-studied Landau equation. Finally we derive an exact expression for the dielectric function given an
arbitrary distribution function and consider its various physical limits.
51. Dynamical Model of Consumer Rumor Transmission in a Twitter-like Network
Jessica Bustamante Westminster College
Advisor(s): Janine Wittwer, Westminster College
True or false, consumer rumor (i.e., rumor that targets a company and/or its products) is challenging to control and
poses a real threat to company reputation, earnings, and employment. As our title suggests, our talk will discuss the
construction and application of a dynamical model of consumer rumor transmission in a Twitter-like network. Twitter
is different from most online social networking sites because it allows for communication that is non-symmetric.
Using difference equations, rather than differential equations, we determine under what conditions a consumer rumor
is likely to infect so many nodes as to make a significant impact on a company’s bottom line.
52. A Mathematical Model for Setting Crime Reduction Targets
Anthony Gusman Vanguard University
Advisor(s): Yoon Sik-Cho, University of Southern California
This project investigates the potential of improving the Los Angeles Police Department’s (LAPD’s) ability to forecast
future crime rates along time scales of months to years. The current forecasting approach does not account for changes
in long-term trends or seasonality in crime data. Our study develops models that extract long-term and seasonal factors, offering a data-informed method to support crime reduction target setting. The patterns detected can be used to
predict future crime rates. These predictions are compared with those of the department’s current method. Accuracy
is quantified using standard goodness-of-fit tests. Furthermore, a graphical user interface is developed to provide an
environment for comparing forecast model predictions. This can also serve as prototype software for LAPD personnel.
Our approach could be extended to other agencies as well.
18
Titles, Authors, Advisors and Abstracts
53. A Simulation Model For Backup-System Capacity Planning
Caleb Miller California Polytechnic State University San Luis Obispo
Advisor(s): Erim Kardes, UCLA RIPS Program
Due to the growing amount of data that clients need to store, effective procedures to predict the future size of backup
storage systems are becoming increasingly important. This report introduces a simulation approach to forecast the
growth of a backup-system. Time series models and machine learning techniques such as autoregressive integrated
moving-average and clustering algorithms, respectively, are used to characterize the behavior of clients’ workloads
and then forecast future development. We developed forecasting software for predicting the storage size given parameters such as backup sizes and deduplication rates. The model gives a distribution of time for when capacity will be
reached. Using synthetic datasets, we compared our results with those of Mark Chamness of the EMC Corporation,
who published an alternative method to predict backup-system storage.
54. Calculating the Channel Capacity of Satellite Communication Systems
Peter Morfe Cooper Union
Kathreen Yanit University of Guam
Elizabeth Spencer University of Maryland - College Park
Advisor(s): Stephen DeSalvo, University of California - Los Angeles
The Aerospace Corporation sponsored this Research in Industrial Projects for Students (RIPS) project in order to
compute the channel capacity of satellite communication systems, particularly the Mobile User Objective System.
The research can be divided into two categories: mathematical modeling of the channel and development of numerical methods for computing the channel capacity. First, the channel models incorporated two and three sources of
fading, as well as frequency selectivity. The two-source fading channel models a cascade of ionospheric scintillation
and terrestrial multipath. The three-source fading channel models ionospheric scintillation, terrestrial multipath and
shadowing, based on the Lutz 2-state model. Three methods were developed for computation of channel. The first
method computes the capacity using the pdf of the combined fading process. The other method computes the capacity
by integrating against the pdfs of the individual sources of fading. Lastly, the capacity is approximated using a time
average.
55. The Black-Litterman Model: Exploring the Sensitivity of Inputs
Timothy Dombrowski Saint Leo University
Jessie Byrnes Morningside College
Jahlani Clarke The University of the District of Columbia
Dorian Kandi Morehouse College
Advisor(s): Tao Pang, North Carolina State University
In modern portfolio optimization theory, the goal is to identify the optimal portfolio. Optimal portfolio weights are
obtained by either maximizing mean portfolio return or minimizing portfolio risk. The classical mean-variance optimization (MVO) model uses historical mean returns and a historical covariance matrix as inputs, but the optimal
portfolio tends to be infeasible in practice and lacks diversification. The Black-Litterman (BL) model integrates investors’ views and their confidence in these views into the MVO model through Bayesian estimation. In this project,
we investigate the sensitivity of investors’ views in the BL model by using real market data. We also determine the
parameters’ effects on the optimal portfolio. Our results are validated by back testing with actual market returns and
by simulating returns with multivariate normal distributions. Through our research, we discover that the sensitivity of
the parameters in the BL model is based on many factors, which falls inline with the intuition behind the model. We
also prove the convergences of several parameters from a theoretical standpoint. Using our research, investors are able
to better understand the impact of their views, as well as other parameters, on the optimal portfolio in the BL model.
56. Self-organized criticality for optimal random search
Michelle Randolph UNC Asheville
Kathleen Donahue Harvard University
Advisor(s): Aliki Mavromoustaki, UC Los Angeles
Self-organized criticality (SOC) is a phenomenon found in dynamic systems ranging from rice piles to solar flares.
HRL Laboratories discovered that shapes generated by these systems produce search patterns that are well suited for
Titles, Authors, Advisors and Abstracts
19
approximating solutions to certain hard optimization problems. Improving the efficiency of random search could have
widespread application (supply chains, allocation of resources, transportation networks). We investigated the properties and performance of shapes generated by an SOC model, the Bak-Tang-Wiesenfeld sandpile model and compared
them to other search patterns for optimizing the Ising spin glass. Specifically, we examined: fractal dimension of the
boundary, size-to-boundary ratio, variation in shape, size distribution and boundary structure.
57. Evolutionary games on cycles
Christopher Paoletti Emmanuel College
James Withers Emmanuel College
Advisor(s): Christine Sample, Benjamin Allen, Yulia Dementieva, Emmanuel College
Evolutionary game theory is a mathematical approach to studying the evolution of social behavior. Interactions that
affect reproductive fitness are conceptualized as games and spatial population structures are represented as graphs.
Using this framework, we investigate the effects of self-interaction on the evolution of cooperative strategies. In our
model, the population is represented as a cycle with self-loops and individuals can employ one of two competing
strategies in a matrix game. Interaction and replacement can be governed by the same graph or by two different
graphs, which can be weighted or un-weighted. We calculate fixation probabilities for these strategies and derive exact
conditions for natural selection to favor one strategy over the other. We determine the behavior of these conditions
in the limits of weak selection and large population size, and show that the limiting conditions do not depend on the
order in which limits are taken. We also consider evolutionary success when mutation is present. All calculations are
performed for two different update rules (birth-death and death-birth). We conclude that self-interaction can promote
the evolution of cooperative strategies in spatially structured populations.
58. A Geometric Approach to Voting
Lee Fisher Appalchian State University
Advisor(s): Vicky Klima, Appalachian State
Arrow’s theorem informs us that while some voting methods may be more fair than others, a certain set of reasonable
fairness conditions cannot simultaneously be satisfied by any one voting system. Well-known voting systems include
the plurality method, where voters are told to pick just one candidate and the candidate who receives the largest number
of votes wins; and the Borda count, where candidates are ranked and given points in equally-spaced descending order.
This presentation proposes a new voting system which functions as a generalization of the Borda count. In our new
system, we assign a value of one to first place, zero to last place, and intermediate rankings are assigned values in
arbitrary intervals of non-increasing order. We then use geometric methods to consider the winning results space
associated with each candidate. We will compare this voting system against the fairness criteria stated in Arrow’s
theorem. We will show this voting procedure is monotone, equivalent to the Borda count only in the three-candidate
case, and like both the plurality and Borda count methods does not necessarily satisfy the Condorcet winner criterion;
which selects the candidate who wins each pairwise comparison to win the election.
59. Determining Sufficiency for Comparison in the Friction Ridge Impression Examination Process
Jason Zellmer University of Wisconsin-Milwaukee
Advisor(s): Gemechis Djira, South Dakota State University
Friction ridge impressions, better known as “fingerprints” , have gained recognition for their use in identification
purposes due to their uniqueness among individuals and their persistence throughout an individuals’ lifetime. Two
characteristics that are used in the identification process are: the quality of the print and the quantity of unique characteristics of the print. Previous research has focused on the lack of examiner consistency in the determination of
sufficiency for comparison based on the quality and the number of minutiae (unique characteristics) of the fingerprint.
Sufficiency of determination is a subjective calculation of the examiner that states whether the print in question has
enough unique characteristics and is of good enough quality that it will likely yield a match when comparing prints.
This research developed an ordinal logistic regression model and examined the predicted usefulness of fingerprints
(measured by their ordered categorical rank when searching a database for a potential match) in an effort to determine
a sufficiency standard for a fingerprint to be considered useful in casework. The results of this project determined
that the minimum standards for the quality and quantity of minutiae of a print may be higher than those currently
recommended.
20
Titles, Authors, Advisors and Abstracts
60. Matching video segments with relevant documents
Ilan Morgenstern Instituto Tecnol´ogico Aut´onomo de M´exico
Lingxin Zhou Rutgers University
Advisor(s): Bin Bi, UCLA
The USC Shoah Foundation has collected over 52,000 video testimonies from survivors and other witnesses of the
Holocaust. The goal of this research project is to improve the learning experience for the people who watch the
testimonies by integrating external knowledge to the archive. Since the videos are indexed by keywords for each minute
of video, we use these data to extend various information retrieval methods to match each segment to a Wikipedia
article; these methods include the vector space model approach, latent semantic indexing, and statistical approaches.
Based on the existing techniques, we propose a new method that exploits the metadata of the document corpus to
match the video segments with relevant documents.
61. Parameter Selection Techniques for Nonlinearly Parameterized Models
Jared Cook Asbury University
Nicholas Myers University of Wisconsin-Milwaukee
Nina Ning George Washington University
Advisor(s): Ralph Smith, North Carolina State University
In this research, we apply parameter selection techniques to nonlinearly parameterized models for disease dynamics.
The objective is to develop techniques to determine the sets of identifiable or influential parameters in these models.
We first illustrate the use of randomized singular value decomposition (SVD) and QR techniques to quantify these
sets for linearly parameterized problems. We then demonstrate the use of the global Morris screening, Sobol analysis,
and active subspace techniques for nonlinear problems. For the considered example, we also show that non-identifiable
parameters can also be determined by considering pairwise correlation plots obtained through Bayesian model calibration. Parameter selection in this manner is a necessary full step before employing Bayesian and uncertainty propagation
techniques to quantify the uncertainties associated with model outputs.
62. Computational Modeling for Optical Coherence Tomography of the Human Eye
Camille Sanchez Mount St. Mary’s University
Asha Cyrs California State University Fullerton
John Maxwell III North Carolina A&T State University
Leah Frederick Pennsylvania State University
Advisor(s): Mansoor Haider, North Carolina State University
Optical coherence tomography (OCT) is a medical imaging technique that uses light waves to produce high-resolution
images of internal tissue microstructures. Using data obtained from OCT images, it should be possible to compute local
curvature maps of the human cornea that exceed the accuracy of existing keratometric technology. However, accurate
three-dimensional reconstruction of the cornea is insufficient to ensure comparable accuracy in curvature maps because
curvature computation is extremely sensitive to small errors in the surface measurements. This project goal was to
develop theoretical cornea models to validate the curvature mapping algorithms and analyze the propagation of certain
types of error. Models were sampled on a hexapolar grid to simulate data from an OCT scanner. Interpolation of the
simulated data was done using Zernike polynomials, which represent a complete orthogonal basis on a unit disc. The
main sources of error considered were perturbations in the surface representation and noise in the OCT image data.
The perturbations that were considered were a result of polynomial over-fitting, and the noise in the data was simulated
using Gaussian white noise. Adjusted models accounting for different corneal disorders were also taken into account
and implemented.
63. Estimating Liquidity Risk Using Exchange Traded Funds
Claire Kelling Virginia Tech
Hannah Li Pomona College
Evan W Witz Wisconsin Lutheran College
Rachel Crowell University of Missouri- Kansas City
Advisor(s): Marcel Blais, Worcester Polytechnic Institute
The financial crisis of 2008 starkly illustrated the perils of ignoring liquidity risk. This risk is particularly crucial to
index and ETF providers that explicitly guarantee the provision of underlying shares through time and implicitly agree
Titles, Authors, Advisors and Abstracts
21
to absorb the associated liquidity risks and costs. In this paper we propose an ETF specific liquidity measure. The
proposed model isolates liquidity risk from other factors by forming a portfolio of buying the ETF and shorting the
underlying basket of securities using the weighting system employed by the ETF issuer. This work is closely related to
a similar measure proposed by Chako et al which we find while useful provides counterintuitive results during times
of fast changing liquidity environments. Since liquidity crises typically evolve rapidly, the time sensitive nature of our
measure is a particularly important feature. Our analysis is performed on JNK and its underlying securities. We also
compare the proposed liquidity metric to those existing in the literature and industry standard measures. We find that
our measure captures the shifts in liquidity in a conservative and timely fashion making it particularly applicable to
risk management purposes for ETF investors and providers.
64. Applying Queuing Theory to a Queuing System
Su Ji Hong California Lutheran University
Advisor(s): Villalpando, California Lutheran University
Queuing theory uses flow balance equations derived from steady states to determine the probability of each state;
in queuing theory a state represents the number of customers present at the system. I modeled the queuing system
at the Centrum caf´e on California Lutheran University campus assuming that the arrival and service rates follow
exponential distribution. The queue system at the register followed M=M=1=FCFS=1=1 : the arrival and service
times are independent, there is one server, it has a first come first serve queue discipline, infinitely many people can
be in the queue, and there are infinite many people from which customers are drawn. I figured out the arrival rate, ,
and service rate, . I computed the waiting time and compared it to the data. I simplified the model at the kitchen to
M=M=1=FCFS=1=1. Then I connected them using the fact that depends on of the register. After calculating the
probability of each steady state, I calculated the queue time at the register and the kitchen.
65. Model Refinement Through Verification and Validation Methods and Anomaly Detection
Jordan Menter Smith College
Hannah Schwarze Saint Mary’s University of Minnesota
Advisor(s): Hien T. Tran, North Carolina State University
Mathematical models can be assessed and refined using two separate procedures: verification and validation (V&V),
and anomaly detection. Together they determine a model’s reliability. Verification assures a model implements the
developer’s concepts correctly, and validation assures the model is an accurate representation of a real world system.
The purpose of anomaly detection is to identify features of the model or data set that throw off the model’s reliability.
We looked at a mechanistic model that takes a system of differential equations describing the dynamics of Hepatitis
C and its treatment system. The system includes the triple drug combination of Ribavirin, Pegylated Interferon-alpha,
and Telaprevir (a new Hepatitis C drug). In addition, we validated and verified a data-based model created by NCSU
REU students in the summer of 2013. By using PITCH f/x data from Major League Baseball and machine learning
methods, they created a pitch prediction system that decides whether or not the next pitch will be a fastball. Through
V&V and anomaly detection, we have developed a level of confidence in both models and are able to make suggestions
for future refinement.
66. A Locally Adaptive RBF-FD Method
Wade Meyers University of Wisconsin-Stout
Talin Masihimirzakhanian California State Polytechnic University-Pomona
Advisor(s): Keith Wojciechowski, University of Wisconsin-Stout
Conventional Radial Basis Function (RBF) methods for numerically solving partial differential equations use global
approximations resulting in dense matrices that grow in size if data refinement occurs. The Radial Basis Function –
Finite Difference (RBF-FD) approach is a local approximation method that utilizes nearest neighbor data and yields
a sparse implementation. Unfortunately RBF-FD differentiation matrices have fixed stencils and the approximations
can lose accuracy under refinement. In this paper we propose using local approximations with locally adaptive stencils
that take advantage of the features of both global and local approximations. In this approach the stencil sizes stay
fixed where the solution is smooth but grow in size only where refinement is needed. The advantage of this method is
that it is computationally efficient and stable much like the RBF-FD method but offers comparable accuracy to global
approximations with significantly lower computational cost.
22
Titles, Authors, Advisors and Abstracts
67. Recognition of Textural Differences in Infrared and Ultraviolet Imagery Using Fractal
Characteristics
Jack Ryan North Central College
Advisor(s): Charles Tolle, South Dakota School of Mines and Technology
Due to the complexity of the world, there are many aspects of nature that cannot be completely or correctly analyzed using Euclidean geometry. For that reason, the study of fractals is of great interest when attempting to study
the natural world. This project utilized an algorithm written by Troy Thielen which is used to estimate the fractal dimension and lacunarity of gray-scale images. New ultraviolet and infrared images were taken throughout the
project. An extensive portion of the project was devoted to the process of aligning the multi-spectral images so that
they line up correctly. This was necessary due to the way that light bends in the different spectrums; the images do
not line up correctly without intervention. The fractal characteristics of these images were then to be analyzed to
recognize textural differences, a useful method when attempting to recognize objects or locate targets, however the
algorithm was found to be insufficient in dealing with the complexities of the multi-spectral imagery. Instead, this
research project provides a foundational background that suggests a method for image registration that can be used
for future research projects that include revising the algorithm in order to handle the higher-dimensional nature of the
multi-spectral images.
68. The Impact of the Effectiveness of Needle Exchange Programs on Addiction-Treatment Dynamics
Natalia de la Torre Mary Baldwin College
Mayra Coronado Regis University
Advisor(s): Muntaser Safan, Aprillya Lanz, Arizona State University, Norfolk State University
The aim of this research is to determine the impact of needle exchange programs as motivation for intravenous drug
users to seek treatment for addiction. A mathematical model of the dynamics of a population of drug users that
incorporates a needle exchange program is formulated. We define the basic addiction reproduction number for the
proposed model and explore its role in the prevalence and control of needle-sharing drug addiction. Specifically, the
local stability of the injection-addiction free and endemic equilibria are determined.
69. Artificial Intelligence of Modern Board Games: Battle Line
Daniel Bladow Gonzaga University
Advisor(s): Sean McCulloch, Ohio Wesleyan University
Many turn based board games played by computers use game state trees to determine what moves to make. Traversing
through game state trees is time consuming, especially with random factors such as a deck of cards. Battle Line is
a game where two players are competing over nine different three card poker hands using a 60 card deck with six
suits and number values one through ten. Each hand is called a flag and the type of hand (such as a straight flush) is
called a formation. Together, the deck and flags create an exponentially large number of game states. We created an
artificial intelligence using a probabilistic approach to approximate future game states. We estimate future game states
by determining the top formations we could make by playing a card from our hand on each flag. Then we calculate
the odds of completing each formation, of each formation winning, and finally of any of the top formations resulting
in a win on that flag. There are two main advantages to this approach. First of all, it is less time consuming and
can be played on weaker computers, possibly cell phones. Secondly, the computer’s thinking avoids ad-hoc strategic
knowledge. Therefore, the computer is both more time efficient and capable of devising unique strategies based on
mathematical principles.
70. Dynamics of Co-orbital Moons Near Collision
Kimberly Stubbs UNC Asheville
Advisor(s): Samuel R. Kaplan, UNC Asheville
This project is about celestial mechanics and dynamical systems. Specifically, the goal is to explore the techniques
used in modern celestial mechanics to analyze near-collision dynamics and chaos. The model we’re working with is
a 3-body co-orbital system. Josep Cors and Glen Hall wrote a paper on 3-body co-orbital systems and determined
when the moons will pass each other and/or change orbits. They were only interested in these two occurrences, and
so they left out the dynamics of near-collision. We’re interested in finding out what happens near collision of the two
Titles, Authors, Advisors and Abstracts
23
moons and have done the necessary change of variables to allow analysis of the dynamics and chaos. We’ll look into
the dynamics and what they mean for the entire system.
71. Mathematical Modeling and Optimal Control of Alternative Pest Management for Alfalfa
Agroecosystems
Cara Sulyok Ursinus College
Julia Senkevich Ursinus College
Advisor(s): Mohammed Yahdi, Ursinus College
Alfalfa is the most cultivated forage legume in the world and is used primarily to feed cattle. The pest, potato leafhopper
(PLH), damage to the host-plant alfalfa is costly and chemical pesticides are unsafe. The goal is to investigate alternative farming techniques for pest management to minimize the alfalfa damage. More precisely, based on data and
results from field experiments that integrate enemies hypothesis (predators) and polyculture farming (plant diversity)
approaches, this project developed a mathematical model for designing cost-effective and environmentally-safe control
strategies to minimize the plant damage. A mathematical model of a system of non-linear differential equations was
created and shown to accurately fit results from open-field experiments and to predict outcomes for scenarios not covered by experiments. Optimal control theory was used to determine optimal and practical pest management strategies
to minimize the plant damage and the impact on the revenues from milk production for farmers. Tools used include
non-linear systems of differential equations, Bang-Bang controls, Pontryagin’s Maximum Principle, and computer
simulations.
72. Improving the Error-Correcting Code Used in 3-G Communication
Tahseen Rabbani University of Virginia
Sung-Min Park University of Virginia
Advisor(s): James Davis, University of Richmond
Consider a digital message sent through a noisy channel with an undesirable amount of interference. Error-correcting
codes encode messages in such a way that the recipient is able to correct bit errors up to a certain threshold. In 2011,
Samsung Electronics Co. filed a complaint against Apple Inc. for alleged infringement of patents concerning errorcorrecting codes. One of the primary patents in question was a Œ30; 10; 10 non-cyclic code which was implemented
in all devices communicating on the 3-G network, including many Apple products. However, in the patent, Samsung
leaves the derivation of the basis for their code rather ambiguous. We first present an explicit construction of a similar
Œ30; 10; 10 idempotent code. Using elementary manipulations of our code, we then develop a novel and improved
Œ30; 10; 11 non-cyclic code with stronger error-correcting capabilities.
73. 3-D Data Fitting and Inverse Kinematics Algorithms for Robot Mimicking of Human Action
Jared Brown St. Olaf College
Christie Mauretour Florida Gulf Coast University
Julian Sass University of Maryland Baltimore County
Ziyue Shuai Bryn Mawr College
Advisor(s): Thomas H¨oft, University of St. Thomas
The development of techniques allowing artificial intelligences to learn and adapt in real time remains an open question
in Machine Learning, Robotics, and related fields. In this vein, we approach the problem of instructing a general robotic
frame to mimic actions as demonstrated by humans. The solution to this problem requires the solving of two separate
systems. The first, dubbed the Forward Kinematics Problem, involves describing the motion of the human. To this end,
we develop a robust linear interpretation of a geometric model of the human frame as well as the transformations on that
model allowing motion through time. We additionally apply regression techniques to fit this model to 3-dimensional
point-cloud data about the current orientation of a particular human subject. The second problem to be solved, the
Inverse Kinematics Problem, entails mapping end-effector position (usually the tip of a hands on the human subject)
to the robot. This being a problem without a unique solution, we apply constraints to find the single optimal path which
the robotic frame should follow.
24
Titles, Authors, Advisors and Abstracts
74. Relationship among FKBP5 genotype, serum cortisol, and cognitive function in aging humans: a
Project FRONTIER study
Cody Tucker Texas Tech University
Advisor(s): Breanna Harris, Texas Tech University
Stress induces physical responses in organisms, one being activation of the hypothalamic-pituitary-adrenal axis, resulting in release of glucocorticoids (cortisol). Prolonged elevated levels of cortisol have been shown to negatively
affect cognitive function in humans. Cortisol exerts its effects by binding to intracellular receptors, therefor factors
associated with receptor sensitivity may affect cortisol actions. One of these factors is FKBP5, a cortisol receptor
co-chaperone, a single nucleotide polymorphism of this gene, rs1360780 (T allele), has been associated with impaired
negative feedback and prolong cortisol exposure. We predict that high levels of cortisol in addition to the T allele of
FKBP5 will lead to an enhanced decline in cognitive function over time. Using human data from Project FRONTIER
(n D 193, 140 females, 53 males, age 40–87), funded and maintained by the F. Marie Hall Institute for Rural and
Community Health, we will measure serum cortisol via radioimmunoassays at two time points (baseline and 3 years).
We will determine FKBP5 genotype by PCR. We will assess how these two variables relate to cognitive function
at baseline and 3 years as measured by the Mini-Mental Status Examination (MMSE), which assesses five areas of
cognitive function.
75. Exploring Phase Space Localization Using a Coherent State Basis in Bargmann Space
Jacob Goudreau Westfield State University
Advisor(s): Brian Jennings, Westfield State University
In 1961 Valentin Bargmann published a paper in which he outlined the structure of a new kind of Hilbert space.
“Bargmann Space,” as it came to be known, contained complex-valued analytic functions. A decade later Bargmann,
et. al, published a new paper wherein they established a basis for Bargmann Space. Building upon the work of the
great John Von Neuman, they asserted it was possible to form a minimal spanning set using a subset of coherent
states. They considered a lattice superimposed over the complex plane and showed that by constructing this lattice
such that each cell bounded an area equal to 2 „, a set of coherent states centered at each lattice point formed the
minimal spanning set they were looking for. While it is known that the coherent states about the lattice points do form
a basis in Bargmann Space, a rigorous study regarding its utility had not yet been undertaken. I began by examining
the approximating capability of the coherent state basis with respect to well-known wave functions, like the harmonic
oscillator basis states. After working with the harmonic oscillator states, I shifted my focus to the coherent states basis’
ability to approximate the more general class of semiclassically-localized wave functions.
76. Sediment Transport at Wallops Island, Virginia
Joshua Updike Shippensburg University
Advisor(s): Ben Galluzzo, Shippensburg University
Wallops Island, Virginia is home to NASA’s principal suborbital research facility and the United States Navy’s Surface Combat Systems Center; it is also extremely prone to shoreline erosion. Due to over $1 billion in facilities and
resources on the island, shoreline maintenance and restoration is of utmost importance. The poster presentation will
cover the motivation for the research, data collection techniques and the mathematical processes used to create a
model for sediment transport on Wallops Island and the region. In particular, we will highlight the interaction between
geology, physics and mathematics as well as the engineering of a new single-wave sediment collection device.
77. Correcting Holographic Images Distorted by Atmosphere
Thomas Welle University of St Thomas
Anne Fink University of St Thomas
Advisor(s): Thomas Hoft, University of St Thomas
When a holographic image of a distant object is recorded it is often blurred due to atmospheric turbulence. Phase
screens constructed from radial disk harmonics can be used to correct for these aberrations. Since in digital holography systems the atmospheric turbulence phase distortion is not know, the correcting phase screen must be found
computationally. We use optimization methods and an image sharpness metric to compute approximate corrections
Titles, Authors, Advisors and Abstracts
25
which substantially improve the blurred imagery. This paper discusses the implementation and effectiveness of ten optimization methods, including gradient descent, simultaneous perturbation stochastic approximation, ALOPEX, and
simulated annealing. Simulated data and results are presented.
78. Linear long wave propagation over discontinuous submerged shallow water topography
Mark Brandao Virginia Polytechnic Institute and State University
Megan Golbek California State University Monterey Bay
Tucker Hartland California State University Chico
Ravi Shankar California State University Chico
Yan Sheng Emory University
Advisor(s): Sergei Fomin, California State University Chico
The dynamics of an isolated long wave passing over underwater obstacles are discussed in this paper within the framework of linear shallow water theory. Areas of practical application include coastal defense against tsunami inundation,
and harbor protection. Three sea-floor configurations are considered: an underwater shelf, a flat sea-floor with a single
obstacle, and a series of obstacles. A piecewise continuous coefficient is used to model the various sea-floor topographies. A simple and easily implementable numerical scheme using explicit finite difference methods is developed to
solve the discontinuous partial differential equations. The numerical solutions are verified with the exact analytical
solutions of linear wave propagation over an underwater shelf. The scope of this simplified approach is determined
by comparison of its results to another numerical solution and experimental data available in the literature for wave
transmission and reflection coefficients. The efficacy of approximating more complicated continuous underwater topographies by piecewise constant distributions is determined. As an application, a series of underwater obstacles are
modeled and reflection/transmission coefficients determined.
79. Modeling Microscopic and Macroscopic Traffic Flow Utilizing the Particle Filter and Ensemble
Kalman Filter
Melissa McGuirl College of the Holy Cross
Advisor(s): Bjorn Sandstede, Brown University
Mathematicians and engineers have been studying traffic flow for decades. In this project we combined new methods
of researching traffic flow in hope of creating a more effective way to model traffic and predict when traffic jams will
occur. We have studied both macroscopic and microscopic traffic flow. In the microscopic traffic flow model we track
the velocity, headway, and position of individuals cars, whereas in the macroscopic traffic flow model we study the
density and flux of cars over a segment of the road. We use the Lighthill-Whitham equations to model traffic flow,
but we have expanded this existing model to include various road conditions for more realistic settings. We also use
data assimilation for our microscopic and macroscopic traffic flow models. Implementing both the particle filter and
the ensemble Kalman filter has allowed us to compare data assimilation techniques in diverse road conditions. Our
research shows that data assimilation works in dynamic traffic settings for both the microscopic and macroscopic
traffic models. To test the efficacy of our methods we applied our models and data assimilation techniques to real data
from a traffic study in Minnesota.
80. Smoothing a Simulated Three-Dimensional Brain Image using an Edge Map
Emmanuel Estrada California State University Stanislaus
Tiffany Chu California State University Stanislaus
Advisor(s): Jung-Ha An, California State University Stanislaus
Three-dimensional images can be represented as a three-dimensional matrix of byte values, with each cell representing
the intensity of a single ‘pixel.’ During data transmission, values can become corrupted with noise. The goal of this
project is to correct for these errors, maximizing image fidelity while eliminating noise. To accomplish this, we apply
our image smoothing algorithm combined with an edge map to simulated brain images. Numerical results show the
effectiveness of the presented algorithm.
26
Titles, Authors, Advisors and Abstracts
81. Performance Sensitivity in Vertical Geothermal Energy Harvesting Systems
Sophia Novitzky Virginia Tech
Mahalia Sapp
Samuel Naden Macalester College
Keenan Hawekotte Nebraska Wesleyan University
Advisor(s): Burt Tilley, Suzanne Weekes, Worcester Polytechnic Institute
Residential geothermal energy systems have the potential to provide a cost-effective, low carbon footprint technology
for heating and cooling. The systems use the soil beneath a residence to store thermal energy in the summer and harness
energy in the winter. The soil exchanges heat with the coolant that flows through a pipe inserted into a vertical bore.
The energy transfer in the soil changes the coolant temperature, setting the efficiency of the residential heating/cooling
system. We are interested in finding how the cross-sectional arrangement of pipes in the bore affects the temperature
of the coolant as a function of depth. For a given cross-section, we describe the temperature functions of the coolant
and the surrounding soil through a system of time-dependent partial differential equations. From these equations, we
find a Sturm-Liouville eigenvalue problem in each cross-sectional variable, whose eigenvalue determines the decay
rate of the eigenfunction over depth. We find a numerical solution via the finite element method for the eigenpair
corresponding the lowest decay rate. Our results can be used to determine the borehole depth necessary to meet the
energy needs of a residence given a desired geometry of the system.
82. Quantum Audio Embedding Using M-Band Wavelets
Cameron Sakurai Western Connecticut State University
Alex Potocki Western Connecticut Statue University
Advisor(s): Xiaodi Wang, Western Connecticut State University
Information hiding is a modern method of privately transmitting information from sender to recipient. In recent there
have been developments in the field of quantum computing. It is therefore important to develop means of encoding
information that will resist attacks from both standard and quantum computers. We will utilize wavelet transforms and
pseudo-quantum encoding to transmit audio data hidden inside of images. Wavelet transforms offer energy conservation and the separation of energy levels, allowing our audio to be embedded with the portion containing the highest
energy. We are exploring the use of M-band wavelets, however for demonstration purposes, we will utilize 4-band
wavelets.
83. Symmetries of the Gross-Pitaevskii Equation
William Barker University of Central Arkansas
Advisor(s): Danny Arrigo, University of Central Arkansas
We consider the classical and nonclassical symmetries of the Gross-Pitaevskii equation (GPE). We will show that the
nonclassical symmetries are more general than the classical symmetries. We further consider first order compatibility
and the GPE and show a class of compatible equations exists that are not obtainable by the symmetry method.
84. Fractal and Wavelet Image Compression
Brian O'Connor Western Connecticut State University
Advisor(s): Xiaodi Wang, Western Connecticut State Univerity
In the application of mathematics, image compression has seen advancements both from fractals and wavelets. Fractal
methods use self-similarity accross different scales to reduce stored information. Segmentation probably plays the
most significant part in fractal image compression. The success of acheiving a large compression ratio effectively rests
on an efficient segmentation algorithm. Wavelet methods exploit redundancies in scale to reduce information stored
in the wavelet transform domain. Hybrid methods apply fractal techniques to information in the wavelet transform
domain to provide even greater compression performance. In our research, we will investigate compression of fractal
images using the Daubechies 4 wavelet.
27
Titles, Authors, Advisors and Abstracts
85. On Globally Defined Solutions of the Generalized CLM Equation
Gopal Yalla College of the Holy Cross
Samantha Davies Carnegie Mellon University
Advisor(s): Johnny Guzman, Brown University
We consider a generalized one dimensional model of the three dimensional vorticity equation, proposed by Constantin,
Lax, and Majda. Results for the one dimensional model may provide insight to the behavior of solutions of Euler’s
equation. Specifically, we studied the effect the parameter that controls the amplification of the convection term has
on solutions to this equation. Cordoba and Okamoto showed that nonpositive parameter values lead to singularities
in finite time. While the behavior of solutions for positive values of the parameter are unknown, it is conjectured by
Okamoto et al. that the solutions undergo a bifurcation from blow up to global existence for some parameter value
between 0 and 1 (around 0:6). Several different numerical and analytic methods are utilized to analyze whether or not
solutions form singularities in finite time, and results reveal evidence contrasting the conjectured value.
86. An Agent-Based Approach to Modeling Police Patrol Strategies and Their Effects on Crime
Alejandro Camacho California State University, Fullerton
Hye Rin Lin Lee California State University, Fullerton
Advisor(s): Laura Smith, California State University, Fullerton
In many urban communities crime is an unfortunate reality for its inhabitants. High levels of crime require law enforcement agencies to optimize their resources to address criminal behaviors. The goal of this project is to extend
an agent-based model for crime pattern formation to incorporate police. This approach allows us to simulate criminal and police behaviors. Through these simulations, we can provide a means to test hypothetical policing strategies
without costly or unethical experiments. The theories we test incorporate realistic patrolling scenarios and the effect
it has on both criminal behavior and the total crime in a region. We compare our results with existing agent-based
approaches, such as random patrols and “hot spot” policing. Using more realistic models, we can test theories to help
law enforcement mitigate crime.
87. Dealing with Damage: A New Approach to Image Segmentation with Dynamic Artifact Detection
Nicholas Takaki Carnegie Mellon University
Jing An University of California, Los Angeles
Advisor(s): Dominique Zosso, UCLA
In this research, we introduce a new model for segmenting damaged images. A segmentation of an image I W  ! R
is a two-color piecewise constant approximation representing “object” and “background.” Traditionally, segmentation
involves evolving a level set function and colors c1 ; c2 to minimize the Chan-Vese functional
Z
Z
2
2
EC V .c1 ; c2 ; / D .I c1 / H./ C .I c2 / .1 H.// C jrH./j;


where H is the Heaviside function and is a weight term. This approach works well on undamaged images, but fails
when artifacts are introduced.Our research has two parts. First, we introduce a dynamic artifact class X W  ! f0; 1g
with weight , which prevents outliers from skewing the segmentation. Our new functional is
Z
Z
Z
2
2
Enew .c1 ; c2; ; X/ D .1 X/..I c1 / H./ C .I c2 / .1 H./// C jrH./j C X;



which has been very effective at segmenting both damaged and undamaged images. Second, we develop a minimization scheme based on diffusion and thresholding, which runs significantly faster than traditional gradient descent
techniques. We also include results and comparisons with existing methods.
88. Quantum-Inspired Robust Audio information embedding algorithms in M-band wavelet Domain
Zhang Mingyang Affiliated High School to Jilin University Changchun China
Yuting Liu
Xuanhe Liu
Advisor(s): Ren Yulian, Affiliated High School to Jilin University
The research is in the area of information embedding has recently gained substantial attention. This is mainly due
to the increased interest in digital watermarking technology that potentially can solve copyright infringements and
28
Titles, Authors, Advisors and Abstracts
data integrity disputes. As digital music has become increasingly popular, there is a great need to further develop a
method that could be used to enhance copyright protection in the music industry. This paper addresses this problem
by providing a way to protect against unauthorized copying of digital music by the use of M-band discrete wavelets
to transform the audio signal into the wavelet domain, then create corresponding “pseudo quantum signal,” and finally
insert watermark into such “pseudo quantum signal.”
89. Nonparametric Regression Algorithm Based on M-band Wavelet Transform
Matt Nejame Western Connecticut State University
Brian Bucknam Western Connecticut State University
Advisor(s): Xiaodi Wang, Western Connecticut State University
Accurately predicting the future changes in the stock market is highly sought after by many businesses and people.
Many corporations need and pay lots of money for accurate analysis of stock market behavior. In this research we
will be applying wavelet based-nonparametric regression models to predict changes in the stock market. Wavelets can
analyze stock market cycles over a short period of time in order to predict future behavior. A nonparametric regression
model can, by putting a line of best fit, also give you a representation of stock market behavior in order to predict
changes. This paper will use these mathematical tools, to predict stock market prices, and compare them to actual
stock market prices.
90. Spatial Knowledge and Relatedness
Rebecca Harper Willamette University
Advisor(s): Shilad Sen, Macalester College
Semantic relatedness (SR) is a burgeoning field in computer science that attempts to replicate the human process of
determining the relationship between two concepts. The current SR algorithms, however, do not attempt to incorporate the effect of differences in human knowledge on the perceived relatedness. Furthermore, the assessments of the
accuracy of SR algorithms depend on a universal, human-based “gold standard” data set collected from people with
unknown levels of knowledge. This study predicted a positive correlation between spatial or geographical knowledge
and evaluated SR of spatial concepts. After analyzing survey data from 1000 participants, this study found that the
greater the respondent’s spatial knowledge of two spatial concepts, the more related the respondent perceived the
concepts to be. The data set contains 34,422 SR judgments of spatial concepts that include respondents’ spatial knowledge and valence toward the concepts. This is an important find for the field of semantic relatedness in computer
science because it uncovers a need to reassess the “state-of-the-art” techniques used to emulate human perception of
semantic relatedness.
91. Determining Optimal Level of Unemployment to Maximize Growth of Real GDP
Angela Hanson Rose-Hulman Institute of Technology
Andrew Kipp Rose-Hulman Institute of Technology
Advisor(s): Wayne Tarrant, Rose-Hulman Institute of Technology
There is much debate over the effects of unemployment on the economy. Some people believe that 0% unemployment
is the best level because everyone should be able to find a job. The Federal Reserve believes that it should be about
5%, or the natural unemployment rate. However, a definitive percent of unemployment has not been mathematically
calculated to maximize the output of the economy. The purpose of this research is to determine if there is a level of
unemployment which maximizes percent growth in real gross domestic product (GDP). This will be determined by
creating an optimization model which considers the effects of employee compensation, consumer price index, labor
force participation rate, and unemployment on real GDP. Also, each contributor to change in real GDP listed above
must be given a value to proportionally weight its effects. This model will use real GDP growth as a reference for an
improved economy because it applies a measure to economic change even if the annual value is imprecise. The data
gathered will be from the various time spans across different countries including the United States. Ultimately, the
goal is to create a model which can be tailored for multiple countries around the world in order to determine a more
precise goal for the unemployment rate.
Titles, Authors, Advisors and Abstracts
29
92. Musical Exploratory Data Analysis
Rachael Fountain Westfield State University
Advisor(s): John Judge, Westfield State University
Founded by John Tukey, the most influential statistician of the second half of the twentieth century, exploratory data
analysis (EDA) is the process of creating new and original methods of analyzing data through visual graphics. The
power of EDA is that it allows the brain to approach and view data in an entirely different way; it lets the data speak
for itself. It is common knowledge that music is so much more than just the notes, pitches, and beats we hear and in
fact there are surprising connections between music and mathematics that can be uncovered through exploratory data
analysis. In this project we will discuss how we used EDA to create two visual graphics which we call will variability
graphs and contour graphs. These graphs allow for pitches, scales, and distances between notes to be examined quickly
and they also allow for the standardization of music in all genres, keys, and tempos. Each of these graphics will then be
used along with techniques from statistics and real analysis to investigate the relationship between musical melodies
and mathematics.
93. Cellular Automata and Shrinking Cities: A Mathematical Model of Youngstown, Ohio
Eric Shehadi Youngstown State University
Advisor(s): Alicia Prieto Langarica, Youngstown State University
Youngstown, Ohio, a city found in the affectionately labeled “rust belt,” is a shrinking city plagued by thousands of
vacant homes. Since the crash of its steel economy in the 1970s, Youngstown has lost over half of its population. This
has led to mass abandonment and blight, which severely inhibits the health of several of Youngstown’s neighborhoods.
The city has taken measures to battle blight by developing a demolition program for blighted properties and investing
resources in a community development corporation. This, however, has sparked a debate as to where the extremely
limited demolition and revitalization funds should be spent within a city. Individual based models, specifically cellular
automata models, are employed to model the dynamic systems that are found in a shrinking city. The goal of the model
is to offer insight as to what demolition and revitalization strategy is a better use of city funds — a suggested prioritized
approach or the current “scattershot” approach. The model also investigates the spatial spread and concentration of
blight. The model results compare the effectiveness of city programs, policies, and strategies.
94. Binomial Solutions to Smale's 17th Problem and their Application to Chemical Reaction Networks
Caleb Bugg Bugg Morehouse College
Advisor(s): J. Maurice Rojas, Texas A&M University
In 1998, Stephen Smale proposed a list of eighteen questions for the mathematical community. Smale’s 17th problem
is concerned with the development of a deterministic algorithm that can approximate a root of a random polynomial
system in polynomial-time. This project in its current form provides a positive answer to Smale’s 17th problem for
binomial systems, and explores a concrete application of the algorithm in chemistry. The main result of this project
is an algorithm that approximates a root of an entire binomial system (n variables, n equations) in polynomial-time.
The algorithm utilizes matrix exponentiation and the Smith Normal Form of an integer matrix in order convert the
system to a simpler system. Binomial systems are used in a variety of mathematical modeling situations. In particular,
there are sufficient, easily verifiable conditions for the expression of a chemical reaction network (CRN) as a binomial
system. We utilize our algorithm to solve these systems to determine the steady-state concentrations of the species in
chemical reaction networks.
95. Algorithmic and Theoretical Exploration of Tiling Harmonic Functions
Yilun Du PRIMES-USA
Advisor(s): Sergiy Merenkov, City College
Consider a square tiling of a region D in the complex plane as a square tiling (finite or infinite) with edges parallel to
the coordinate axes and mutually disjoint interiors whose union is all of D. For each tile, we define the oscillation upon
the tile as the maximum value of vertex on the tile minus the minimum value squared(vertexes may not necessarily
be the actual vertices of the squares). A function u on the grid is defined as tiling harmonic if interior points are
assigned such that oscillation over the entire grid is minimized. We present a computationally efficient algorithm that
determines the values of the interior points from the value of exterior points in square tilings. In addition, we generate
an algorithm for computing graph harmonic functions(values as averages of neighboring values) defined on boundary
30
Titles, Authors, Advisors and Abstracts
values. We then use 3D images comparisons of the values of graph and tiling harmonic function and explore the general
behavior of tiling harmonic functions given boundary values. We find in general that graph harmonic functions are not
equivalent to tiling harmonic functions. In addition, we prove that the function f .x C iy/ D cy as tiling harmonic,
and find criteria for bounded tiling harmonic functions on infinite tilings.
96. Applying Dynamic Amino Acid Interaction Preferences to GNM with X-ray Crystallography and
Solution NMR Data
Robert Marek Houghton College
Advisor(s): Kristin Camenga, Advisor/Houghton College
Modeling the folding structure in proteins may improve drug manufacturers’ ability to design effective medical inhibitors for proteins involved in the treatment of diseases by allowing more precise and accurate prediction of the
folded protein structure. The Gaussian Network Model (GNM) is one method to model these folding patterns. Our
research modified the GNM model to take into account amino acid interaction preferences. Using X-ray crystallography and NMR data from the Protein Data Bank, we predicted B-factors, experimental values that correlate with amino
acids’ mean square deviation, of a set of proteins. The predicted B-factors from our modification of the GNM showed
a higher correlation to the experimental B-factors than did a simple form of the GNM. In the future, incorporating additional variables such as bond torsional forces and multiple conformations of the same protein would likely increase
the efficacy of the model.
97. A Model of Sedimentation and Consolidation in Colloidal Suspensions With Applications to
Industrial Paint Products
Eric Johnson University of Wisconsin Milwaukee
Advisor(s): Eric Johnson, University of Wisconsin Milwaukee
Pigments settling out of paint can destroy the product when settled pigments are consolidated into a hardened cake.
Our industrial collaborators have found that this process seems to be amplified during transport via cargo ship. We
have chosen to model the paint as a colloid instead of a discrete particle system. We assumed the mixture to be
continuous and the fluid to be Newtonian and satisfy the point Navier-Stokes equations. We then developed a twophase PDE model. It extends work done by Kynch and others. We added diffusion effects in the flux term during
hindered settling, and account for both effective stresses and pore pressure in consolidation. Taking these terms into
account gave us a degenerate parabolic PDE. We used a finite volume scheme to run several numerical experiments
and have partially verified previously obtained results. However, we were not able to obtain results for Peclet numbers
of order 105 , a realistic estimate for our mixture. We suspect we will need to implement fraction step methods into
our finite volume scheme. Future work will include allowing gravity to vary to capture the effects of waves, finding
physically principled parameter ranges and quantifying the added effects of high frequency vibrations from the ship’s
engines on consolidation.
98. Probing the Early Universe with Deuterium and Helium-4 utilizing Second Order Runge-Kutta
Method
Jeremy Ariche Morehouse College
Advisor(s): George Fuller, University of California - San Diego
Sterile Neutrinos are a primary candidate for dark matter. Their origin is unknown, yet it is plausible that a lepton
asymmetry could generate sterile neutrinos. In addition, observations of the Cosmic Microwave Background suggest
the presence of extra radiation energy density. We aim to investigate how lepton asymmetry and extra radiation energy
density affect the abundances of Helium-4 and Deuterium as determined in Big Bang Nucleosynthesis (BBN). We
will utilize the Wagoner-Kawano code which employs a second order Runge-Kutta differential equation solver to
time evolve the electron chemical potential, temperature, and baryon density as well as nuclear abundances. The first
three quantities are significant because they are the thermodynamic variables that describe the early universe, hence
setting the environment for the production of the light elements during BBN. We will compare the deviation of these
abundances due to lepton asymmetry and extra radiation against the standard model for primordial nucleosynthesis.
Titles, Authors, Advisors and Abstracts
31
99. An Aggregate Stochastic Model incorporating individual dynamics for Predation Movements of
Anelosimus Studiosus
Alex John Quijano East Tennessee State University
Advisor(s): Michele L. Joyner, East Tennessee State University
In this poster, we discuss methods for developing a stochastic model, which incorporates behavior differences in the
predation movements of Anelosimus Studiosus (a subsocial spider). Stochastic models for animal movement and, in
particular, spider predation movement have been developed previously; however, this paper focuses on the development and implementation of the necessary mathematical and statistical methods required to expand such a model in
order to capture a variety of distinct behaviors. A least squares optimization algorithm is used for parameter estimation
to fit a single stochastic model to an individual spider during predation resulting in unique parameter values for each
spider. Similarities and variations between parameter values across the spiders are analyzed and used to estimate probability distributions for the variable parameter values. An aggregate stochastic model is then created which incorporates
the individual dynamics. The comparison between the optimal individual models to the aggregate model indicates the
methodology and algorithm developed in this paper are appropriate for simulating a range of individualistic behaviors.
100. Barcode Decoding Using M-band Wavelet Transform
Jordan Marquis Western Connecticut State University
Josh Torres Western Connecticut State University
Advisor(s): Xiaodi Wang, Western Connecticut State University
A barcode is a pattern of parallel black and white lines with varying widths and distances. It is printed on products and
readable by a barcode scanner or machine in order to identify them. The varying widths and distances distinguish the
product so that it is a unique identifier. With new technology, there has been a demand for barcodes to be interpreted
by mechanisms other than designated barcode scanners. In our research we will use wavelet transforms to show that
a two-dimensional image can be taken of a one-dimensional barcode and transformed back into the unique identifier.
This can be applied to taking a picture of a barcode with a cellphone camera, transforming it into a 3-level wavelet
transform, and finally performing horizontal sampling in order to allow the information to be read for decoding.
101. puzzleJAR: Automated Constraint-based Generation of Puzzles of Varying Complexity
Justin Kaashoek Massachusetts Institute of Technology
Amy Chou
Advisor(s): Rishabh Singh, Microsoft Research
Engaging students in practicing a wide range of problems facilitates their learning. However, generating fresh problems
that have specific characteristics, such as using a certain set of concepts or being of a given difficulty level, is a
tedious task for a teacher. In this paper, we present puzzleJAR, a system that is based on an iterative constraint-based
technique for automatically generating problems. The puzzleJAR system takes as parameters the problem definition,
the complexity function, and domain-specific semantics-preserving transformations. We present an instantiation of our
technique with automated generation of Sudoku and Fillomino puzzles, and we are currently extending our technique
to generate Python programming problems. Since defining complexities of Sudoku and Fillomino puzzles is still an
open research question, we developed our own mechanism to define complexity, using machine learning to generate a
function for difficulty from puzzles with already known difficulties. Using this technique, puzzleJAR generated over
200,000 Sudoku puzzles of different sizes (9 9, 16 16, 25 25) and over 10,000 Fillomino puzzles of sizes ranging
from 2 2 to 16 16.
102. Destabilizing a Convergent Molecular Dynamics Simulation of the Hen Egg White Lysozyme
Protein
William Frazier East Tennessee State University
Advisor(s): Jeff Knisley, East Tennessee State University
Molecular Dynamics (MD) is a computational tool used to simulate the folding dynamics of a protein. Typically,
one begins with Protein Data Bank coordinates file for the crystallized protein, and after a sequence of preprocessing
steps, subsequent coordinates for each atom are predicted for femtosecond time steps for up to nanoseconds. If an
MD simulation tails, perhaps because the crystal structure is unstable or flawed, then the failure tends to occur in the
32
Titles, Authors, Advisors and Abstracts
preprocessing steps. In general, it is difficult to use MD to predict the destabilization (i.e., “unfolding”) of a flawed
crystal structure. In the project, we develop a working MD simulation of Lysozyme, after which we explore how the
crystal structure can be modified so as to cause the failure of the protein without causing premature termination at the
MD run. The goal is to suggest how MD simulations can be used to simulate the destabilization of a protein, which
would be useful information in applications of MD to drug design and similar areas.
103. Analysis of the Discrete Maximum Principle
Arman Green Morehouse College
Cassidy Krause University of Wisconsin-Platteville
Advisor(s): Johnny Guzman, Brown University
The maximum principle is a fundamental property of solutions to Laplace’s equation with Dirichlet boundary conditions. To numerically solve the equation we use the Finite Element Method (FEM). However, for certain triangulations
of the domain, the numerical solutions will not satisfy the maximum principle. The triangulation of the mesh affects
the values of the global stiffness matrix, which is directly used in calculating the numerical solution to the equation. If
all the entries of the inverse of the global stiffness matrix are positive, then the maximum principle will hold, (Ciarlet,
1970). An interest that naturally arises is identifying conditions of the triangulation of the mesh to ensure a positive
inverse of the global stiffness matrix. Recently Korotov et al. showed that an obtuse triangulation produces a global
stiffness matrix whose inverse has negative entries. We investigate the location and magnitude of the negative entries
of the inverse of the global stiffness matrix for arbitrary obtuse triangulations.
104. Topic Point Process Models for Twitter Data
Eric Lai University of California, Irvine
Advisor(s): Blake Hunter, Alexandre Robicquet, Eric Fox, Claremont McKenna College
The widespread use of social media as a form of communication tool has enabled automated detection of trends and
extraction of useful information. Social media is attractive for such purposes because its data clusters geographically
around events such as sports competitions. However, transforming raw, free-form text into meaningful information
remains a challenging task. Confounding factors include the scale of posted data, recovery of event information, and
the temporal behavior of messages. In the present work, we consider 500,000 timestamped and geolocalized Twitter
messages from the East Los Angeles area collected over a ten-month period. After cleaning the Tweets by removing
stop words and correcting misspellings, we then apply topic modeling methods such as non-negative matrix factorization to discover prevalent topics as well as frequent words associated with each topic. We can then model Tweets with
topics over time as Hawkes and Poisson processes and compare their fit by computing its Akaike Information Criterion
value. This allows identification of self-exciting topics and Tweets. We plan to adopt this technique to an algorithm
that can automatically summarize large collections of Tweets, recover lateen relationships, and predict future events
from Tweets.
105. The Schelling Model: Simulation and Analysis
Michael Roberts University of Maryland
Advisor(s): Pieter Tibboel, HKUST
We studied the Schelling model, which is a well-known model of segregation. Our primary focus was analyzing its
one-dimensional variants using simulation and rigorous analysis. The variants we chose to study include those we’ve
called the standard model and the simple model. The standard and simple models consider agents split among discrete
types, with agents’ happiness determined by neighbors. In the standard model, a pair of unhappy agents of different
type will swap as long as both will be no worse-off. In the simple model an agent will change its type if it will
become happy after the swap. We have focused on results about “tipping points”, values of agents’ tolerance levels
(a parameter of these models), around which model behavior changes qualitatively. For the standard model, we have
extended results about the “tipping point” , the threshold of the agents’ tolerance level at which the segregation level
qualitatively changes in the one-dimensional Schelling model, from two types of agents to m types. For the simple
model, we similarly extended results from a model with two agent types to one with m types, focusing on tipping
points. We furthermore combined results about two tipping points to analyze a type of scenario which can occur only
for m > 2.
Titles, Authors, Advisors and Abstracts
33
106. Population dynamics of the Tasmanian devil and modeling strategies for the containment of Devil
Facial Tumor Disease (DFTD).
Ashley Sichak University of St. Francis
Advisor(s): Megan Powell, University of St. Francis
DFTD is causing a massive decline in the Tasmanian devil population in a 20-year time-span. In this poster, I will
present a system of ordinary differential equations used to understand the impact of disease on the population. I will
show how the model predicts future population numbers based on disease factors. The success of potential strategies
to help fight the population decline such as selective culling, isolation, and vaccination are evaluated using the model.
107. A Mathematical Model for the Interaction of the Proteins MMP-1, TIMP-1, and ECM in a Wound
Hannah Pennington Western Kentucky University
Nitin Krishna University of Chicago
Advisor(s): Richard Schugart, Western Kentucky University
This study aims to formulate, quantify, and analyze a mathematical model describing interactions among matrix metalloproteinases (MMP-1), their inhibitors (TIMP-1), and extracellular matrix (ECM) in the healing of a diabetic foot
ulcer. De-identified patient data for modeling were taken from Muller et al. (2008), a research outcome that collected
average physiological data for two patient subgroups: “good healers” and “poor healers,” where classification was
based on rate of ulcer healing. Model parameters for the two patient subgroups were estimated by minimizing the
least-squares residual between the data and model output using Matlab’s GlobalSearch and fmincon routines. The
model and parameter values were analyzed by conducting steady-state and sensitivity analyses. Both global and local
sensitivity analyses were conducted and compared. The global sensitivity analysis was performed using Latin hypercube sampling, while local analysis was conducted through a classical sensitivity analysis followed by an SVD-QR
subset selection. Model and clinical implications are discussed based on our results.
108. Incorporating Exposure Periods, Hospitalization, and Mosquito Control into Dengue Fever Models
Jared Bland The Richard Stockton College of New Jersey
Advisor(s): Brandy Rapatski, The Richard Stockton College of New Jersey
Dengue Fever is an epidemic in many third world nations and warrants the attention of mathematical modeling to
understand its dynamics and propose solutions to combat the deadly disease. Our model incorporates hospitalization
effects, which greatly reduce the deadliness of the disease, exposure periods, which more realistically model the
phenomenon, and active participation by the community to reduce the ability for mosquitoes to breed. Here, I present
the results of a system incorporating these effects. The exposure period and hospitalization result in a longer lasting
epidemic, but with a less pronounced peak. Our model also aimed to confirm a study from Sri Lanka, which used
community involvement to reduce the mosquito breeding grounds, thereby reducing the number of Dengue cases. Our
model compared different results corresponding to the timing of community involvement, and, as expected, the sooner
the community is involved, the less Dengue cases; this confirms the intuition that year-round maintenance is required
to help prevent epidemics.
109. Glial Cell Defense Mechanisms in Response to Ischemic Hypoxia in the Brain
Tiffany Reyes Whittier College
Hasan Sumdani The University of Texas at Arlington
Matthew Buhr University of South Dakota
Advisor(s): Benjamin Morin, Anuj Mubayi, Arizona State University
Three models are introduced that explore the dual role of glial cells in the formation of scar tissue and in the neural
repair following hypoxia ischemia in the brain. Scar tissue helps protect the brain during the acute phase of injury
by limiting the spread of secondary damage, but limits recovery by inhibiting the repair of damaged neurons. Scar
formation is not ideal, however repaired neurons are still susceptible to damage and do nothing to halt the spread of
ischemic injury. A stochastic, spatially explicit Cellular Automaton (CA) model is used to capture the dynamics of
neural tissue repair and the containment of damage by scar tissue. In addition, two deterministic models are developed
to approximate the stochastic process namely, Mean Field (MFA) and Pair Approximation (OPA) models. We show
that the MFA neglects all spatial dependence among state variables. The OPA models the dynamics of state variables
34
Titles, Authors, Advisors and Abstracts
evolving as pairs, where spatial adjacency matters. Our results compare how the trade-off between scar tissue formation
and neural repair impacts future brain health.
110. The Effect of Prey Dispersal on a Two-Patch Predator-Prey System
Emma Talis Marist College
Conor Wells Drake University
Victoria Kelley James Madison University
Brennan Bean Brigham Young University - Idaho
Juliana Richardson Smith College
Advisor(s): Glenn Ledder, University of Nebraska - Lincoln
We consider the effect of prey dispersal in a two-patch predator-prey model in which the two patches are qualitatively
different. In particular, we assume patch two has a significantly smaller carrying capacity and a correspondingly
higher predation rate. Scaling the model under these assumptions introduces a parameter of arbitrarily small order,
allowing for an asymptotic analysis. We show that the predator and prey will always coexist for biologically reasonable
parameter values. Furthermore, we prove the existence and uniqueness of a coexistence equilibrium and determine the
stability regions in the parameter space. Using numerical simulations, we illustrate the varying effects of prey dispersal
on the stability of the coexistence equilibrium and find parameter values for which a Hopf bifurcation occurs.
111. An Individual-Based Model of Chaparral Vegetation Response to Frequent Wildfires
Dayna Mann Pepperdine University
Reanna Dona Pepperdine University
Advisor(s): Timothy Lucas, Pepperdine University
In recent years, the Santa Monica Mountains (SMM) have been plagued by frequent wildfires which threaten the
native chaparral species. Nonsprouting chaparral species are completely killed by a fire, but their seeds germinate in
response to fire cues. Facultative sprouters both resprout after a wildfire and release seeds that germinate post-fire. This
project is based on data collected since 1986 at a biological preserve adjacent to the Malibu campus of Pepperdine
University with an average fire return interval of 7.25 years. We present a spatial model that simulates the growth, seed
dispersal and resprouting behavior of individual shrubs that compete for space and resources in a domain similar to
our study site. The model incorporates varying rainfall and fire frequency as well as the competition between plants for
scarce resources. Our simulation reproduces the change in plant community structure at our study site such as the local
extinction of Ceanothus megacarpus due to short fire return intervals. Our simulations also predict frequent wildfires
will drastically reduce the number and size of individual shrubs. A mathematical model of this system is important
because a reduction of vegetation cover can cause a decrease in slope stability that leads to rock and mudslides.
112. Combinatorics of k-Interval Cospeciation for Cophylogeny
Jane Coons State University of New York at Geneseo
Advisor(s): Joseph Rusinko, Winthrop University
The subfield of phylogenetics known as cophylogenetics is concerned with discerning evolutionary relationships between taxanomical units that are evolving concomitantly, such as hosts and parasites and genes within a species. We
show that the cophylogenetic distance, k-interval cospeciation, is distinct from other metrics and accounts for global
congruence between locally incongruent trees. We define the size of the neighborhood of trees which satisfy the largest
possible k-interval cospeciation with a given tree. The growth of this neighborhood indicates that k-interval cospeciation is useful for analyzing simulated data.
113. A Model for Stripe Rust Growth with Two Fungicidal Effects
Josean Velazquez University of Puerto Rico at Humacao
Advisor(s): Leon Arriola, Derdei Bichara, University of Wisconsin, Arizana State University
Stripe rust, also known as yellow rust, is a disease caused by the fungus Puccinia striiformis that affects host crops,
primarily barley and wheat. It is the most prevalent strain of wheat rust within North America, and can account for a
significant amount of yield loss within a harvest. In order to estimate the dynamic yield loss of a wheat crop infected
by stripe rust undergoing fungicide application, a single-host, single-pathogen compartmental model is introduced
Titles, Authors, Advisors and Abstracts
35
using a system of nonlinear ordinary differential equations. Two fungicidal effects are considered: preventative and
anti-sporulant, which reduce susceptibility and inhibit the further spread of an infection, respectively. The stability
conditions associated with the disease-free and endemic equilibria are identified. Numerical simulations are then used
to examine the time-dependent behavior of fungicide efficacy and approximate yield loss. Finally, sensitivity indices
are calculated to study the impact of fungicide efficacy and retention relative to plant growth.
114. Modeling the Interaction Between Inflammatory and Cardiovascular Dynamics
Jessica Perez The College of New Jersey
Ailaura Donahoe Pennsylvania State University
Chelsea Ball California State University, Monterey Bay
Advisor(s): Mette Olufsen, North Carolina State University
The more quickly a patient begins to mobilize after surgery, the shorter their recovery time is. However, postoperative
patients are prone to syncope, or fainting. This is due to inflammation lowering blood pressure. If a patient’s heart is
unable to compensate for this change in pressure, they can faint. To better understand this phenomenon, three models
were developed to predict blood pressure, heart rate, and the inflammatory response. The models were formulated
as systems of differential equations that were solved using MATLAB. They were then tested using blood pressure
and heart rate data measured from a healthy subject who was given a dose of endotoxin to stimulate an inflammatory
response. The ultimate goal of the project is to combine the models for heart rate and blood pressure in the cardiovascular system and the model of the inflammatory response, in order to achieve a better grasp of why patients faint after
surgery and to shorten recovery time by expediting mobility.
115. Alcoholism-A Mathematical Model with Media Awareness Programs
Erik Ander Rollins College
Advisor(s): Zeynep Teymuroglu, Rollins College
In this paper, we address the role of media awareness on the spread and persistence of a drinking behavior in a
community. Our model integrates a media awareness component (Huo et al., 2014) into a modified version of an SIR
drinking model developed by Sanchez et al. (2007). Here, the number of problem drinkers that exist in the community
depends on the effectiveness of media awareness programs and peer pressure without the possibility of direct relapses.
Our model mimics the dynamics of media awareness programs such that the density of awareness increases as the
percentage of problem drinkers increase in the population. However, as mentioned in Huo et al. (2014), maintaining
continuous support for media awareness is not possible. The peer influence is modeled as a factor to encourage drinking
among the susceptible population as well as to cause relapses among recovered individuals. A threshold value, R0 ,
is proposed for the drinking-free equilibrium. Numerical simulations are presented to study the stability conditions in
the case that alcoholism culture is established in the community.
116. The Effects of Regional Vaccination Heterogeneity on Measles Outbreaks with France as a Case
Study
Nicholas Roberts Arizona State University
Elaine Alexander
Savanah McMahon
Advisor(s): Sergei Suslov, Kamuela Yong, D. Burkow, M. Manning, Arizona State University
The rubeola virus, commonly known as measles, is one of the major causes of vaccine-preventable deaths among
children worldwide. This is the case despite the fact that an effective vaccine is widely available. Even in developed
countries elimination efforts have fallen short as seen by recent outbreaks in Europe, which had over 30,000 cases
reported in 2010. The string of measles outbreaks in France from 2008-2011 is of particular interest due to the documented disparity in regional vaccination coverage. The impact of heterogeneous vaccine coverage on disease transmission is a broad interest and the focus of this study. A Susceptible-Exposed-Infectious-Recovered (SEIR) multi-patch
epidemiological model capturing the regional differences in vaccination rates and mixing is introduced. The mathematical analysis of a two-patch system is carried out to help our understanding of the behavior of multi-patch systems.
Numerical simulations are generated to aid the study of the systems qualitative dynamics. Data from the recent French
outbreaks was used to generate parameter values and to help connect theory with application. Our findings show that
36
Titles, Authors, Advisors and Abstracts
heterogeneous vaccination coverage increases controlled reproduction number compared to comparable homogeneous
coverage.
117. Importance of Remineralization of Cladophora-Epiphyte Assemblages on Silicate Cycling in Lake
Michigan
Kolin Konjura University of Wisconsin-Milwaukee
Advisor(s): Gabriella Pinter, University of Wisconsin-Milwaukee
Nuisance growths of benthic filamentous algae, particularly Cladophora (Cl), support dense diatom epiphytes. These
algae can impact the drawdown of silica from the water column to the benthos. However, at the end of the growing
season the fate of the Si in the Cl is unknown. To further examine Si dynamics in the nearshore, Si pools in Cl-diatom
branches were measured during the degradation of Cl branches in triplicate 1 L tanks. Algal branches that were still
attached to the wharf (attached) were compared with algae found on the beach (sloughed), which had presumably
started degrading. Water samples were collected from tanks for determination of dissolved Si (dSi), particulate Si
(pSi), and algal subsamples were collected to determine biogenic Si (bSi) in the Cl-diatom branches. Over 58 days,
there was a large flux of Si from Cl and diatom pools into the dSi fraction; dSi within the attached increased, while
mean bSi decreased. In tanks with sloughed, dSi also increased while mean algal bSi decreased. Suggesting that
attached lost more Si into the dSi pool, possibly because the initial diatom populations had higher Si. A system of
ordinary differential equations was used to model the predictability of the system, and analyze further constraints that
needed to be considered.
118. DNA Sorting with Game Theory
Samuel Simon Carnegie Mellon University
Caroline Jansen Notre Dame
Elizabeth Tatum Rutgers University
Advisor(s): Marion Scheepers, Boise State University
Ciliates have two copies of their genomes, one an encrypted version, the “ciphertext” version, of the other. Ciliate
genome maintenance involves decryption of the ciphertext genome to replace their current “plaintext” genome. According to the current biological model this decryption is accomplished through two specialized permutation-sorting
operations. Not all permutations are sortable by these. For one of these sorting operations random applications to even
a sortable permutation can result in a failed sorting. This suggests that either ciliates have a strategy to sort a permutation, or else there is a rescue mechanism to recover from errant sorting attempts. Strategic ciliate permutation sorting
can be modeled by two-player games. Using newly developed graph theoretic tools we prove a strict bound for when
a player has a winning strategy in some of these games. Using graph-based techniques we also prove that the current
biological model provides a rescue mechanism for ciliate permutation sorting.
119. A Periodic Matrix Population Model for Monarch Butterflies
Emily Hunt James Madison University
Advisor(s): Anthony Tongen, James Madison University
The monarch butterfly (Danaus plexippus) exhibits a unique migration phenomenon leaving Michoacan, Mexico in
the spring and traveling as far north as Southern Canada; later generations return to the same location in Mexico
the following fall. However, there is concern within the scientific community about the long-term stability of this
impressive annual journey. We use periodic population matrices to model the life cycle of the eastern monarch butterfly
and find that this unique migration is not currently at risk. We extend the model to address the three primary obstacles
for the long term survival of this unique migration: deforestation in Mexico, increased extreme weather patterns, and
milkweed degradation.
120. Stochastic Delay and Signal Propagation in Gene Regulatory Networks
Sarah Stanley University of Houston
Advisor(s): William Ott, University of Houston
The overall goal of this project is to understand the flow of information through gene regulatory networks and to create
a model that effectively simulates stochastic signaling during protein transcription. To accomplish this, relevant models
Titles, Authors, Advisors and Abstracts
37
must be created to accurately represent the signal’s path and intensity. My project sought to improve current modeling
paradigms such as ODEs, which oversimplify signal complexities and exclude important dynamical features such as
delay, by ignoring the time involved during transcription. As an alternative, signaling was modeled using the M=G=1
queue, where M represents a memoryless arrival process (transcription initiation); G, the service time distribution
(protein formation); and infinity, the number of service channels. In this model, the queue, a virtual holding tank,
incubates proteins until maturation with succeeding queues continuing synthesis, producing a cascade. My project
expanded upon two models in which the effects of delay were observed in an unmodified system as well as a system
with limited initiation events. In turn, I analyzed the effects of a finite number of servers on propagation speed,
concluding that as the number of servers increases, signaling resembles the limited initiation system.
121. Modeling Chronic Immune Activation in HIV-1 Infection
Kenneth Goettler College of the Holy Cross
Advisor(s): David B. Damiano, College of the Holy Cross
Chronic immune activation has been implicated as a major cause of the progression to AIDS in HIV-1 infection and
in the limited immune reconstitution of HIV-1 patients on antiretroviral therapy. The healthy human immune system
maintains low levels of rapidly proliferating and short-lived activated CD4+ (helper) and CD8+ (killer) T lymphocytes.
In HIV-1 infection, however, the presence of virus increases recruitment into this activated population and effectively
burns out’ the immune system over time. Here we employ a system of twelve non-autonomous ordinary differential
equations, modeling CD4+ and CD8+ T lymphocyte populations, to explore the effects of chronic immune activation
and how they shape the course of HIV-1 infection. Activated cells are divided into HIV-specific and non-specific
compartments. Heaviside functions are used to turn on increased recruitment of non-specific cells during infection.
Parameter and target values for each compartment are estimated from literature and optimized using a simplex method
and Latin hypercube sampling.
122. Population Dynamics of the Green Iguana in Puerto Rico: a Pest Control Method.
Dayanara Lebron Universidad Metropolitana
Advisor(s): Luis F. De La Torre, Universidad Metropolitana
The specie Iguana iguana, commonly named as green iguana, is a reptile specie native from Central and South America,
and introduced in Puerto Rico as a domestic exotic mascot in 1970. The density and abundance of the I. iguana
in Puerto Rico surpass the populational ranges of their native habitat due to the lack of natural predators; thereby
being identified as a pest. The consequences of their extensive population growth in the island has had a negative
ecological impact; primordially, causing a depletion of mangroves. I. iguana population is believed to currently display
an exponential growth, however there are no preview mathematical models to proof this observation, since a gap exist
in studies that explain the population dynamics of the I. iguana in Puerto Rico. Therefore, this research aims to explain
the population growth behavior of the green iguana through the development of mathematical models for the dynamic
population of green iguana. These models produce the estimations of optimal parameters that could yield population
equilibrium; these equilibrium must be achieved in real life by the conversion of these parameters in alternatives
policies as hunting season or animal exportation, ultimately modifying the future growth of this population in the
island.
123. Optimizing Anti coagulation Therapy
John Williams University of Wisconsin-Milwaukee
Advisor(s): Peter Tonelatto, University of Wisconsin-Milwaukee
Nearly 31 million people per year are prescribed anticoagulation therapy to prevent thromboembolic complications.
Despite this, anticoagulation therapy lacks a standard method of treatment. This is due in part to the sensitivity of
anticoagulation therapy and to the impact of patient characteristics such as age, genotype, and race on therapeutic dosing. This results in an excess of distinct dosing algorithms, all of which require regular monitoring and individualized
dosing. In fact, more than 12 different dosing algorithms have been proposed and used. The aim of this project is to
simulate clinical trials in order to test and identify an optimal dosing algorithm according to standard anti-clotting
outcome metrics. To do this, a Bayesian network model was used to create a statistical representation of a clinical trial
population, called clinical avatars. The model proposed by Hamberg was then used to Pk/Pd simulate treatment of the
38
Titles, Authors, Advisors and Abstracts
clinical avatars. As a secondary objective we will use a Markov model to predict the ability of dosing algorithms to
increase the cost effectiveness of anticoagulation therapy.
124. Physiologically Based Pharmacokinetic Modeling for Acetone: How Much Do We Really Breathe
In?
Vy Nguyen Baylor University
Ashley Edwards North Carolina State University
Benjamin Freedman North Carolina State University
Advisor(s): Marina Evans, EPA
We are interested in developing a PBPK model of inhaled acetone. The interest in acetone derives from two sources:
its hydrophilic behavior and its ketotic structure. Since acetone is hydrophilic, it can be absorbed into the tissue of the
upper respiratory tract after it is inhaled. There are other more toxic chemicals that are similarly structured to acetone,
so if our PBPK model for inhaled acetone accurately represents the data, then the results can be applied to study those
other more toxic chemicals. The model was developed with the respiratory tract separated into three different compartments along with a respiratory metabolizing constant. When the parameters were globally optimized, the model
accurately estimated concentrations for the chamber with four varying initial conditions. Respiratory metabolism was
found to be driving the elimination of acetone in the body, while the liver metabolized the chemical much more slowly.
With the development of a model that is mathematically and physiologically accurate, the model can now be adapted
to perform toxicological analysis on chemicals that are similarly structured to acetone, yet more toxic.
125. Modeling Latency of Thalmocortical Fast-Spiking Interneurons in Schizophrenia
Jennifer Houser East Tennessee State University
Advisor(s): Talitha Washington, Howard University
Neural thalamocortical circuits relay external sensations from the thalamus to the cortex where sensory information
is then processed. Feedforward inhibition involving a subtype of fast-spiking interneurons, which are marked by the
calcium-binding protein parvalbumin, reduce the chance that a postsynaptic neuron will fire an action potential. Consequences on the circuit due to the absence of parvalbumin expression in fast-spiking neurons in schizophrenia patients
are caused by fast spiking latency. In this presentation, we present a conventional neuron model. We will show how to
develop a mathematical model to incorporate a latency effect as well as show numerical simulations.
126. Estimating Bacterial Diversity in Defined Regions
Krista Kernodle University of California, Irvine
Advisor(s): Katherine Pollard, University of California, San Francisco
Scientists have made many predictions regarding the number of species on the planet today using varying methods,
all of which produce a wide range of possibilities. In particular, the number of bacterial species (also known as OTUs)
has been difficult to estimate and attempts are generally regarded as producing very low estimates. To address this
problem, we present a method that uses information on the areas of bacterial ranges to estimate the number of bacteria
in a defined region. The robustness of the method was tested using simulations before being applied to a dataset
of OTUs observed in soil samples collected from Central Park. Further applications to real-world data may provide
improved estimates of total bacterial diversity
127. A Bio-mathematical Approach to the Stability of HIV and AIDS
Talon Johnson Morehouse College
Advisor(s): Shelby Wilson, Morehouse College
HIV is a sexually transmitted disease that weakens one’s immune system allowing other pathogens to affect one’s
body, ultimately resulting in the development of AIDS. A nonlinear mathematical model of differential equations with
piecewise constants will show us the rate that HIV spread in a homogenous population. The solutions will be analyzed
the systems of differential equations.The population of the model will be separated into three subclasses: the HIV
negative class, the HIV positive class that don’t know they are being affected, and HIV positive classes that know they
are infected.
Titles, Authors, Advisors and Abstracts
39
128. A meta-analysis of coastal populations' genetic diversity of species throughout their range
Fangyuan Hong Mount Holyoke College
Advisor(s): Sean Hoban, Julia Earl, Suzanne Lenhart, NIMBioS, University of Tennessee-Knoxville
High genetic diversity in the centers of many species’ ranges have been thought to account for the abundant-center
hypothesis (ACH), which states that populations found in the center of their natural distribution are more abundant
than those found at the edges. Despite the prevalence of the ACH in population studies, relatively few studies have
tested the assumption of differential genetic diversity throughout ranges by accounting for variation in sample sizes.
To better assess the validity of the ACH in coastal habitats, this study is the first attempt at using a meta-analysis
to investigate how genetic diversity of populations living along world’s coastlines is affected by location within the
species’ range. An effect size of correlation coefficient (Pearson’s r) for each genetic measurement of each species
versus relative distance to center was calculated from the meta-analysis. Results showed positive correlations for most
taxa and insignificant correlations for others. Multivariable models between taxa, range dimension, relative direction
of sample sites showed significant r values, or associations. This study suggests for developing a more effective and
feasible marine conservation practice that targets individual populations rather than the whole species.
129. Two Novel Non-Convex Approaches for Learning Near-Isometric Linear Embeddings
Hao-Jun Shi University of California, Los Angeles
Jerry Luo University of Arizona
Kayla Shapiro University of California, Berkeley
Qi Yang University of Southern California
Kan Zhu University of California, Los Angeles
Advisor(s): Ming Yan, Wotao Yin, University of California, Los Angeles
The large size of data acquired and processed by diverse modalities poses a challenge to current information processing systems. Thus, we propose two non-convex approaches for learning near-isometric linear embeddings of finite sets
of data points. Given a set of training points X, we consider the secant set S.X/ that consists of all pairwise difference
vectors of X, normalized to lie on the unit sphere. The problem can be formulated as finding a symmetric and positive
semi-definite matrix that preserves the norms of all the vectors in S.X/ up to a distortion parameter ı. Motivated by
non-negative matrix factorization, we reformulate it into a Frobenius norm minimization problem, which is solved by
Alternating Direction Method of Multipliers (ADMM) and develop an algorithm, FroMax. Our other method seeks a
projection matrix such that the restricted isometry property (RIP) constraint is minimized directly. Using ADMM
and a Moreau decomposition on a proximal mapping, we develop an algorithm, NILE-Pro, for dimensionality reduction. Both methods are then demonstrated to be more computationally efficient than previous convex approaches for a
number of applications in machine learning and signal processing.
130. Properties of Self-Organized Criticality for Optimal Random Search
Kathleen Donahue Harvard University
Michelle Randolph University of North Carolina - Ashville
Advisor(s): Aliki Mavromoustaki, Heiko Hoffman, David Payton, UCLA, HRL
HRL Laboratories recently discovered that self-organized critical (SOC) systems produce search patterns that are well
suited for approximating solutions to certain hard optimization problems. Self-organized criticality is a phenomenon
found in dynamical systems ranging from rice piles to solar flares. Our team investigated which properties of SOC
patterns led to improved random search. Improving the efficiency of random search could have widespread application
(supply chains, allocation of resources, transportation networks). We investigated the following properties of search
patterns: fractal dimension of the boundary, size-to-boundary ratio, variation in shape, size distribution and boundary
structure. We found that smoothed shapes based on SOC-produced shapes performed better than original SOC patterns,
as well as tree structures and standard shapes such as squares, rectangles and disks.
131. Raising the Bound on the Circulant Hadamard Matrix Conjecture
Brooke Logan Rowan University
Advisor(s): Michael Mossinghoff, Davidson College
A circulant Hadamard matrix is an n n matrix with a number of properties: its rows are mutually orthogonal, its
entries are all ˙1, and each row after the first is a circular shift of the prior row. It is widely conjectured that no
40
Titles, Authors, Advisors and Abstracts
circulant Hadamard matrices exist with order n > 4. Many algebraic restrictions are known on the order of such a
matrix, and as a result there are less than 1400 values less than 4 1026 which have not been eliminated as a possible
order of such a matrix. We describe a project from the [email protected] REU program at Brown University in 2014,
in which we raised this bound by a factor of 2500 through some careful computations. In addition to describing
our computational methods, which involved searching for elusive double Wieferich prime pairs, and our new list of
possible orders, we also discuss the connection between this problem and the Barker Conjecture, and we pose some
directions for further research.
132. K -Knuth Equivalence for Increasing Tableaux
Michelle Mastrianni Carleton College
Colleen Robichaux Louisiana State University
Christian Gaetz University of Minnesota, Twin Cities
David Schwein Brown University
Advisor(s): Rebecca Patrias, University of Minnesota, Twin Cities
A K-theoretic analogue of RSK insertion and Knuth equivalence relations was first introduced in 2006. The resulting
K-Knuth equivalence relations on words and increasing tableaux on Œn has prompted investigation into the equivalence classes of tableaux arising from these relations. Of particular interest are the tableaux that are unique in their
class, which we refer to as unique rectification targets (URTs). Here we give several new families of URTs and a bound
on the length of intermediate words connecting two K-Knuth equivalent words. In addition, we describe an algorithm
to determine if two words are K-Knuth equivalent and to compute all K-Knuth equivalence classes of tableaux on Œn.
133. Touch-based Continuous Authentication
Wendy Rummerfield University of Redlands
Advisor(s): Kiran Balagani, New York Institute of Technology
Smartphone continuous authentication seeks to add a layer of defense beyond existing entry-point authentication
systems. Most pre-existing research focuses on data gathered from the touch screen itself, such as the characteristics of
tapping and swiping. However, the focus of this project discovering novel features from different sources, specifically,
the device’s motion sensors. I hypothesized that a person’s heartbeat could be detected using the smartphone’s built-in
accelerometer, which could be utilized as a method to differentiate between users. After collecting data from an app
created last year, I analyzed the data using an algorithm I wrote in Matlab to detect peaks in the acceleration signal
that could correspond to heart beat. The algorithm also creates a feature vector to hold calculated statistics which will
be used to create a template in order to compare different users.
134. Barcode Deconvolution via Wiener and Dictionary Analysis of Barcode Subsections
Bohyun Kim University of California, Irvine
Advisor(s): Yifei Lou, University of Texas Dallas
While current systems already provide reliable results in correctly reading barcodes at close range, we want to increase
the limits of these systems so they will be able to read these codes with extreme levels of blur and noise. We discover
that we can retrieve a complete blurring function from a subsection of a blurry barcode and created a system that can
analyze this subsection to deliver an estimated blurring function and a clean version of the barcode subsection that
when convoluted together can recreate the blurred barcode. This function uses the Wiener filter to analyze and filter
out noise and blur from a binary image: in our case, a picture of a blurry barcode, using a brute force method to try
every possible combination of the first two digits of a barcode and finding a function that when convoluted with a
clean subsection of the barcode, would return the result that best matches the original barcode. Through testing, our
method successfully analyzes a subsection of a blurry barcode and returns an estimate kernel that is very similar to
the actual function that blurred the image, possibly allowing future processes to accept even more limited data and
correctly reconstruct the original data.
135. Improving plotting algorithms in SAGE
Nathan Dasenbrock-Gammon Northern Kentucky University
Advisor(s): Steven Wilkinson, Northern Kentucky University
Computers plot graphs by selecting points at which to evaluate the function, and then connecting the dots. If too
few points are evaluated then the resulting graph is not smooth, too many points and the computer has to run more
41
Titles, Authors, Advisors and Abstracts
calculations than necessary, which can take extra computing time. This project worked to develop and test different
adaptive algorithms to determine where the computer needs to plot points so that a smooth graph is produced, but
while having to evaluate a minimum number of points. All of our algorithms were written and run in SAGE, which is
a python based, free and open source computer mathematics software.
136. The Accessibility Polynomial of a Sandpile Monoid
Bryan Oakley University of Georgia
Elizabeth Herman Marian University
Advisor(s): Luis Garcia, Sam Houston State University
Let € be a directed graph with a global sink s. A sandpile c is a vector of nonnegative integers indexed by the non-sink
vertices of €, where c.v/ represents the number of grains of sand at vertex v. If a sandpile has too much sand, the
unstable sandpile c may be stabilized by a sequence of vertex topplings where an unstable vertex v topples sending one
grain of sand through each of its outgoing edges. A stable sandpile c is accessible from a sandpile b if one can reach
c from b by a series of sand additions and topplings. The accessibility number of a stable sandpile c is the number of
stable sandpiles that access c. This naturally leads to the univariate accessibility polynomial of a sandpile graph €
A.x/ D
m
X
ai x i ;
i D1
where m is the total number of stable sandpiles in the graph € and ai is the number of stable sandpiles with accessibility
number i . In this talk we will focus on the second largest nonzero term in A.x/. It can be checked that the degree of
this term is at most m am . We present exact conditions for undirected graphs to have a.m am / D 1.
137. Parallel Implementation of Time-Decomposition Preconditioner for the Dirac Equation
Arthur Kurlej University of Massachusetts Amherst
Advisor(s): Jung-Han Kimn, South Dakota State University
We have implemented a fully implicit numerical approach based on space-time finite element methods for the onespace plus one-time dimensional Dirac equation. The purpose of this talk is to present a stable parallel implementation
of the physical system. The proposed numerical method is applied to generate a successful simulation results of
gauge free, massless solution of the Dirac equation. The time additive Schwarz method is vital to make successful
simulations with KSP (Krylov Subspace Methods) solvers. We show that proper time parallel implementation allows
for physically intuitive boundary conditions, improvement of numerical efficiency,and reduces the overall error of
the computed solution. The time parallel algorithm is implemented through PETSc (Portable, Extensible,Toolkit for
Scientic Computation, developed by Argonne National Laboratory).
138. Experimental and Numerical Studies of Oceanic Overflows
Frederick Hohman University of Georgia
Thomas Gibson Baylor University
Theresa Morrison San Diego State University
Advisor(s): Shanon Reckinger, Fairfield University
Oceanic overflows occur when dense water flows down a continental slope into less dense ambient water. The resulting density driven plumes occur naturally in various regions of the global ocean and affect the large-scale circulation.
The work presented here involves a direct qualitative and quantitative comparison between physical laboratory experiments and lab-scale numerical simulations. Laboratory experiments are conducted using a rotating square tank
customized for idealized overflow and a high-resolution camera mounted on the table in the rotating reference frame
for data collection. Numerical simulations are performed using the MIT general circulation model (MITgcm). The
governing equations are a set of coupled nonlinear partial differential equations called the non-hydrostatic primitive
equations, which are numerically solved using the finite-volume method. Resolution and numerical parameter studies
are presented to ensure accuracy of the simulation. The results are analyzed using various calculated metrics including
plume direction, plume velocity, and plume area. Laboratory and computational experiments are compared across a
wide range of physical parameters, including Coriolis parameter, inflow density anomaly, and dense inflow volumetric
flow rate.
42
Titles, Authors, Advisors and Abstracts
139. North Carolina and Pennsylvania for Boardwalk? Trade Values for Monopoly Real Estate
Crisel Suarez St. Edward’s University
Advisor(s): Michael Kart, St. Edward’s University
The objective of the game Monopoly is to become the wealthiest player through buying, renting and trading properties.
To improve a player’s chance of winning, a player can optimize it’s trading strategy. In this research a computer
program written in Python 3.2 is able to model, simulate 10 million Monopoly games and collect data on the expected
values of properties. For example, when two players are trading who would have a better deal? The data collected is
able to determine the expected value for each property, where a player can compare other player’s properties and decide
if they want to make the trade. Future works for this research include optimization of the model, trading properties for
money and applications to analytics.
140. Usage of FPGAs to implement complex algorithms used in linear systems and differential
equations solvers
Mai Dao Texas Tech University
Advisor(s): Brock Williams, Texas Tech University
FPGAs are very flexible computer chips, and they are used to enhance the performance of complex mathematical
solvers. The Verilog language in Xilinx program is used to write parallel hardware Conjugate Gradient implementation that executes numerical methods like the Finite Difference Methods and Finite Element Methods. Such methods
can be used to solve various partial differential equations like fluid dynamics, heat transfer, and Laplace transformation equations and to locate the areas of possible solutions for various linear systems. Not only is the cost of such
technologies greatly reduced, but this method is also super fast and can present novel approaches to complex mathematical challenges.
141. Motif-based clustering of directed networks
Thomas Reith Davidson College
Advisor(s): Laurie J. Heyer, Davidson College
The study of networks has become increasingly relevant in a variety of scientific fields, including biology, sociology,
and computing. It has been shown that many real networks divide naturally into clusters, or communities, of related
vertices. A large amount of research has focused on the detection of such clusters in undirected networks; their detection in directed networks, however, is a less well-studied problem. Here we introduce a novel clustering algorithm for
directed graphs based on the concept of network motifs—recurrent, statistically significant subgraphs found at higher
frequencies in real networks than ones that are randomly generated.
142. A Split-Explicit High Order Numerical Scheme for Atmospheric Applications
Juan Mora Arizona State University
Advisor(s): Mohamed Moustaoui, Arizona State University
A new numerical method is developed and applied to solve time-dependent partial differential equations involving
advection problems. This method is based on a recently published numerical scheme which uses leapfrog and a 4thorder implicit time filter. This scheme is split-explicit and uses small and large time steps applied to terms responsible
for fast and slow propagating waves respectively. Formal analysis of the stability and accuracy of this scheme will be
presented. This scheme uses only one function evaluation per time step, is of 3rd-order accuracy for amplitude error,
and damps the spurious waves due to computational modes inherent in the leapfrog scheme. The proposed scheme is
compared to the 3rd- Order Runge-Kutta method which also has 3rd order accuracy for amplitude errors but requires
three times as many function evaluations per time step. Examples demonstrating the performance of the proposed
scheme will be presented.These includes application of the proposed scheme to the Global Shallow Waters Spectral
Model on the Sphere, and nonlinear mathematical models used in regional Numerical Weather Prediction (NWP).
143. Coarse-Grain Model for Knotted Glueball Creation
Joseph Spitzer University of St. Thomas
Advisor(s): Eric J. Rawdon, University of St. Thomas
In the Standard Model, a quark and anti-quark can be bound together by the strong force to form a meson. In quantum
chromodynamics (QCD), a meson can be modeled as a flux tube connecting a quark and anti-quark. We can think of
the meson as a long cord with velcro at the two ends. If a fuzzy end touches a rough end, then they join together. If
Titles, Authors, Advisors and Abstracts
43
the fuzzy and rough ends come from different mesons, then a new meson is formed with a longer cord. If the fuzzy
and rough ends come from the same meson, then this connection forms a closed tube that can be knotted and/or linked
with other closed tubes. Our ultimate goal is to simulate the creation of these glueballs to compare to experimental
data. Currently, we employ a coarse-grain model where the mesons initially are represented by directed line segments
randomly placed in a cube. Polygonal arcs are then formed by connecting heads and tails that are “close” to each other.
The net result is a sort of bowl of polygonal spaghetti. The density of the initial set of directed line segments within the
cube affects how many of these noodles are closed vs open and the knotting/linking of the noodles. We present initial
results on this density effect. Future work will include adding velocities or Brownian motion to the polygonal noodles.
144. Encrypting Data For Web Applications
Jassiem Ifill Morehouse College
Advisor(s): Elaine Shi, University of Maryland, College Park
The primary goal of the research presented in this paper is to enhance the capabilities of the ShadowCrypt web
extension in order to provide more privacy and authentication when sharing information within web applications.
Throughout this summer, the section of the research in the Laboratory that was focused on enhancing ShadowCrypt
attempted to improve the previously established asymmetric implementation, instead of the symmetric implementation
which was not ideal. As such, this project builds upon the previous version of the ShadowCrypt extension, in order
to make the asymmetric implementation more practical and easier to use in addition to increasing the extension’s
richness of features. The final result of this project was an improved version of the asymmetric implementation of
the ShadowCrypt web extension, which allowed users to input their personal private keys, and manually input public
keys. Also, this version of ShadowCrypt also possesses the feature to search, authenticate, and grab an individual’s
public key from Keybase.io, which is a public directory of public keys. Currently, this version of ShadowCrypt has
been tested to work with gmail.com, Facebook, and Twitter among a few other web applications/domains.
145. Large Scale Computer Visualization
Omar Monarrez Texas Tech University
Advisor(s): G. Brock Williams, John Calhoun, Texas Tech University
High performance computers are designed to deal with big data sets, but oftentimes visualization tools need to be
combined with high performance computing in order to render the data intelligible. Our new GPU cluster provides
the raw computing power to run large scale simulations and analyze the resulting large data sets. We will describe
our design for a large scale visualization system to complement the computer cluster. For example, when the disease
simulation written for the cluster is running, each compute node will produce a graphical representation of the current
extent of the outbreak for the area under that node’s control. Combining these into a coherent whole will allow the
large scale behavior of the disease to be visualized.
146. Implementing Multi-Agent Simulations
Catherine Nicosia Texas Tech University
Advisor(s): G. Brock Williams, John Calhoun, Texas Tech University
We will describe our work on the design and implementation of a large scale GPU cluster for multi-agent simulation. The new Schoenberg GPU cluster in our LAZARUS lab is designed for conducting large scale simulations and
analyzing the resulting big data sets. For example, the multi-agent disease simulation software models the spread of
a disease by simulating the movement of healthy, infected, and immune individuals throughout a city or continent.
However, constructing and operating the cluster created numerous hardware and software challenges. For example,
we will discuss our work creating a model for understanding the temperature and humidity around the computer and
developing software and circuitry for the needed sensors and environmental controls.
147. Averages in Tree Space
Amber Holmes LaGrange College
Mercedes Coleman Lamar University
Emily Smith Kenyon University
Cody FitzGerald University of New Hampshire
Advisor(s): Seth Sullivant, North Carolina State University
As part of the North Carolina State University Research Experience for Undergraduates program in the Department of
Mathematics, we explored a variety of means in tree space. Tree space is a subset of Euclidean space in which every
44
Titles, Authors, Advisors and Abstracts
point represents a tree. In this space, there exist numerous metrics, each of which lead to a different notion of a mean
tree. We have implemented the algorithm for the .1; 1/, .2; 2/ and .1; 1/-mean in MATLAB.
The .1; 1/-mean is a weighted majority rule consensus tree, the .2; 2/ mean is generated using Sturm’s algorithm.
Unlike the .1; 1/ and .2; 2/-mean, the .1; 1/ mean is not as well studied. The .1; 1/-mean algorithm we propose
incorporates traversing the L2 geodesic using the L1 midpoint as a mechanism for obtaining the mean tree’s orthant.
Once the orthant has been found, the mean tree can be optimized using linear programming.
148. Do Polygons Become Asymptotically Regular under Flow by Curvature?
Ryan Gallagher University of Connecticut
Jessie De La Cruz Santos College of the Holy Cross
Sarah Hadaidi University of Kentucky
Advisor(s): Andrew Cooper, North Carolina State University
The process of deforming a curve by the curvature vector at each point is known as the curve-shortening flow (CSF).
Grayson showed that CSF averages the shape of a curve, causing simple, closed curves to become asymptotically
circular. Our research applies the idea of curvature flow to polygons. We give a novel definition for the ’curvature
vector’ at a vertex of a polygon. Deforming in the direction of this curvature vector yields a flow of polygons, the
polygon curvature flow (PCF). We numerically investigate the behavior of this flow and show that it exhibits several
of the qualitative properties of CSF. We conjecture that PCF makes polygons asymptotically regular.
149. Centers of Triangles: Parameterization of the Distribution
Andrew Arnold University of Alabama at Birmingham
Advisor(s): John Mayer, University of Alabama at Birmingham
In response to Mueller’s investigation of the loci of centers of triangles for a fixed center, we introduce a reparameterization in rectangular coordinates of the original problem in search of rigorous results. To study the distribution
of triangle centers, we consider two families of triangles: those with fixed circumcircle, and those with fixed incircle,
which represent all triangles up to rotation and dilation. Our goal is to find both synthetic and analytic arguments for
the distribution of centroids, circumcenters, incenters, and orthocenters, when applicable, for the two families and discuss unexpected behavior of various distributions. In addition, we discuss the use of the dynamic geometry software
package GeoGebra in the process of investigating the distribution, making conjectures, and arriving at both synthetic
and analytic proofs of results.
150. Constant Vector Curvature in Three Dimensions
Albany Thompson Central Washington University
Advisor(s): Corey Dunn, California State University San Bernardino
Differential geometry is the use of the techniques and tools of calculus to study the geometric properties of manifolds.
One of the most commonly studied properties of manifolds is its curvature. After a metric is defined on the manifold,
an algebraic curvature tensor is a pointwise measurement of this curvature, and a model space is this collective information at one point. At any given point of the manifold, there are several curvature conditions that it can satisfy. This
research is concerned with the necessary and sufficient conditions for a model space in three dimensions with positive
definite inner product to have the specific curvature condition known as constant vector curvature. This poster outlines
these conditions and provides examples of model spaces that fall into several different curvature classifications.
151. On the Levi graph of point-line configurations
Jessica Hauschild Kansas Wesleyan University
Jazmin Ortiz Harvey Mudd College
Advisor(s): Oscar Vega, California State University, Fresno
Given a certain point-line configuration C one may construct its Levi graph (AKA incidence graph), Levi(C ), by
connecting point-vertices with line-vertices if and only if they are incident in C . In this work we assume C to be
a .vr ; bk / configuration, and we investigate the vector space of vertex weightings of Levi(C ) that are constant on all
maximal independent sets of the graph, and prove that its dimension, known as the well-covered dimension of Levi(C ),
is always equal to 0, as long as r > 2.
Titles, Authors, Advisors and Abstracts
45
152. Geometric Analysis of Chimu Ceramics
Joseph Hopkins Georgia College and State University
Matthew Kennedy Georgia College and State University
Advisor(s): Ryan Brown, Georgia College and State University
The Chimu Empire, an ancient pre-Columbian culture in Peru, had significant collections of ceramics and textiles.
The ceramics were used for communication, to display societal status, and for religious ceremonies. Most research
into these ceramics has focused on artistic and historical meaning. In this research we examine the ceramic works of
the Chimu culture using mathematical analysis of their symmetries. We compare the geometry and symmetries in a
sample of Chimu ceramics. We characterize structural symmetry, iconographic symmetry, and symmetry with respect
to color. One goal of this project is to understand the relationship between the geometry of the Chimu in relation to
other pre-Columbian cultures.
153. Investigating Square Hyperbolas and Dirichlet Domains under a non-Riemannian Metric of the
Bidisk.
Hassan Nasif Wake Forest University
Advisor(s): Melissa Macasieb, William Goldman, University of Maryland, College Park
The Dirichlet Domain admitted by cyclic Fuchsian Groups on hyperbolic 2-space has been shown to only have two
faces. In hyperbolic 3-space, the Dirichlet Domain admitted by the same isometries can have an infinite amount of
faces. The bidisk, a cross of hyperbolic 2-space with itself, exhibits properties of both hyperbolic 2-space and 3space, While it has been proven that a two-faced Dirichlet Domain can be admitted, computational evidence suggests
that Dirichlet Domains in the bidisk can have more than two faces, but a constructive proof is not yet known. In
further understanding the Dirichlet Domains and bidisk as a geometric space, we investigated the bidisk under a nonRiemannian metric and determined which properties and theorems remained true under the new metric.
154. Geometric Analysis of Chimu Textiles and Metalwork
Sidney Kennedy Georgia College and State University
Advisor(s): Ryan Brown, Georgia College
In this poster we examine geometric patterns and symmetries found in ancient textiles and planar metalwork of the
Chimu, a pre-Columbian civilization of northern Peru. There are few extant Chimu textiles, but we study several pieces
and compare patterns with several examples of planar Chimu metalwork still in existence. We observe reflection,
rotation, and translation symmetries, and show how the Chimu incorporate subtle disruptions of symmetry.
155. Geometric Properties of Conformal Transformations on R p;q
Surya Raghavendran University of Texas at Austin
Advisor(s): Matvei Libine, Indiana University
We show that conformal transformations on the semi-Euclidean space Rp;q map hyperboloids and affine hyperplanes
into hyperboloids and affine hyperplanes. We also show that this action is transitive when p or q is 0, and that this
action has exactly three orbits otherwise. We then extend these results to hyperboloids and affine planes of arbitrary
dimension. These properties generalize the well-known properties of fractional linear transformations on the complex
plane C.
156. Sphere Packings in Non-Euclidean Spaces
Sarah Stoops Butler University
Advisor(s): Prem, Sharma
Imagine emptying a sack containing small equal-sized ping pong balls to fill a large container. Experiment shows that
the balls will occupy nearly 65% of the space in the box. However, a much higher packing density is achieved by filling
the box layer by layer: Arrange the balls in the bottom layer as closely together as possible in a hexagonal lattice. Then
in each succeeding layer, place the balls in the indentations created by the layer directly below it. If the container is
the whole 3-dimensional space, the problem of finding a densest packing is the celebrated Sphere Packing Problem of
Johannes Kepler (1611). Now suppose that instead of perfectly round spheres (ping pong balls), one were to fill the
46
Titles, Authors, Advisors and Abstracts
box with identical prolate spheres (like apples) or oblate spheres (like American footballs). In this study, we introduce
non-Euclidean geometries in the plane and 3-space to provide some good estimates for such packings.
157. Mean Curvature Zero Surfaces in Three-dimensional Lie Groups
Joseph Gills Longwood University
Advisor(s): Thomas Wears, Longwood University
A translation surface in a three-dimensional Lie group G is a surface that is generated from two curves ˛ W R ! G
and ˇ W R ! G by using the group operation of G to “multiply” alpha and ˇ together. Namely, one has the
parametric surface S parametrized by r W R2 ! G, where r .u; v/ D ˛ .u/ ? ˇ .v/ and ? is the group operation of G.
We investigate translation surfaces in the three-dimensional Heisenberg group and the three-dimensional group Sol3 ,
where the groups are equipped with a left invariant (Riemannian or Lorentzian) metric. Building off of the work of
P. Piu and A. Sanini we focus on the case where our generating curves are integral curves of left invariant vector fields
and/or geodesics and we present partial classification results on mean curvature zero surfaces of the indicated form.
158. Number Sequences for Rectified Cross Polytopes and 4-D regular Polytopes
Ryan Brown Grove City College
Advisor(s): Michael Jackson, Grove City College
Polytope numbers are a non-negative number sequence constructed from the geometry of a polytope. Rectification
is the process of cutting of a vertex to the center of each edge connected to the vertex. We extend previous results
to show a formula for the polytope numbers of an rectified Cross Polytope in any dimension. We use a stepping up
inclusion exclusion and a geometric argument called Kim’s process to cut the rectified cross polytope into smaller,
same dimensional, cross polytopes. Our proof takes advantage of a number of identities relating cross polytopes and
simplexes to each other and smaller or lower dimensional of themselves.
159. Graphs with minimal well-covered dimension
Gabriella Clemente The City College of New York
Advisor(s): Oscar Vega, California State University, Fresno
A graph is well-covered if all of its maximal independent sets have the same cardinality. The well-covered space of
any graph, G, is the vector space of all functions on V .G/ that ‘make’ G well-covered, and the dimension of this
space is the well-covered dimension of G, denoted by wcd i m.G/. In 2005, Brown and Nowakowski proved that
.1/ wcd i m.G/ sc.G/, where sc.G/ is the number of simplicial cliques of G, and that .2/ wcd i m.G/ D sc.G/
if G is a chordal graph. In this poster, we present our main theorem, a generalization of .2/ that yields a family of
graphs with minimal well-covered dimension that contains chordal graphs. We also present our principal theorem on
the well-covered dimension of Sierpinski gasket graphs, Sn with n 2, which states that wcd i m.Sn / D 3, where
sc.Sn / D 3.
160. Radio Number for Seventh Powered Paths
Ilia Gonzales California State University
Gilbert Felix California State University San Beranardino
Antonio Saucedo California State University San Bernardino
Osvaldo Gonzalez California State University San Bernardino
Advisor(s): Min Lin-Lo, California State University San Bernardino
Let G be a connected graph. For any two vertices u and v, let d.u; v/ denote the distance between u and v, which is
the smallest length of any u v path in G. The diameter of G, denoted by diam.G/, is the greatest distance between
any pair of vertices in G. A rad i o label i ng (or multi-level distance labeling) of G is a function f that assigns to
each vertex with a label from the set f0,1,2,3,. . . g such that the following holds for any two vertices u and v:
j f .u/
f .v/ j 1 C d i am.G/
d.u; v/
The span of f is defined as maxu;v2V .G/f j f .u/ f .v/ jg. The radio number of G, denoted by r n.G/, is defined
as the minimum span of all radio-labelings for G. The goal of this presentation is to discuss the progress we made
towards finding the radio number for the seventh power of any path during a 2014 research program which is an MAA
activity funded by NSF (grants DMS-1156582 and DMS-1359016).
Titles, Authors, Advisors and Abstracts
47
161. Visualizing Dessins D'Enfants
Susan Maslak Ave Maria University
Mary Kemp Occidental College
Advisor(s): Naiomi Cameron, Lewis and Clark College
Our poster summarizes the results of a summer research project motivated by the theory of dessins. Dessin is short for
dessin d’enfant which means child’s drawing. Mathematically speaking, a dessin is a connected bicolored graph where
the edges around every vertex are cyclically ordered. Dessins can be realized by Belyi maps which are meromorphic
functions f W X ! P 1 .C/ such that X is a Riemann surface and f is unramified outside f0; 1; 1g: One of the
goals of this project is to determine Belyi maps that realize a given loopless, connected bipartite graph on a compact
Riemann surface X. We report on our considerations of certain classes of such graphs, explorations of computational
methods for finding associated Belyi maps and related applications.
162. Graphs with Characteristic-dependent Well-covered dimension
Joseph Burdick Humboldt State University
Advisor(s): Oscar Vega, California State University Fresno
Given a graph G and a field F , the F -vector space of all vertex-weights that are constant on the maximal independent
sets of is called the well-covered space of G. The dimension of this space is the well-covered dimension of G (relative
to F ). For some graphs, their well-covered dimension can depend upon the characteristic of the field used for scalars.
We investigate graphs that have this property and show how more graphs with this property can be constructed.
163. Reliability of Recursively Constructed Graphs
Hanah Goetz The University of Texas at Tyler
Jon Erickson Rice University
Cecile Cornelus Adelphi University
Advisor(s): Christina Graves, The University of Texas at Tyler
The probability that a given graph G remains connected if each edge is included with probability p is the all-terminal
reliability polynomial of G. A graph whose reliability polynomial is greater than all other graphs on the same number
of vertices and edges is call a uniformly most reliable graph (UMR). We first investigate the UMR for some specific
number of vertices and edges. We then apply a recursive process to the Petersen graph and compare the reliability
polynomials of the resulting graph families.
164. Counting Symmetric Fullerene Patches with 4 Pentagons
Armando Grez Florida Gulf Coast University
William Linz Texas A&M University
Heather Chiros Massachusetts College of Liberal Arts
Advisor(s): Stephen Graves, University of Texas at Tyler
This study examines a method for constructing fullerene patches in the hexagonal tessellation of the plane. We extend
a result of Graves and Graves (2013) by producing an exact process for drawing fullerene patches with 4 pentagonal
faces embedded in them. Attempts to formulate a closed equation for the number of symmetric patches, up to isomorphism, are shown. These include direct counting by predetermining the placement of two pentagonal faces and noting
graphical symmetries that cause overcounting, using the Principle of Inclusion-Exclusion approach and by recoordinatizing our depiction from Coxeter coordinates to Euclidean in order to analytically count the patches within the largest
convex polygon interior to all the boundary lines we construct.
165. The Number of Seymour Vertices in Random Tournaments and Digraphs
Yiguang Zhang Johns Hopkins University
Advisor(s): Anant Godbole, East Tennessee State University
Seymour’s distance two conjecture states that in any simple digraph with no anti-parallel arcs, there exists a vertex
(a “Seymour vertex”) that has at least as many neighbors at distance two as it does at distance one. We explore the
validity of probabilistic statements along lines suggested by Seymour’s conjecture, proving that almost surely there
are a “large” number of Seymour vertices in random tournaments and “even more” in general random digraphs.
48
Titles, Authors, Advisors and Abstracts
166. Knight's Tours on Boards with Odd Dimensions
Stephanie DeGraaf Iowa State University
Advisor(s): Steve Butler, Iowa State University
A popular form of recreational mathematics deals with chess problems, such as a knight’s moves on a chess board.
In particular, a closed knight’s tour is a sequence of knight moves where each square is visited exactly once and the
sequence begins and ends with the same square. The problem of determining when a board has a closed knight’s tour
was solved by Schwenk, with the condition that the size of the board must be of even dimension. This leads to the
question of how to construct a closed knight’s tour on boards of odd dimension. To do so, we must delete a square in
order to have an equal number of black and white squares, so for boards of size m n where m and n are odd, we
determine which square to remove to allow for a closed knight’s tour.
167. Saturation Number of Trees in the Hypercube
Kavish Gandhi Newton North High School
Advisor(s): Chiheon Kim, Massachusetts Institute of Technology
A graph H 0 is .H; G/-saturated if it is G-free and the addition of any edge of H not in H 0 creates a copy of G. The
saturation number sat.H; G/ is the minimum number of edges in an .H; G/-saturated graph. We investigate bounds
on the saturation number of trees T in the n-dimensional hypercube Qn . We first present a general lower bound on
the saturation number based on the minimum degree of non-leaves. From there, we suggest two general methods for
constructing T -saturated subgraphs of Qn , and prove nontrivial upper bounds for specific types of trees including
paths, generalized stars, and certain caterpillars under a restriction on minimum degree with respect to diameter.
168. The Abelian Sandpile Model on Fractal Graphs
Rafael Setra University of Maryland
Advisor(s): Robert Strichartz, Cornell University
The Abelian Sandpile Model is a process where chips are placed on a graph’s vertices. When the number of chips on a
vertex exceeds its degree, chips are distributed to each neighboring vertex by a process called toppling. On the integer
lattice the resulting stable configurations form interesting fractal patterns. Using these fractal patterns as motivation,
we examine resulting configurations on graph approximations of post critically finite fractals. In particular, we look
for connections between the boundary growth of the stable configurations and the initial number of chips. We also
prove that the resulting configurations are periodic with respect to the number of chips. We may further generalize this
periodic behavior on arbitrary graphs.
169. Characterizing Veto Interval Graphs and Related Graph Classes
Jessica Kawana Willamette University
Advisor(s): Joshua D. Laison, Willamette University
Given a set S of intervals on the real line, we can construct a graph with a vertex for each interval in S , and an edge
between two vertices if and only if their corresponding intervals intersect. Conversely, given a graph G, G is an interval
graph if it has such a representation. We define a variation of interval graphs called veto interval graphs, where each
interval contains a veto mark, and two intervals do not count as intersecting if either contains the veto mark of the
other. We explore properties of veto interval graphs and their relation to other classes of geometric intersection graphs.
170. A study of competition graphs induced by permutations
Elizabeth Yang Princeton University
Advisor(s): Brian Nakamura, Rutgers University
This project introduces a means to unite two active areas of research namely, the study of competition graphs and
the study of patterns in permutations. We extend the work of Cho and Kim on the competition graphs of doubly
partial orders of sets of points in R2 . Here, we study the competition graphs arising from permutations (using the
doubly partial order on the permutation’s graphical representation). We find many interesting connections especially
between the graphs of permutations that avoid the pattern 1 3 2 and permutations that avoid the pattern 1 2 3. These
connections involve both the structure of the graphs involved, and the enumerative aspects of these graphs (i.e. how
many permutations form a certain type of graph). This was joint work with Brian Nakamura.
Titles, Authors, Advisors and Abstracts
49
171. Construction and Optimality of Undirected de Bruijn Sequences
Christie Burris Colorado State University
Advisor(s): Patrick Shipman, Colorado State University
For positive integers k and n, what is the minimal length of a word over an alphabet of size k which contains every
length-n word as a subword? Clearly the minimum possible length of such a word is k n C n1, as this length is
required to see all k n subwords without repetition. Words that achieve this bound are commonly referred to as de
Bruijn sequences. Applications of de Bruijn sequences arise as optimally random sequences in coding theory as well
as in the design of DNA microarrays. De Bruijn sequences on the alphabet corresponding to the set of amino acids,
fA; D; C; Gg, which comprise DNA can be used to see a great variety of binding patterns relative to the length of the
sequence. We introduce a variation on the idea of a de Bruijn sequence, as exemplified by the sequence 00010111.
This sequence sees each of the binary words of length 3 as subwords with the minimal number or repetitions when
read both left-to-right and right-to-left. We call such a sequence an undirected de Bruijn sequence. The purpose of this
research is to determine the lengths of undirected de Bruijn sequences and develop methods for their construction.
172. Studying Brain Connectivity using Weighted Graph Comparison
Alexander Berger New York City College of Technology
Thierno Diallo CUNY City Tech
Advisor(s): Urmi Ghosh-Dastidar, New York City College of Technology
Studying brain components and its connectivity is an important field in neuroscience. While concepts of weighted
graphs are widely used in many areas including computer, social, biological pathways, and air traffic networks, application of weighted graphs to study brain connectivity pattern is relatively new in the field of graph theory. In
this project we focus on anatomical connectivity that connects nodes representing regions of interest (ROIs) and the
weighted edges associated with structural connectivity, i.e., the density fibers connecting ROIs. An adjacency matrix
using connectivity weights between ROIs is created and Laplacian spectrum analysis and spectral clustering method
are used to study the connectivity strength between and within two cerebral hemispheres.
173. A Generalization of Higher Dimension Fibonacci Matrices and Their Graphs
Zachery Lancto Westfield State University
Advisor(s): Karin Vorwerk, Westfield State University
The 2 2 Fibonacci matrix is well studied in mathematics. One can interpret this Fibonacci matrix as a transition
matrix and thus define the associated directed graph. The concept of the Fibonacci matrix and its associated graph can
easily be extended to higher dimensions using concepts and techniques from linear algebra, graph theory, and discrete
mathematics. We define nxn Virahanka matrices as the generalization of the Fibonacci matrix. A Virahanka matrix has
the property that, when raised to integer powers, all its entries are Fibonacci numbers. We examine n n Virahanka
matrices where n > 2 and discuss their underlying graphs and recurrence relationships. We also provide an existence
argument for certain higher dimensions and show that Virahanka matrices in higher dimensions are neither unique nor
necessarily possess isomorphic associated graphs. By studying these matrices and graphs, we discover recurring subgraphs and patterns in the graphs. We propose an upper limit to the minimum number of edges needed for a graph to
define an nxn Virahanka matrix and establish and prove many of its general properties. We built on work first presented
by Brian Bowen in 2008.
174. Properties of the Generalized Type A and Full Flag Johnson Graphs
Linhan Chen Boston University, Program in Mathematics for Young Scientists (PROMYS)
Jongwon Kim Boston University, PROMYS
Advisor(s): Dylan Yott, University of California, Berkeley
We present a study of a family of graphs called the Johnson graphs, as well as one of its generalizations, the full flag
Johnson graphs. Briefly mentioned in algebraic graph theory texts, Johnson graphs contain structural and algebraic
properties that are analogous to those of the Grassmann graphs, the Hamming graphs, and the Kneser graphs. While the
Johnson graphs have been well-studied with regards to their applications in coding theory, the full flag Johnson graphs
have been relatively unexplored. Apart from the fact that they are regular, are connected, form permutohedra in some
cases, and are Cayley graphs over Sn , not many of their properties are known. In our research, we focus on previously
unexplored properties of both the Johnson and full flag Johnson graphs through novel approaches. Specifically, with
50
Titles, Authors, Advisors and Abstracts
the aid of the Sage program to generate examples, we derive a recursion for the regularity of the full flag Johnson
graphs through combinatorial considerations. We also exploit a structural property within the full flag Johnson graphs
to determine the planarities of all cases. We then proceed to derive the clique size of the Johnson graphs, before proving
the existence of a graph isomorphism between different classes within the Johnson graph family.
175. Combinatorial Analysis of the Game Grim
Richard Adams California State University, Fresno
Janae Keys Fresno State University
Advisor(s): Oscar Vega, California State University, Fresno
The game Grim is a subtraction game played on a graph. A move consists of deleting an available vertex (one that
has not already been deleted), which also deletes vertices that become isolated as a result of this move. As usual, the
last player who can make a move wins. Some was already known about Grim played on certain families of graphs
(certain paths, cycles and wheels, and complete bipartite graphs and grids). Not surprisingly, the analysis of the game
gets really complex when the number of vertices grows, even for very simple graphs, such as even paths. We focused
on two problems:
textbf(1) Large graphs: What are winning strategies when Grim is played on graphs with many vertices.
textbf(2) Randomized playing: What happens when Grim is played randomly (no strategy, or just using a ‘local’
strategy) on a given graph.
176. k-dependence on Hexagonal Boards
Robert Doughty Miami University
Berkeley Bishop Anderson University
Jessica Gonda University of Akron
Adriana Morales University of Puerto Rico at Rio Piedras
Josiah Reiswig Minnesota State University Moorhead
Katherine Slyman University of Delaware
Advisor(s): Reza Akhtar, Miami University
Combinatorial chessboard problems have been studied extensively in the literature. A particular question of interest
concerns determining the maximum number of kings that can be placed on an m n chessboard such that no king
attacks more than k others; similar questions have also been studied on triangular boards. We investigate this question
when the chessboard is in the shape of a rhombus tiled by regular hexagons of the same sidelength. We provide upper
and lower bounds for this quantity, using vertex discharging methods to improve the upper bound in certain cases.
177. On 2
2 Graph Achievement Games
Curtis Clark Jr. Morehouse College
Advisor(s): Curtis Clark, Morehouse College
Let F be a graph with no isolated vertices. The 2 2 F -achievement game on the complete graph Kn is described
as follows. Player A first colors two edges of Kn green. Then Player B colors two different edges of Kn red. They
continue alternately coloring the edges with Player A coloring at most two edges green and Player B coloring at most
two different edges red. The graph F is achievable on Kn if Player A can make a copy of F in his color. The minimum
n such that F is achievable on Kn is the 2 2 achievement number of F , a.F /. The 2 2 move number of F , m.F /,
is the least number of edges that must be colored by Player A to make F on the complete graph with a.F / vertices.
The numbers a.F / and m.F / are determined for some small graphs and paths.
178. Colored strongly regular, and Colored quasi-random graphs
Mark Rychnovsky University of California, Santa Barbara
Advisor(s): Padraic Bartlett, University of California, Santa Barbara
Strongly regular graphs are often used for their powerful symmetries. Here we generalize the notion of a strongly
regular graph to k colors giving what is often an even stronger symmetry group. We will see that several natural
examples of strongly regular graphs extend cleanly to k colors, particularly those arising from affine spaces, and finite
fields. We will also examine the similarities between these graphs and the random graph on k vertices.
Titles, Authors, Advisors and Abstracts
51
179. An Excursion into the Collatz Conjecture and Related p-adic Topologies
Joseph Hoover Butler University
Advisor(s): Prem Sharma, Butler University
Let G be the graph with the set of odd positive integers D C as its vertex set and with a directed edge from each vertex
n to g.n/, where g.n/ is the largest element of D C that divides 3n C 1. Furthermore, let P be the component of G
containing 1. We elegantly embed G in the Euclidean plane as such: place the vertex 1 at the origin and each vertex
v 2 P on the circle of radius equal to the graph distance of v from 1. This partition of P into points on concentric
circles reveals many interesting relationships among the odd integers by inducing a topology on D C which we extend
to p-adic topologies (for p D 2 and p D 3) on the set of rational numbers. These relationships certainly possess
intrinsic value regardless of whether the Collatz Conjecture is true or not.
180. Accelerating Stochastic Local Search Algorithms with Markov Clustering
Timothy Goodrich Valparaiso University
Advisor(s): Erik Ferragut, Oak Ridge National Laboratory
The Travelling Salesperson Problem (TSP) occurs frequently in modern applications such as traffic routing, power grid
design and circuit layout, and yet the NP-completeness of the problem suggests that a quick algorithm does not exist for
finding an optimal solution to any given problem instance. Our purpose is to develop quick approximation algorithms
for TSP instances with Euclidean distance between vertices (Euclidean TSP). Specifically, our work concentrates on
implementing clustering techniques such as Markov Clustering (MCL) to provide the algorithms with valuable metadata. Repeating this process by clustering the clusters themselves, we gain access to an innate ontology on the data set.
We analyze the use of this ontology on both a theoretical and an experimental level. Theoretically, we analyze run time
and solution quality when given certain properties of a clustering. Experimentally, we verify that our algorithms work
well in practice when compared to current state-of-the-art algorithms. By examining the existing, natural clusters in
the data, we define structures that reflect and utilize the nuances of each data set. We find that this algorithm design
choice drastically reduces run time while maintaining solution quality.
181. Monomization of Power Ideals and Generalized Parking Functions
Brice Huang MIT
Advisor(s): Wuttisak Trongsiriwat, MIT
A power ideal is an ideal in a polynomial ring generated by powers of homogeneous linear forms. For a graph G and
an abstract simplicial complex , we define a power ideal JG; and a monomial ideal IG; . We prove that IG; has
the same Hilbert series as JG; , and that their quotient algebras AG; and BG; share a monomial basis. This result
generalizes and interpolates between the known monomization theory of the central and external zonotopal algebras by
Postnikov-Shapiro and Desjardins, respectively. We also interpret the common Hilbert series of the algebras AG; and
BG; in terms of generalized parking functions. We construct a bijection between a generalization of the G-parking
functions and a class of forests of G, extending the known bijection between G-parking functions and spanning trees
of G.
182. Forbidden Subgraphs of Competition Graphs
Aquia Richburg Morehouse College
Advisor(s): Gene Fiorini, Rutgers University
Since the introduction of competition graphs much work has been done to study their properties. In this paper we look
at classifying forbidden subgraphs of competition graphs of doubly partial order. These results also extend to n-tuply
partial order. This paper also looks at the correlation of Dyck paths, Catalan numbers and partially ordered sets.
183. On the Minimal Reset Words of Synchronizing Automata
David Stoner South Aiken High School (RSI MIT research)
Advisor(s): Chiheon Kim, Massachusetts Institute of Technology
Cerny’s Conjecture is a 50-year old question which concerns the combinatorial field of synchronizing automata. In
particular, it postulates that the maximal length of the minimal reset word among all n-state automata is .n 1/2 . A
proof is presented for Pin’s Theorem, which applies Cerny’s conjecture to p-state automata consisting of a cycle and a
52
Titles, Authors, Advisors and Abstracts
non-permutation, where p is an odd prime. Also, families of automata of the form F .p; k/ are introduced; they consist
of a cycle and a group of k disjoint merging arcs. C.p; k/ is defined to be the maximal length of minimal reset words
within these families. A lower bound of C.p; k/ for general k is demonstrated, and the exact value of C.p; 2/ is found.
184. A Game of Cops and Robbers on Generalized Petersen Graphs
Jonathan Guzman California State University, Long Beach
Nikolas Schonsheck Vassar College
Advisor(s): Robert Bell, Michigan State University
The game of Cops and Robbers is a perfect information, vertex-pursuit game in which a set of “cops” and a “robber”
occupy vertices in a graph G. The cops place on vertices of the graph, followed by the robber and then both alternate
turns, either moving to adjacent vertices or passing. The copnumber of a graph, c.G/ is the least number of cops
required to catch the robber. We study the game on Generalized Petersen graphs. Let n and k be positive integers
such that n 5 and 1 k b n 2 1 c. The Generalized Petersen graph GP .n; k/ is the undirected graph with vertex
set fa1 ; : : : ; an ; b1 ; : : : ; bn g and the edges .ai ; ai C1 /; .ai ; bi /, and .bi ; bi Ck / for each i D 1; : : : ; n with indices read
modulo n. We lift the game to an infinite cyclic cover, GP .1; k/, which has the vertex set fai j i 2 Zg [ fbi j i 2 Zg
and edges defined analogously. There is a graph homomorphism W GP .1; k/ ! GP .n; k/ defined by reducing
indices modulo n. The copnumber of every Generalized Petersen graph is less than or equal to 4.
185. Connected Matchings in Graphs of Independence number 2
Jung Yoon Kim Thomas Jefferson High School for Science and Technology
Advisor(s): Peter Csikvari, Massachusetts Institute of Technology
In this project, we study connected matching in graphs of independence number 2. This problem is motivated by
Hadwiger’s conjecture. We conjecture that if G is a graph on 4n 1 vertices with independence number 2, then it
contains a connected matching of size n. We confirm our conjecture for n less than or equal to 13.
186. Cyclic Jacobians of Nearly Complete Graphs
Jeffrey Davis University of South Carolina
Joshua Klarmann Kansas State University
Advisor(s): Luis Garcia, Sam Houston State University
The Jacobian group of a graph is defined as the cokernel of the reduced Laplacian of the graph. The order of this finite
Abelian group equals the number of spanning trees in the graph. Its structure is an isomorphism invariant that has
received much attention recently in diverse areas of mathematics including algebraic and tropical geometry, combinatorics, and theoretical computer science. A nearly complete graph is obtained by removing n 3 edges from a complete
graph Kn on n vertices. In this paper, focus is given to removing families of graphs on n 3 edges from Kn that yield
cyclic Jacobian groups. Among these families of graphs are forests of vertex disjoint paths, and almost-paths, i.e., the
vertex-join of a path and a single edge. Necessary and sufficient conditions for the cyclicity of the groups associated
with the complements of vertex-disjoint paths on Kn are presented. We also determine the cardinality of these groups
as well as for those associated with almost-paths based on tree counting methods for graphs. These cardinality results
are given in terms of Chebyshev polynomials of the second kind. Finally, an alternate proof is given for a previously
held result about an ordering on almost-paths.
187. Laplacians Associated with Regular Bipartite Graphs and Application to Finite Projective Planes
Jesse Zhang Fairview High School
Advisor(s): David B Rush, Massachusetts Institute of Techonology
Internet search technology is a pervasively used utility that relies on techniques from the field of spectral graph theory.
We present a novel spectral approach to investigate an existing problem: the critical group of the line graph has been
characterized for regular nonbipartite graphs, but the general regular bipartite case remains open. Because of the
ineffectiveness of previous techniques in regular bipartite graphs, our approach provides a new perspective and aims
b We obtain a
to obtain the relationship between the spectra of the Laplacians of the graph G and its line graph G.
theorem for the spectra of all regular bipartite graphs and demonstrate its effectiveness by completely characterizing
the previously unknown critical group for a particular class of regular bipartite graphs, the incidence graphs of finite
53
Titles, Authors, Advisors and Abstracts
projective planes with square order. This critical group is found to be Z2 ˚ .Z2qC2 /q
q is the order of the finite projective plane.
3
1
˚ .Zq 2 CqC1 /q
2 Cq
1
; where
188. Diameters of polytope graphs and an improved upper bound on subset partition graphs
John Gallagher University of Wisconsin-La Crosse
Advisor(s): Edward D. Kim, University of Wisconsin-La Crosse
We study polytopes, which are generalizations of three-dimensional polyhedra. The graphs of polytopes are studied
in mathematical optimization because of their relation to the simplex method for linear programming. Due to the
relationship between the diameters of polytopes and the efficiency of the simplex method, we analyze diameters of
abstract polytope graphs. Abstract polytope graphs are graphs which encode combinatorial data about polytopes. We
present an analogue of Todd’s recent bound for polytopes in the setting of subset partition graphs, a generalization of
abstract polytopes.
189. Teaching Approximation Using Technology: A Survey of Calculus Instructors' Perceptions
Samantha MacMillan Nichols College
Advisor(s): Nicholas Gorgievski, Nichols College
This paper will discuss the ways and extents to which U.S. higher education calculus instructors use technology to
teach approximation in first-year calculus courses. The research is a part of a larger study designed to investigate
calculus instructors’ perceptions of approximation as a central concept and possible unifying thread of the first-year
calculus. A survey was administered to 279 calculus instructors who were asked to share information about their use
of technology to promote student understanding of approximation in the first-year calculus. Qualitative and categorical
measures were employed to extract themes and patterns from the data. Four major themes emerged which focused on:
(a) the types of technology currently being used by college and university calculus instructors to teach approximation
concepts, (b) coverage of specific topics using the same devices, (c) the frequency with which calculus instructors
report using technology, and (d) calculus instructors’ perceptions of the role of technology in teaching and learning
approximation ideas. The findings of this research are of significance to all instructors of first-year calculus courses
and have the potential to impact their instructional practices.
190. Online Professional Learning Networks
Justin Gomez Montana Tech of the University of Montana
Kaitlin Rudy Montana Tech of the University of Montana
Advisor(s): Hilary Risser, Montana Tech of the University of Montana
As technological advancements are made in the social media world, more people are connecting for professional development this way. This study served as an update to a 2011 study conducted by Dr. Hilary Risser. The previous
study established a base network of teachers that used blogs to communicate educational practices. The purpose of
this investigation is to analyze the differences, similarities, and benefits of online versus face-to-face communication.
Interviews with multiple math and science teachers were conducted first with an online survey, and followed up via
Skype. Their blogs were examined to identify connections between these teachers so that a new network of communication could be established. Preliminary results show that since 2011, networks have grown. Moving forward, the
contents of each blog will be assessed. One future goal is that the conclusion of this study could lead to better equipped
online social media for education professionals to grow.
191. Calculus: Conceptual or Procedural?
Brandon Finney Texas Tech University
Advisor(s): Gary Harris, Texas Tech University
The literature is rich with articles addressing and assessing the effects of conceptual versus procedural approaches
to teaching and learning in Calculus. We are interested in just what these terms mean in practice and how to gauge
the extent to which the teaching and expected student outcomes in our calculus classes can be viewed as conceptual,
procedural, or some combination of both. To this end, using definitions motivated by the early work of Hiebert and
Lefevre (1986), we have created a Conceptual/Procedural rubric (CP rubric) designed to assess the extent to which our
Calculus exams are entirely conceptual (C), primarily conceptual with procedural knowledge needed (CP), primarily
procedural with conceptual knowledge needed (PC), or entirely procedural (P). In this poster we present the data from
our use of the CP rubric applied to the final exams given in our Calculus I and II classes over multiple semesters.
54
Titles, Authors, Advisors and Abstracts
192. Coffee Wars: Bringing Data About Real World Competition into the Math Classroom
Mary Long Shippensburg University
Advisor(s): Ben Galluzzo, Shippensburg University
The recent addition of a Dunkin’ Donuts caf´e to the Shippensburg University campus was welcomed by many students,
faculty, and staff. However, behind the scenes personnel wonder what the effect of this added competition will have
on the Starbucks caf´e, which has been operating on the campus for seven years. For example, questions such as: “are
more individuals purchasing coffee and/or food from the university?” , and “will Starbucks on campus lose business?”
are being asked. With recent implementation of Common Core Standards in many states, mathematics teachers are
under pressure to incorporate real world connections into course curriculum. This study, a data driven exploration on
the effect of competition between two coffee “giants” on campus, provides background for numerous applications
of mathematics appropriate for students in algebra and introductory statistics courses and addresses questions that
matter. This poster presentation will cover the data collection process, some preliminary results, and sample lesson
plans developed from this study.
193. Mobile Devices as a Component of a Student's Response System for STEM Majors
Shontrice Garrett Jackson State University
Todderrick Robinson Jackson State University
Advisor(s): Jana Talley, Jackson State University
Mobile devices have become increasingly more accessible to students and instructors at the post-secondary level. With
the implementation of Jackson State University’s iPad Initiative, all first time freshmen now have access to iPads.
This study is designed to investigate how iPads and other mobile devices can be used in undergraduate mathematics
courses to engage students in mathematical dialogue, which has been shown to facilitate conceptual understanding.
Specifically, the mobile devices will serve as components of a student response system to engage students in meaningful classroom discussions. The investigation will be conducted at Jackson State University, an urban minority-serving
university and the participants will be the instructors of and the students enrolled in a freshman level mathematics
course.
194. The Impact of Leading Supplemental Instruction Sessions on the SI Leader
Christina Tran California State University, Fullerton
Kelly Hartmann CSU-Fullerton
Advisor(s): Todd CadwalladerOlsker, California State University, Fullerton
The impact of Supplemental Instruction (SI) has been shown to be positive for students taking gateway STEM courses;
we focus instead on the impact of leading SI workshops on the SI leaders. While the impact of SI on the leaders themselves has long been thought to be positive, evidence for this has been largely anecdotal. This study conducted at
California State University, Fullerton attempted to examine this impact with more quantifiable evidence and on a
larger scale than past studies may have reported. Variables included gender, first generation college status, and underrepresented minority status (URM). While the subgroups were not significantly different on variables linked to
academic background and achievement such as grade point average, several significant differences emerged. For example, results showed that a higher proportion of men than women reported increased confidence and effectiveness
in communicating with professors, peers, and students. Additionally, URM leaders reported increases dealing effectively with student conflict and communicating with peers than did their non-URM counterparts. This information may
help inform how to improve the structure of supplemental instruction programs and training to better benefit both SI
students as well as the SI leaders.
195. Primality in Numerical Monoids Generated by an Arithmetic Sequence
Jonathon Spaw Millikin University
Ashley Mailloux Transylvania University
Meghan Malachi Providence College
Advisor(s): Roberto Pelayo, University of Hawaii: Hilo
In cancellative, commutative, atomic monoids, the !-primality function measures how far an element is from being
prime. Since irreducible elements which are not prime give rise to non-unique factorizations, understanding the primality of individual elements gives us a deeper understanding of the factorization theory of these monoids. In the
Titles, Authors, Advisors and Abstracts
55
context of numerical monoids generated by an arithmetic sequence with three generators, we present a thorough analysis of the !-primality function. In particular, we provide a conjectured closed form for the !-value of the generators,
a description of the dissonance points, and closed forms for the quasi-linear portion of the function.
196. Lattice Embeddings of Planar Point Sets
Dara Zirlin Mount Holyoke College
Michael Knopf University of California, Berkeley
Jesse Milzman Georgetown University
Dantong Zhu Lafayette College
Advisor(s): Derek Smith, Lafayette College
In the Euclidean plane, let S be a set of points whose pairwise distances are integers. If the area of each triangle with
vertices in S is also an integer, it is not hard to find a congruent copy of S that embeds in Q2 . It is more surprising
that S also embeds in Z2 . What can be said if the area of some triangle is not an integer? Then S certainly will not
embed in Z2 , but there are other natural lattices of points to consider, namely those corresponding to the maximal
orders of quadratic extensions of Q. We provide a number-theoretical condition sufficient for S to embed in one of
these p
lattices, and we conjecture that the condition is sufficient as well. It is then natural to ask whether S embeds
in ZŒ
d . The equilateral triangle with side length 1 provides a counterexample
for d D 3; but the triangle does
p
embed in the maximal order ZŒ! of Eisenstein integers, where ! D .1 C
3/=2. Our main resultpdetermines the
values of d for which all S with characteristic d embed in the maximal order of the quadratic field Q.
d /. We also
provide similar results for point sets whose pairwise distances need only be square roots of integers.
197. Traces of singular values of Hauptmoduln
Lea Beneish Indiana University
Advisor(s): Ken Ono, Emory University
In an important paper, Zagier proved that certain half-integral weight modular forms are generating functions for
traces of polynomials in the j -function. It turns out that Zagier’s work makes it possible to algorithmically compute
Hilbert class polynomials using a canonical family of modular forms of weight 23 . We generalize these results and
consider Haupmoduln for levels 1; 2; 3; 5; 7; and 13. We show that traces of singular values of polynomials in Haupmoduln are again described by coefficients of half-integral weight modular forms. This realization makes it possible
to algorithmically compute class polynomials.
198. Design of Knapsack Cryptosystems Using Certain t-Superincreasing Sequences
Rob Rexler Baello Montclair State University
Donald Coleman New Jersey Institute of Technology
Advisor(s): Aihua Li, Montclair State University
Superincreasing sequences have been used widely in designing Knapsack cryptosystems. Generalizing the concept,
we define a new type of sequence, t-superincreasing sequence. In this presentation, we report our result on designing
knapsack cryptosystems using certain t-superincreasing sequences. Methods of creating t-superincreasing sequences
and how to use them to construct knapsack cryptosystems are provided. This research is funded by MAA NREUP
through NSF grants DMS-1156582 and DMS-1359016.
199. Metacommutation of the Hurwitz Quaternions and the Projective Line over F p
Shakthi Shrima Boston University - PROMYS
Adam Forsyth Boston University - PROMYS
Jack Gurev Boston University - PROMYS
Advisor(s): Raffael Singer, Oxford University
In 2013, Cohn and Kumar investigated the metacommutation mapping for pairs of primes in the ring of Hurwitz
integers. Specifically, Cohn and Kumar showed that the sign of the permutation of Hurwitz primes of norm p induced
by the Hurwitz primes of norm q under metacommutation is given by the quadratic character of q modulo p, where
p and q are rational primes. Here, we show that this permutation is equivalent to the permutation induced by the right
standard action of PGL2 .Fp / on P 1 .Fp /. Using our equivalence, we give simple proofs of the results of Cohn and
Kumar and characterize the cycle structure of the aforementioned permutation.
56
Titles, Authors, Advisors and Abstracts
200. The Diophantine Frobenius problem over fields.
Yan Sheng Emory University
Advisor(s): Ricardo Conceicao, Oxford College of Emory University
It is a well-known result in number theory that given a collection of positive co-prime integers fa1 ; : : : ; an g, we can
express any integer k as
x1 a1 C C xn an D k;
with x1 ; : : : ; xn integers. The Diophantine Frobenius problem asks for the largest positive integer not expressible in
the above form, where x1; : : : xn are non-negative. Recently, this problem has been extended to polynomials over an
arbitrary field by R. Conceicao, R. Gondim and M. Rodriguez. In our poster, we provide different proofs of certain
results in the aforementioned work. In addition, we reformulate some theorems related to the classical problem in this
new context for polynomials.
201. Density of Primes Dividing Terms in the Somos-5 Sequence
Bryant Davis Wake Forest University
Rebecca Kotsonis Wake Forest University
Advisor(s): Jeremy Rouse, Wake Forest University
The Somos-5 sequence is a recursive sequence defined by an D
an
1 an 4 Can 2 an 3
an 5
with a0 D a1 D a2 D a3 D
1 .x/
a4 D 1, with the interesting quality that all values of the sequence are integers. We wish to determine limx!C1 .x/
where .x/ is the number of primes less than or equal to x and 1 .x/ is the number of primes less than or equal to
x that divide a term in the Somos-5 sequence. To do this, we study the relationship between the Somos-5 sequence
and the elliptic curve E W y 2 C xy D x 3 C x 2 2x. There is an isogeny between the elliptic curve E 0 W y 2 C xy D
x 3 C x 2 C 8x C 10 and E, and the basis for calculating 1 .x/ comes from determining the relationship between the
point of order 2 on E and the point of order 2 on E 0 in Fp . To determine this relationship, we make extensive use of
Galois Theory, Algebraic Number Theory, and the Chebotarev Density Theorem.
202. Shopping For Nu Hats: A Lower Bound for OfZ n ; m; 2g
Alice Mitnick Gettysburg College
Advisor(s): Bela Bajnok, Gettysburg College
In our research, we explored the maximum value of restricted sumsets with a fixed number of terms added. In particular, given a subset A Zn consisting of m elements, one can ask how many distinct elements can be written as the
sum of two elements of A. As A ranges over all subsets of Zn with m elements, we define OfZn ; m; 2g to be the largest
number of distinct elements that can be generated in this manner. Here, we find an explicit formula for a lower bound
of OfZn ; m; 2g.
203. Representations by Ternary Quadratic Forms
Edna Jones Rose-Hulman Institute of Technology
Advisor(s): Matthew Young, Texas A&M University
How can you represent integers by ternary quadratic forms? For example, can the integer 2015 be represented as a
sum of a square plus three times a square plus five times a square? A few kinds of representations over the integers
(such as global representation and local representation) will be discussed. To better understand these representations,
we count how many solutions there are to equivalences involving ternary quadratic forms using quadratic Gauss sums
and Hensel’s Lemma.
204. On the Distribution of Discriminants over a Finite Field
Soonho Kwon Princeton University
Jonathan Chan Bergen County Academies
Advisor(s): Keith Conrad, University of Connecticut
For an odd prime power q, we prove a sufficient condition for the equal distribution of discriminants of monic polynomials in Fq Œx with a given degree. We also prove an analogous result for even q. In addition, we show for each degree
57
Titles, Authors, Advisors and Abstracts
m greater than or equal to the characteristic of Fq that every number in Fq is the discriminant of a monic polynomial
in Fq Œx of degree m.
205. Exploring the Calkin-Wilf Tree: Subtrees and the Births of Numbers
Kayla Javier Bridgewater State University
Advisor(s): Shannon Lockard, Bridgewater State University
Over a hundred years ago Georg Cantor proved that the set of rational numbers is countably infinite. In 1999, Neil
Calkin and Herbert Wilf gave an alternate proof of this fact using a fraction tree which became known as the CalkinWilf tree. The tree presents a way to list the rational numbers so that every fraction is listed exactly once in the
tree, creating a bijection from the set of rational numbers to the set of natural numbers. Since then, mathematicians
have been examining this tree closely, looking for interesting patterns and properties. For this project, we investigated
subtrees and the birth, or the first occurrence, of numbers within the tree. We used Maple to generate the first 1,000
births and looked for similarities within these first occurences. While investigating the births, we found a number
of patterns that give us insight as to where births occur within the tree. We also studied subtrees, which are smaller
fraction trees within the Calkin-Wilf tree. While examining subtrees, we determined which numbers do and do not
occur in a subtree. In the future, we would like to use this information about subtrees to help us determine whether the
birth of a number can be found in a given subtree.
206. Combinatorial Properties of Rogers-Ramanujan-Type Identities Arising from Hall-Littlewood
Polynomials
Claire Frechette Brown University
Madeline Locus University of Georgia
Advisor(s): Ken Ono, Emory University
Here we consider the q-series coming from the Hall-Littlewood polynomials,
R .a; bI q/ D
X
1 a
q cjj P2 1; q; q 2; : : : I q 2bCd :
These series were defined by Griffin, Ono, and Warnaar in their work on the framework of the Rogers-Ramanujan
identities. We devise a recursive method for computing the coefficients of these series when they arise within the
Rogers-Ramanujan framework. Furthermore, we study the congruence properties of certain quotients and products of
these series, generalizing the famous Ramanujan congruence
p.5n C 4/ 0 .mod 5/:
207. Exact p-divisibility of exponential sums of polynomials over finite fields
Daniel Ramirez University of Puerto Rico, Rio Piedras Campus
Ramon Collazo University of Puerto Rico, Rio Piedras Campus
Julio De la Cruz University of Puerto Rico, Rio Piedras Campus
Advisor(s): Ivelisse Rubio, University of Puerto Rico, Rio Piedras Campus
The p-divisibility of the exponential sum of a polynomial F over a finite field with p elements, p a prime, has several
applications in coding theory and cryptography. The covering method of [1] is an elementary method to compute the
exact p-divisibility of this sum. The results presented in [2] give the exact p-divisibility (or improves its bound) of the
exponential sum of F when the polynomial might have several minimal coverings. We simplify the conditions of [2]
and apply the results to the estimation of the p-divisibility of exponential sums of deformations of F .
1. Francis N. Castro and Ivelisse Rubio. “Exact p-divisibility of exponential sums via the covering method” to
appear in Proc. Amer. Math. Soc.
2. ———. “Construction of systems of polynomial equations with exact p-divisibility via the covering method,”
Journal of Algebra and Its Applications, 13(06):1450013, 2014.
58
Titles, Authors, Advisors and Abstracts
208. Special Values in Leamer Monoids
Nicholas Rojina University of North Carolina, Chapel Hill
Jackson Rebrovich Angelo State University
Crystal Mackey Youngstown State University
Advisor(s): Brian Loft, Sam Houston State University
Much is known about the factorization theory for arithmetical Leamer monoids which are generated by an arithmetic
sequence where the step size and the distance between the generators are equivalent. However, much is still unknown
about Leamer monoids when the step size does not equal the distance between the generators. In this paper, we focus
on the values of x0 and xf in relation to the step size of the Leamer Monoid for embedding dimension 2 numerical
monoids. In particular, we present a bound on the step size for which x0 D xf and present a method to compute
x0 in this case. We also present a closed form for the value x0 with any step size. We apply our results to prove the
Huneke-Wiegand conjecture in the case of an arithmetically generated Leamer monoid with embedding dimension 2.
209. Catalan Numbers Modulo 2˛
Gianluca Pane Brown University
Advisor(s): Victor Moll, Tulane University
Catalan numbers, defined by the explicit formula Cn D 1=1 C n 2n
, have been studied since the eighteenth century
n
due to their frequent appearance in various fields from set theory to combinatorics. For example, Cn counts the number
of permutations of f1; 2; : : : ; ng that avoid a three-term increasing subsequence. However, there are few results about
the properties of Catalan numbers modulo prime powers. In particular, we examine the number of residues obtained
by viewing Catalan numbers modulo powers of 2.
210. Improving the Speed and Accuracy of the Miller-Rabin Primality Test
Shyam Narayanan MIT PRIMES-USA, Blue Valley West High School
Advisor(s): David Corwin, MIT
Currently, even the fastest deterministic primality tests run slowly, with the Agrawal-Kayal-Saxena (AKS) Primality
6
Q
Test runtime O.log
.n//, and probabilistic primality tests are still highly inaccurate. In this paper, we discuss the
accuracy of the Miller-Rabin Primality Test and the number of nonwitnesses for a composite odd integer n. We also
extend the Miller-Rabin Theorem by determining when the number of nonwitnesses N.n/ equals '.n/
and by proving
4
5
that for all n, if N.n/ 32 '.n/ then n must be of one of these 3 forms: n D .2x C 1/.4x C 1/, where x is an integer,
n D .2x C 1/.6x C 1/, where x is an integer, n is a Carmichael number of the form pqr , where p, q, r are distinct
primes congruent to 3 (mod 4). Finally, we find witnesses to certain forms of composite numbers with high rates of
nonwitnesses and find that quadratic nonresidues and 2 are both valuable bases for the Miller-Rabin test. This work is
expected to result in a faster and better primality test for large integers.
211. Exact 2-Divisibility of Exponential Sums Associated to Elementary Symmetric Boolean Functions
Oscar Gonzalez University of Puerto Rico at Rio Piedras
Raul Negron University of Puerto Rico at Rio Piedras
Advisor(s): Francis N. Castro, Luis A. Medina, Ivelisse Rubio, University of Puerto Rico at R
An n-variable Boolean function F is a function
defined overˇF2n with values in F2 , the finite field with two elements.
ˇ˚
n
ˇ
The Boolean function F is balanced if x 2 F2 jf .x/ D 1 ˇ D 2n 1 . This property is important for many applicaP
tions in cryptography. A function F is balanced if and only if the exponential sum S.F / D x1 ;:::;xn 2F . 1/F .x1 ;:::;xn /
associated to F is zero. A Boolean function F is called symmetric if its value is not affected by a permutation of its
input. On 2008, Cusick-Li and Stˇanicˇa proposed a conjecture about the non-balancedness of elementary symmetric
Boolean functions. In this work, we calculate the exact 2-divisibility of S.F / for some families of elementary symmetric Boolean functions and use it to prove some cases of this conjecture. Our approach is completely elementary.
212. Covering Everything: An Exploration of h-Critical Numbers
Kevin Campbell Gettysburg Colege
Advisor(s): Bela Bajnok, Gettysburg College
We define an h-fold sumset of A as the set that is given by the sum of h not necessarily distinct values in a set A,
where each value is taken modulo n where n is the size of the ambient group. Given an h; m; n 2 N and the group
Titles, Authors, Advisors and Abstracts
59
Zn ; .Zn ; m; h/ denotes the smallest possible size of an h-fold sumset of A, where A is a subset of Zn of size m. An
h-critical number is the minimum value of m such that .Zn ; m; h/ D n. That is, if a subset A of Zn is at least the size
of the h-critical number of Zn , then the h-fold sumset of A will span Zn . We present bounds for all h-critical numbers
and an explicit equation for any h-critical number for all values of h where the ambient group has a size that is an even
number or a power of a prime. We also conjecture an equation for a general value of n.
213. On Numerical Semigroups Generated by Geometric Progressions
Claire Spychalla Taylor University
Advisor(s): Vadim Ponomarenko, San Diego State University
Let n1 ; n2 ; : : : ; nk 2 N such that gcd.n1 ; n2 ; : : : ; nk / D 1. Then the set S D fc1 n1 C c2n2 C C ck nk W ci 2 Ng is a
numerical semigroup, and we say S is generated by hn1 ; n2 ; : : : ; nk i. Certain invariants of numerical semigroups, such
as the delta set, elasticity, and catenary degree, give a distance measure between elements of the semigroup. Here we
explore the invariants of numerical semigroups that are generated by finite geometric progressions. These semigroups
have surprisingly elegant structure that give insight into the structure of all numerical semigroups and allow us to
determine the delta set, elasticity, and catenary degree of many other semigroups.
214. P-adic Limits of Combinatorial Sequences
Alexandra Michel Mills College
Joseph Rennie Reed College
Advisor(s): Victor Moll, Tulane University
The p-adic field is an alternative to the real numbers as a completion of the rational numbers under a different metric
for every prime p. The definition of a p-adically Cauchy convergent sequence can be equivalently stated in terms of
congruences of the terms modulo arbitrarily large powers of p: a sequence is p-adically Cauchy convergent if and only
if it is eventually constant modulo p k for all k 2 N. Using results on factorials modulo powers of primes by Lin and
Yeh (2010) and Granville (1997), we identify a class of p-adically convergent subsequences of Catalan numbers in
terms of the p-ary expansion of the sequence elements. We then discuss the extension of our results to a more general
class of sequences involving factorials and binomial coefficients.
215. Mirror Symmetry in Reflexive Polytopes
Christopher Magyar University of Wisconsin Eau Claire
Advisor(s): Ursula Whitcher, University of Wisconsin Eau Claire
The polar duality transformation takes a polytope with integer lattice points to its polar dual. If the polar dual is
also a lattice polytope, then we refer to the polytopes as reflexive polytopes. Dual varieties defined from pairs of
reflexive polytopes exhibit the phenomenon of mirror symmetry predicted by string theorists. We use concepts from
algebraic geometry, number theory, and combinatorics to examine one-parameter families of elliptic curves obtained
from reflexive polygons. Previously, we have counted points of dual hypersurfaces within these elliptic curves over
finite fields to demonstrate an arithmetic mirror symmetric relationship holds for three pairs of elliptic curve families.
We link these results to the Picard-Fuchs equations, which can be used to describe how these families change as their
parameter is changed.
216. Sequences of p-adic valuations of polynomials: an analysis of aperiodic and non p-regular
behavior
Amber Yuan The University of Chicago
Alyssa Brynes Tulane University
Advisor(s): Victor Moll, Tulane University
In the field of number theory, the p-adic valuation is a useful device in studying the divisibility of an integer by
powers of a given prime p. This paper centers on 2-adic valuations of quadratic polynomials in ZŒx. In particular,
the existence and properties of roots of such polynomials modulo 2l , are determined and assessed. Polynomials of
particular interest are those that yield non 2-regular sequences in Q2 . Such sequences are represented in a novel
infinite tree form, and patterns in such trees are analyzed to classify the sequences by their structure and non 2-regular
properties. Such classification is further refined through an algebraic analysis of the polynomials at hand.
60
Titles, Authors, Advisors and Abstracts
217. Subtraction Makes Little Difference: The Minimum Sized h-fold Span of m-sized Subsets of
Abelian Groups
Ryan Matzke Gettysburg College
Advisor(s): B´ela Bajnok, Gettysburg College
Let G be a finite abelian group written with additive notation. For a positive integer h and a nonempty subset A of G,
we let hA and h˙ A denote the h-fold unrestricted sumset of A and the h-fold unrestricted span of A, respectively; that
is, hA is the collection of sums of h not-necessarily-distinct elements of A, and h˙ A consists of all signed sums of h
not-necessarily-distinct elements, meaning you can add or subtract each element rather than only add. For a positive
m jGj, we let
.G; m; h/ D minfjhAj W A G; jAj D mg
and
˙ .G; m; h/ D minfjh˙ Aj W A G; jAj D mg:
While one might intuitively think that .G; m; h/ is almost always smaller than ˙ .G; m; h/, we find that for almost
every abelian group G, and positive integers h and m, .G; m; h/ D ˙ .G; m; h/.
218. Newman's conjecture for function field L-functions
David Mehrle Carnegie Mellon University
Advisor(s): Steven J Miller, Williams College
De Bruijn and Newman introduced a deformation of the Riemann zeta function .s/, and proved there is a real constant ƒ which encodes the movement of the nontrivial zeros of .s/ under the deformation. The Riemann hypothesis
is equivalent to ƒ 0, but Newman conjectured that ƒ 0. Andrade, Chang, and Miller extended the machinery developed by Newman and Polya to L-functions for function fields. Here we consider a modified Newman’s conjecture:
supf 2F ƒf 0; for F a family of L-functions. We prove this modified Newman’s conjecture for several families of
L-functions. In contrast with previous work, we are able to exhibit specific L-functions for which ƒ D 0, and thereby
prove a stronger statement: max L2F ƒL D 0. Using analytic techniques, we show a certain deformed L-function must
have a double root, which implies ƒ D 0. For a different family, we construct particular elliptic curves with p C 1
points over Fp . By the Weil conjectures, this has either the maximum or minimum possible number of points over
Fp2n . This tells us that the associated L-function satisfies ƒ D 0.
219. Special Elasticity of Numerical Semigroups
Andy Fry Western Oregon University
Advisor(s): Vadim Ponomarenko, San Diego State University
A numerical semigroup is an additively closed subset of the nonnegative integers. Any numerical semigroup is generated by irreducible elements (elements that can’t be written as a sum of two nonzero elements). Each non-irreducible
element can be factored as a sum of irreducible elements, though these factorizations are generally not unique. In factorization theory, invariants are often used to measure the non-uniqueness of factorizations within a given monoid. One
such invariant is special elasticity, which compares the factorization lengths of certain elements. In this project, we investigate the extremal cases of specialized elasticity, and characterize which semigroups with exactly three irreducible
elements have minimal special elasticity.
220. Moduli Space of Heron Triangles
Jason Murphy University of California, Santa Barbara
Advisor(s): Jordan Schettler, University of California, Santa Barbara
We present an invariant, called k 2 , associated to elliptic curves over the complex numbers. Like the well-known j invariant, the k 2 -invariant can be regarded as a function on the complex upper half plane, and, as such, it is modular
with respect to the congruence subgroup €0 .3/. In fact, k 2 generates the function field of the associated modular curve
X0 .3/. Certain rational values of k 2 come from elliptic curves which parameterize Heron triangles with fixed inradius.
Thus, we relate special analytic properties of k 2 and the field theoretic properties of the function field of X0 .3/ to the
geometry of Heron triangles. We present an analogous set of results for an invariant whose rational values come from
a family of elliptic curves which parametrize hyperbolic Heron triangles with fixed inradius.
Titles, Authors, Advisors and Abstracts
61
221. Delta Sets of Numerical Semigroups
Zofia Stanley Stanley Brown University
Advisor(s): Vadim Ponomarenko, San Diego State University
A numerical semigroup S is a subset of the non-negative integers which is closed under addition and has a finite
complement in N. A factorization of an element in S is a linear combination of the generators with non-negative
integer coefficients; the length of this factorization is the sum of the coefficients. The length set of an element is
the set of lengths of all factorizations of this element, and the delta set is the set of differences between consecutive
elements of the length set. We define the delta set of the semigroup to be the union of delta sets of all elements of S .
It is known that semigroups which are generated by arithmetic progressions have delta sets which contain exactly one
element, namely, the step size of the progression. Through the use of mathematical software and algebraic calculations,
we examine semigroups which are generated by subsets of arithmetic progressions and determine when the delta set
contains exactly one element. We establish a lower bound for the number of generators which must be removed from
an arithmetic progression in order to alter the delta set. We also determine when a numerical semigroup with three
generators has a delta set of size one.
222. An Exploration of Subsequences of the Fibonacci Sequence and their Applications to Cryptography
Yi Wang Millburn High School
Advisor(s): Aihua Li, Montclair State University
The subset-sum problem has widespread applications, including but not limited to the design of knapsack cryptosystems. In this paper, I investigate various superincreasing subsequences of the Fibonacci sequence applied to this problem and search for patterns that arise, involving what types of numbers can and cannot have solutions.
223. Counting The Number of Real Roots of Random Polynomials
Joseph Cleveland Sam Houston State University
Jeffrey Dzugan Samford University
Anthony Morse
Taylor Schlenk
Advisor(s): Dhagash Mehta, University of Notre Dame
Though solving high degree univariate random polynomials and determining the precise real roots out of all the roots
are some of the most important problems in mathematics, science and engineering, it has remained a highly challenging
problem in computational mathematics. By combining an efficient implementation of the Lindsey-Fox algorithm that
can compute roots of high degree univariate random polynomials and Smale’s alpha theorem which can certify if a
given numerical root in the quadratic convergence region of a nearby exact solution, we obtain the certified real roots
and certified counting of the number of real roots of the polynomials.
224. The Modified Galerkin Method for the Modified Wave Equation for the Shape of Biconcave Disk
(Bloodcell)
Jill Resh Roger Williams University
Advisor(s): Yajni Warnapala, Roger Williams University
The objective of this work was to find a numerical solution for the Dirichlet problem for the Helmhloltz for a smooth
biconcave disk. The biconcave disk is a shape that is controlled by one parameter. There are some numerical issues in
this type of analysis; any integration method is affected by the wave number, k, because of the oscillatory behavior of
the fundamental solution. I significantly reduced the number of terms in the infinite series needed to modify the original
integral equation and use the Green’s Theorem to solve the problem on the boundary of the surface. One practical value
of this computation can be getting a shape for a part of a space shuttle that one day might go to planet Mars.
225. Applications and mathematical challenges of digital image mosaicking
Erika Ordog Pepperdine University
Hanna Kristensen Pepperdine University
Advisor(s): David Strong, Pepperdine University
We have created an automatic algorithm that creates a mosaic of user-supplied digital images to reproduce a usersupplied single larger digital image. In creating the mosaic of smaller images to reproduce the larger image, the
62
Titles, Authors, Advisors and Abstracts
simplest approach in selecting which image to place in each larger image pixel location is simply to select the one
image of all smaller images whose overall color or grayscale levels best matches that of the larger image at that pixel
location. However, for a higher quality result, other additional considerations and issues arise, including: reshading
of the smaller images to better match the color or grayscale value at each large image pixel location; how to best
reproduce essentially solid (constant-valued) areas in the larger image; creating more smoothness between smaller
images in the mosaic for a more visually pleasing large image; ensuring that all of the smaller user-supplied supplied
images are used at least once in reproducing the larger image, or possibly using all of the smaller images essentially an
equal number of times; and numerical issues such as image and integer types in working in Matlab. Our work thus far
has focused primarily on grayscale images and is the foundation of our ongoing work with color image mosaicking.
226. Boundary value problems for a self-adjoint Caputo nabla fractional difference equation
Kelsey Mitchell Buena Vista University
Lydia DeWolf Union University
Liam Mazurowski Carnegie Mellon University
Tim Rolling University of Nebraska-Lincoln
Dominic Veconi Hamilton College
Advisor(s): Al Peterson, University of Nebraska-Lincoln
In this paper we develop the theory of initial and boundary value problems for a self-adjoint nabla fractional difference
equation containing a Caputo fractional derivative that is given by
rŒp.t C 1/ra
x.t C 1/ C q.t/x.t/ D h.t/;
where 0 1. We begin by giving an introduction to the nabla fractional calculus and then look at a certain type
of initial value problem containing the Caputo fractional derivative. We investigate properties of the specific selfadjoint nabla fractional difference equation given above, where we show existence and uniqueness for both initial and
boundary value problems. We introduce the definition of a Cauchy function which allows us to solve initial value
problems, as well as the definition of a Green’s function that allows us to certain boundary value problems. Finally, we
look at various inequalities regarding the Green’s function for a particular self-adjoint boundary value problem where
p.t/ D 1, q.t/ D 0, and h.t/ D 0.
227. Quantifying Limits on Replication, Death, and Quiescence of Mycobacterium tuberculosis in Mice
Margaret McDaniel University of Tennessee Knoxville
Winode Handagama Maryville College
Advisor(s): Vitaly Ganusov, Shigetoshi Eda, NIMBioS
According to the World Health Organization, an estimated 2.3 billion people are infected with Mycobacterium tuberculosis (Mtb). Current models for Mtb hold that the total bacterial burden approaches a static equilibrium during
chronic infection. A recent study of Mtb-infected mice by Gill et. al used a mathematical model to show that bacterial
replication and death rates do not necessarily remain constant. In our study, we extend this model by investigating the
effects of a time-dependent segregation rate and the inclusion of quiescence for growth rates and bacteria counts. We
address lingering questions related to Mtb pathogenesis, by developing a mechanistic model, to understand population
of macrophages, bacteria, and the host immune response.
228. Economical Extremal Hypergraphs for the Erd˝
os-Selfridge Theorem
Emily Heath Occidental College
Advisor(s): Eric Sundberg, Occidental College
A positional game can be thought of as a generalization of Tic-Tac-Toe played on a hypergraph .V; H /. We study
the Maker-Breaker game in which Maker wins if she occupies all of the vertices in an edge of H ; otherwise Breaker
wins. The Erd˝os-Selfridge Theorem, a significant result in positional game theory, gives criteria for the existence of
an explicit winning strategy for Breaker for the game played on H . The bound in this theorem has been shown to be
tight, as there are several examples of extremal hypergraphs for this theorem. We focus on the n-uniform extremal
hypergraphs on which Maker has an economical (n-turn) winning strategy. We identify two equivalent criteria to
classify these economical extremal hypergraphs.
Titles, Authors, Advisors and Abstracts
63
229. Visualizing Dessins D'enfants
Mary Kemp Occidental College
Susan Maslak
Advisor(s): Naiomi Cameron, Lewis and Clark College
In this presentation, we will discuss the results of a summer research project motivated by the theory of dessins. Dessin
is short for dessin d’enfant which means child’s drawing. Mathematically speaking, a dessin is a connected bicolored
graph where the edges around every vertex are cyclically ordered. Dessins can be realized by Belyi maps which are
meromorphic functions f W X ! P 1 .C/ such that X is a Riemann surface and f is unramified outside f0; 1; 1g:
One of the goals of this project is to determine Belyi maps that realize a given loopless, connected bipartite graph
on a compact Riemann surface X. We report on our considerations of certain classes of such graphs, explorations of
computational methods for finding associated Belyi maps and related applications.
230. Pattern Avoidance in Double Lists
Frehiwet Negassi Saint Joseph’s College
Advisor(s): Lara Pudwell, Valparaiso University
Originally motivated by sorting algorithm discrepancies in computer science, pattern avoidance is a rapidly-growing
area of combinatorics. In particular we study pattern avoidance in restricted forms of a word, i.e. a sequence of positive
integers. The poster will offer an introduction to the topic of pattern avoidance in double lists. Results for enumeration
of double lists of length n which avoid nontrivial patterns up to length 4 are presented.
231. Mixed Smiles: An Analysis of the Coherence Between Experiential and Behavioral Response
During Ambiguous Emotional Stimuli
Marina Massaro State University of New York College at Geneseo
Advisor(s): Charles Collins, University of Tennessee, Knoxville
In 1971, Paul Ekman published that the six basic emotions (anger, happiness, surprise, disgust, sadness, and fear)
have distinct, universal facial expressions. Past studies confirm that high levels of coherence exist between emotional
experience and facial expression when subjects are shown poignant film clips. But what happens to the tie between
emotional experience and facial response when the eliciting event isn’t purely of one sentiment? In this study, data
collected from video recordings of individuals watching ambivalent film clips are used to examine the relationship
between emotional experience and facial response when mixed emotion is elicited. Cross-correlations between data
sources measuring positive and negative emotional response indicate low to moderate amounts of coherence between
response systems. Contributors to coherence are being investigated, including the degree to which emotion is “mixed”
at a given time, individual personality characteristics, and attitude toward various emotions. Furthering the understanding of response coherence and mixed feelings will shed light on various psychiatric disorders, where emotional
dysfunction skews the normal emotional experience and response.
232. Coloring Techniques for Pattern Avoidance over an Infinite Sequence
Paula Egging Benedictine College
Robert Laudone Boston College
Amanda Owens McDaniel College
Advisor(s): David Milan, University of Texas at Tyler
We investigate Grytzcuk’s conjecture and extend the entropy compression method to prove that long-square-free sequences (length 6) can be chosen from lists of size 3. Within this proof we utilize and independently discover a
bijection between plane trees and difference sequences (sequences of integers generated when running an algorithm).
We also investigate  Sets, utilizing new strategies to surpass the results of entropy compression. We then expand
known work on shuffle squares to apply to shuffle long-squares and provide a general formula for further expansion.
233. Elliptic Curve Cryptography
Samantha Sprague Marist College
Advisor(s): Joseph Kirtland, Marist College
An elliptic curve E is the set of solutions to an equation of the form y 2 D x 3 C Ax C B where the constants A and
B satisfy 4A3 C 27B 2 ¤ 0. The collection of points on E, with coordinates in Fp , along with the point at infinity O,
64
Titles, Authors, Advisors and Abstracts
form the group E Fp which is of interest in cryptography. This poster will present how E Fp can be used to create
a secure public-key cryptosystem.
234. Inverses and Invariant Subspaces for a Certain Non--self-adjoint n n Toeplitz
Zach Walter Butler University
Advisor(s): William Johnston, Rebecca Wahl, Butler University
For any size n, when Œar c  is c r 1, for c r , and c r C 1 for c r , each corresponding n n non–selfadjoint Toeplitz matrix turns out to have a remarkably simple inverse with only six distinct nonzero entries. This
result is new, and it points toward fruitful investigation into a broader category of non–self-adjoint Toeplitz matrices,
which this poster will also describe. The inverse also exhibits a pattern, no matter what size n. This poster explicitly
describes each inverse, giving a formula for each of the inverses’ entries, no matter what size n. It also explains why
the calculation of this inverse is not trivial. Finally, it exhibits all of the invariant subspaces for these matrices, no
matter what size n, in terms of the corresponding lattice diagram.
235. Explicit Bounds for Matrix Pseudospectra
Jeremy Meza Carnegie Mellon University
Abigail Ward The University of Chicago
Olivia Meyerson Williams College
Advisor(s): Mihai Stoiciu, Williams College
We study the -pseudospectra .A/ of square matrices A 2 C N N . We give a complete characterization of the pseudospectrum of any 2 2 matrix and describe the asymptotic behavior (as ! 0) of .A/ for any square matrix
A. We also present explicit upper and lower bounds for the -pseudospectra of bidiagonal and tridiagonal matrices, as
well as for finite rank operators.
236. Pattern Avoidance in Forests
Peter Nugent University of Wisconsin Eau Claire
ML Tlachac University of Wisconsin-Eau Claire
Derek Levin University of Wisconsin-Eau Claire
Advisor(s): Manda Riehl, University of Wisconsin Eau Claire
Pattern avoidance in permutations is a growing research specialty, and we widen its application to forests. An increasing forest is a collection of t trees where each of the trees has v labels, and every path from root to leaf is an increasing
sequence of labels. These forests contain associated permutations which are found by conducting a depth first search
on each tree from left to right, starting at the root and moving left to right by level. We enumerate pattern avoidance
classes within the permutations associated to these forests and create bijections to other combinatorial objects. We
have a bijection from forests avoiding the pattern 123 to the set of paths loosely under the line y D vx from .0; 0/
to .t; tv t/ using vertical steps .0; 1/ and horizontal steps .1; 0/. Additionally, the number of forests of n increasing
2-node trees avoiding 321 is the same as the number of 321-avoiding ordered set partitions of f1; : : : ; 2ng where
all
blocks are of size 2. We also found that the number of 321 avoiding binary forests with 2 trees is given by 2n
.
Lastly,
n
in unary forests where each tree has 2 nodes, there is a bijection between forests avoiding each of the patterns 132,
213, 312, and 231.
237. Diagrammatics of Coxeter and Braid Groups
Uma Roy MIT PRIMES
Niket Gowravaram MIT PRIMES
Advisor(s): Alisa Knizel, Massachusetts Institute of Technology
Elias, Khovanov and Williamson recently developed a diagrammatic language to attack many problems of current
interest in representation theory. We examine using diagrammatics, a problem related to the K.; 1/ conjecture for
braid groups, which we refer to as the K.; 1/ conjecturette, and an analogous problem for Coxeter groups. Given
a Coxeter or braid group, one can form a family of colored planar graphs based on the generators and relations of
the group, which we refer to as diagrams. Two diagrams are considered homotopic if one can be transformed into
the other through a series of allowed transformations. For braid groups, proving that any diagram is homotopic to the
65
Titles, Authors, Advisors and Abstracts
empty graph is equivalent showing that 2 of the associated Salvetti complex is trivial, which is a part of the K.; 1/
conjecture. For Coxeter groups, proving that any diagram is homotopic to the empty diagram is equivalent to showing
that 2 of a cell complex introduced by Elias and Williamson is trivial. We present a diagrammatic proof of the K.; 1/
conjecturette for a family of braid groups as well as analogous results for several families of Coxeter groups.
238. SmarTrAC: Activity Detection with Smartphone Sensor Data
Bryan Martin Macalester College
Anya Phillips Macalester College
Mitchell Kinney University of Iowa
Heather Berlin Washington University in St. Louis
Lindsey Schadler Shippensberg University
Advisor(s): Vittorio Addona, Macalester College
As smartphones become more prevalent in everyday society, their technology can be harnessed for new purposes. Our
research develops a model for SmarTrAC, an android application that uses smartphone GPS and accelerometer data
to more fully understand how the urban population travels. Our problem is activity detection; we want the application
to determine, in real time, which mode of transportation is currently being used without the smartphone user using
manual inputs. In order to accomplish this, we collected training data and analyzed it for descriptive features. We then
utilize machine learning modeling techniques, focusing on decision trees and random forests. Our final models achieve
accuracies ranging from approximately 90% to 94%.
239. Tauer of Terror: Not as Scary as One Might Think
Elizabeth Matys Gettysburg College
Advisor(s): B´ela Bajnok, Gettysburg College
Given a subset A in a finite abelian group G and a positive integer h, A is zero-h-sum-free if the signed sum of any h
(not necessarily distinct) elements in A is never equal to zero. Working in Zn , we explore the maximum zero-h-sumfree set size over the set of all integers for odd h. That is, the quantity defined as
˙ .Zn ; h/ D maxfjAj j A Zn ; 0 62 h˙ Ag
We establish the value of ˙ .Zn ; h/ for all even n and construct two lower bounds where n is odd. We also determine
the value for some cases of prime n.
240. On the Minimum Size of Spanning Sets in Cyclic Groups
Tyler Reckner Gettysburg College
Advisor(s): Bela, Bajnok
A subset A of a finite Abelian group G is said to span G if the h-fold span, h˙ A D f1 a1 C 2 a2 C C m am W
j1 j C j2 j C C jm j D hg, is equal to G. Here, we are interested in the minimum size of such a spanning set for
cyclic groups Zn . This quantity is denoted ˙ .Zn ; h/. We discuss several conditions on positive integers h and m for
which ˙ .Zn ; h/ D m for small values of m, offering both foundational propositions and new results.
241. Uptake, Translocation, and Stress Effects of Carbon Nanotubes in Drought Induced Corn
Sabrina Deleon Texas Tech University
Advisor(s): Jaclyn Canas-Carrell, Texas Tech University
Carbon nanotubes are one of the most used manufactured nanomaterials. However, these materials are not regulated
and there are concerns regarding their behavior in the environment and human health. This study was conducted to
evaluate uptake of various types of carbon nanotubes in corn under ideal watering and drought conditions. Corn was
exposed to either non-functionalized carbon nanotubes (CNTs) or functionalized carbon nanotubes (COOH-CNTs).
Plants were grown for 21 days in soil with 10 mg/kg of CNTs or COOH-CNTs in 1 L or 3 L of soil in a greenhouse with
natural day:night conditions. Corn was also grown under conditions simulating a seven-day drought and photosynthesis
measurements were taken using a LI-6400XT Portable Photosynthesis System. Following harvest after 28 days, roots,
stems, and leaves were dried, grounded, and analyzed using a microwave-induced heating technique to quantify CNT
and COOH-CNT concentrations in the corn. Photosynthetic rate declined throughout the duration of the drought
66
Titles, Authors, Advisors and Abstracts
treatments. CNT uptake was only detected in roots of drought-treated plants exposed to CNTs and COOH-CNTs.
Additional plant analyses are currently ongoing.
242. Investigating Properties of Cay.Z; f˙2 kg/ and Cay.Z; f˙3kg/
Daniel Gulbrandsen Utah Valley University
Advisor(s): Violeta Vasilevska, Utah Valley University
Richard E. Schwartz [1] posed the following problem: Are the Cayley graphs C2 D Cay.Z; f˙2k g/ and C3 D
Cay.Z; f˙3k g/, for k D 0; 1; 2; : : : , quasi-isometric? Moon Duchin and Bryan White [2] have shown that C2 and
C3 , with the associated word metric, are not quasi-isometric under the identity map. In this presentation we further
investigate this problem. Namely, a particular class of maps will be considered and proved that these maps are not
quasi-isometries between C2 and C3 . In addition, properties known to be quasi-isometric invariants, such as hyperbolicity and metric ends, are studied. It will be shown that these properties cannot be used to determine if C2 and C3 are
quasi-isometric. Further work will also be discussed. Joint work with Daniel Adams and Kristen Smith.
1. Melvyn B. Nathanson, Bi-Lipschitz Equivalent Metrics on Groups, and a Problem in Additive Number Theory. 18
Feb. 2009
2. Moon Duchin, Bryan White. Two presentations of Z that are not Quasi-Isometric. (preprint)
243. An Algorithm for Maya Division
Nicholas Heil Millersville University of Pennsylvania
Advisor(s): Ximena Catepillan, Millersville University of Pennsylvania
The ancient Maya civilization, which began ca. 1200 BC, had a long history divided into three periods. These people
were sophisticated astronomers who developed a positional number system with base twenty along with the concept
of zero. They used this system to build their roads, construct monumental architecture, and carry out astronomical
computations. This vigesimal system used bars and dots to represent numbers and carry out the four basic operations
with whole numbers. There is research on processes for addition, subtraction, and multiplication; however, there is
limited information for how division was performed. In this presentation a new model for division, using repeated
subtraction and multiplication, is explained. The process was made simpler by using a Maya multiplication table. The
spatial aspect of the division algorithm was expanded to include columns for current product, partial working quotient,
and partial final quotient.
244. Graphs, Probability, and Separating Families
Luke Serafin Carnegie Mellon University
Daniel Condon Georgia Institute of Technology
Cody Stockdale Bucknell University
Advisor(s): Samuel Coskey, Boise State University
Suppose we have a herd of cows, one of which has an illness that can be detected by means of a blood test. We want
to determine which cow is sick using as few simultaneous tests as possible; this is an example of a combinatorial
search problem. We can solve our problem by using a separating family to ensure each cow goes through a unique
series of tests. In addition to combinatorial search, separating families and related concepts have applications to digital
fingerprinting, cryptanalysis, coding theory, and related fields. Our project focuses on generalizations of separating
families. In particular, we define the concept of an n-separating family, which simultaneously separates all collections
of n pairs which are in fact separable. To facilitate use of this definition, we give a graph-theoretic characterization
of separability. Finally, we use the nonconstructive probabilistic method to compute bounds on the minimal size of
n-separating families.
245. Can Governments Curb Bitcoin? How Regulation Affects Virtual Currency
Connor Kispert Rose-Hulman Institute of Technology
Advisor(s): Wayne Tarrant, Rose-Hulman Institute of Technology
One of the most notable qualities of Bitcoin is its decentralized nature and disconnect with banking and regulatory
systems. However, recent government bans and regulations on the virtual currency are being linked to the current
downward trend of Bitcoin. Most notably was the drastic falloff in December of 2013 which began immediately after
Titles, Authors, Advisors and Abstracts
67
China introduced new regulations on the currency. In this paper, we look at whether or not government regulations can
have a significant effect on the trend of Bitcoin in terms of both its price and volatility. By analyzing trends surrounding
the event window of the introduction of new regulations on Bitcoin by a national government, we can measure if there
is any significant change in trend for either measure. By answering whether or not a government’s regulations can
have substantial influence over Bitcoin, we can give insight as to whether a decentralized virtual currency could have
sustained success or stability.
246. Solving Rubik's Revenge
Kyle Lewis Coastal Carolina University
Advisor(s): Thomas Hoffman, Coastal Carolina University
In this poster presentation, we find the average number of moves to solve a 4 4 Rubik’s Cube. To solve this puzzle,
a personal algorithm is used in a computer program that will keep track of the number of moves being made. The
program ran ten thousand times for an accurate average. The presentation will also give a brief explanation of the
algorithm and a comparison to the 3 3 algorithm.
247. 3D Printing and Its Uses in Super-Computing Clusters
Cody Stafford Texas Tech University
Advisor(s): John Carrel, Texas Tech University
In this project, the Texas Tech University Department of Mathematics is working closely with the Department of
Industrial Engineering to build, modify, and test a “homemade” 3D printer. This printer will be used to prototype
and assemble custom parts for the Schoenberg high performance computing cluster, these parts will allow for full
customization of the cluster. Also the 3D printer will be used as a promotion tool for STEM education by showing
prospective students how Mathematics and Engineering majors can work together on a project. The poster will cover
the involvement of an undergraduate student in the assembly of the 3D printer and Computer Aided Drafting/Computer
Aided Manufacturing (CAD/CAM) of prototyped parts. In addition, the poster will show how the parts produced will
be used to customize the Schoenberg cluster.
248. Bathtub and Unimodal Hazard Flexibility Classification of Parametric Lifetime Distributions
Dana Lacey North Central College
Anh Nguyen Texas Christian University
Advisor(s): Kahadawala Cooray, Central Michigan University
There are a number of bathtub and unimodal hazard shape parametric lifetime distributions available in literature.
Therefore, it is important to classify these distributions based on their hazard flexibility to facilitate their use in applications. For this purpose we use the Total Time on Test (TTT) transform plot with two different criterion: I. measure
the slope at the inflection point on the scaled TTT transform curve; II. measure the slope at selected points from the
constant hazard line on the scaled TTT transform curve. We confine our research to classify the flexibility of Weibull
extensions and generalizations and also select one-shape parameter lifetime distributions to exemplify the two criterion
process.
249. Predictive Power of a Generalized Preventive Care Segmentation Model
Sirui Wang Cornell University
Ciaran Evans Pomona College
Parker Hund The University of Texas at Austin
Erics Varley St. Olaf College
Advisor(s): Matthew Willyard, The Pennsylvania State University
As the cost of healthcare rises, health insurance providers are encouraging their customers to take preventive actions
against illness. Using data from a health insurance company, we built several models that predict if a customer will
take preventive healthcare actions. Since contacting every customer can be very costly, insurance companies can use
our predictive models to economically target their outreach to only customers who are likely to respond. We find that
Chi-Squared Automatic Interaction Detection is best at predicting whether a customer will take a preventive action,
and Random Forests using Conditional Inference Trees is best at predicting the number of these preventive actions.
68
Titles, Authors, Advisors and Abstracts
250. Too Many Variables too Many Conclusions? Using Mahalanobis-Taguchi Analysis on Health Risk
Factors
Sarah Appelbaum NYU Polytechnic School of Engineering
Advisor(s): Lindsey Van Wagenen, NYU Polytechnic School of Engineering
The Mahalanobis-Taguchi System (MTS) is a binary classification system that incorporates multidimensional statistics
with industrial engineering practices to determine key variables. The Taguchi aspect of MTS allows us to simulate
many different scenarios, maximizing the variety of variable combinations while minimizing computation time. MTS
is particularly useful when there are large numbers of variables (e.g. medical applications) and it works with any type
of distribution, yielding an ordered sequence of variables in terms of their influence on the multidimensional system.
Many medical conditions result from a combination of factors (e.g., genetic, lifestyle) but can result in widely different
prognoses. The genetic variables can differ in expression and thus contribute to different progressions of the condition.
Environmental and lifestyle factors are thought to contribute to the severity of the condition and have become of great
interest to researchers. Infertility is a prime candidate for analysis by MTS. The research conducted applied MTS to
infertility patient data. This has the ability to correlate different lifestyle factors with genetic conditions to isolate the
human factors at the root cause of infertility and thus improve patient treatment.
251. Estimation of Option Prices with Heston's Stochastic Volatility Model
Robin Dunn Kenyon College
Paloma Hauser The College of New Jersey
Tom Seibold Western Kentucky University
Advisor(s): Hugh Gong, Valparaiso University
An option is a security that gives the holder the right to buy or sell an asset at a specified price at a future time. The
Black-Scholes model is the most widely used option pricing model, but it makes the strong assumption of constant
volatility of the asset’s price. This project focused on deriving and testing option pricing formulas for the Heston
model, which treats volatility as a stochastic process. With the assistance of computational software, R and Maple, the
method of moments estimated parameters of the Heston model. Historical option data provided a basis for comparing
the estimated option prices from the Heston and Black-Scholes models. Based on root-mean-square error values, the
Heston model provided more accurate option pricing estimates than the Black-Scholes model.
252. Mixing Times of Markov Chains
Matthew Sutter University of Michigan - Flint
John Rahmani Virginia Tech
Advisor(s): Peter Otto, Willamette University
The mixing time of a convergent Markov chain measures how fast it converges to its stationary distribution. The Rook’s
walk can be represented by a Markov chain that describes the motion of a Rook (chess piece) moving uniformly
randomly on an N N chessboard. The Restricted Rook’s Walk is a variation on the Rook’s Walk, but limiting the
distance the Rook can move in a single step. We apply spectral and coupling methods in order to bound the mixing
times of these two Markov chains, and explore asymptotic properties of these bounds.
253. The Probabilistic Change Point Algorithm
Marcus` Antonellis College of the Holy Cross
Advisor(s): Eric Ruddieri, College of the Holy Cross
While data sets have often been approximated through the usage of linear models, it is often inappropriate to use just a
single line segment, especially when the data is expected to vary over time. Often a piece-wise function of multiple line
segments is the proper solution, however deciding where to break up the model poses its own set of challenges. Possible
solutions grow exponentially with the size of the data set and this makes finding the “right” change point very difficult.
The analysis of these “change points” in a Bayesian setting was the intention of this study. A probabilistic model was
developed for accurately predicting the location of change points. Using Bayes’ rule suitable model parameters could
be found and from that probabilities could be assigned to the suspected change point. These methods were all tested on
temperature data sets provided by the NOAA on departures from long standing temperature averages the years 1880
to 2012. Moreover the probabilistic model that has been developed can be used to search for change points in a wide
variety of data sets.
Titles, Authors, Advisors and Abstracts
69
254. The Stock Price Effect of Apple Keynotes
Ethan Petersen Rose-Hulman Institute of Technology
Advisor(s): Wayne Tarrant, Rose-Hulman Institute of Technology
Apple’s product announcements have been colloquially termed “Mini-IPOs” for their behavioral similarities to other
companies’ IPOs. Exploring this claim, I analyzed the volatility of Apple’s stock beginning January 3, 2005 up to
October 9, 2014, then focused on a range from 30 days prior to each product announcement until 30 days after the
announcement. Product announcements were filtered; only announcements whose 60 day range was devoid of other
events were used. This filtration was chosen to limit the potential cross-effect of a previous or upcoming announcement
on each event. Analyzing the data within these ranges, there was a clear trend prior to Day 0 and significantly different
trend after the event. To explore any confounding effects, a comparison to the VIX was made to determine if the
changes in trends were simply following the market or were truly deviating from a trend. Again, these results showed
significant activity around Day 0. We believe the findings from this study may raise concerns with the Efficient Market
Hypothesis, providing an insight on how strongly this idea relates to activity surrounding Apple keynotes.
255. Topological Data Analysis | The Use of Non-Parametric Statistical Methods to Distinguish
Between the Topologies of Data Sets
Joshua Kiers Taylor University
Christopher Cericola Seattle University
Advisor(s): Inga Johnson, Jordan Purdy, Willamette University
Homology is a well-established area of algebraic topology that can be used to distinguish between topological spaces.
By forcing a parameterized topological structure on discrete data sets, features of homology within the data can be
detected. As shown in a paper by Robinson and Turner, non-parametric statistical methods can then be employed on
groups of data in order to distinguish them topologically in a meaningful way. This method of hypothesis testing is relevant and more insightful than established clustering methods when certain limitations are considered. The extensions
and limitations of the method when applied to a variety of data sets are explored.
256. Chi-Squared Versus Fisher's Exact Test
Lillian Nicholson UNG
Advisor(s): Brad Bailey, University of North Georgia
When looking at Chi-Squared Test and Fisher’s Exact test one may wonder where differences lie. We used a Maple
procedure to look at all possible two-way tables for a pair of dichotomous variables to compute p-values for Fisher’s
exact test and chi-squared test of independence. From there, we sought to find answers involving the proportion for
which the two tests disagree based on samples sizes 4-1,000. Also, we investigated, which of the two test proves to
be more conservative. Through our research we will reveal the secrets to the relationship of the Chi-Squared Test and
Fisher’s Exact Test.
257. Understanding Neuroticism and its Mediation of Optimism
Samantha Hulsey University of North Georgia
Lillian Nicholson University of North Georgia
Advisor(s): Robb Sinn, University of North Georgia
Optimism is most commonly measured using the Life Orientation Test (LOT). To measure neuroticism, we used
two surveys, a Likert scale instrument and a dichotomous scale, to demonstrate the how neuroticism mediates many
correlates of optimism. The goal of this study is to explore the relationship between how neuroticism instruments
measure optimism.
258. Modeling Related Failures of Vehicle Components using Bayesian Copula Selection
John Kumerow University of St. Thomas
Nicole Lenz University of St. Thomas
Katie Wifvat University of St. Thomas
Kelsie Sargent University of St. Thomas
Advisor(s): Arkady Shemyakin, University of St. Thomas
In the automotive industry it is important to know whether the failure of some car parts may be related to the failure of
others. This project studies warranty claims data for five engine components obtained from a major car manufacturer
70
Titles, Authors, Advisors and Abstracts
with the purpose of modelling the joint distributions of failure of two parts. The one-dimensional distributions of
components are combined to construct a bivariate copula model for the joint distribution that enables us to estimate
the probabilities of two components failing before a given time. Ultimately, the influence of the failure of one part on
the operation of another related part can be described, predicted and addressed. The performance of several families
of copula models (Clayton, Gumbel-Hougaard, Student) is analyzed, and Bayesian model selection is performed.
259. On the Probability of Weak Lattice Point Visibility
Rebecca Rachan North Central College
Advisor(s): Neil Nicholson, North Central College
A specifi c rectangular array of lattice points is investigated. We say that a point in the array is weakly visible from
a lattice point not in the array if no point in the array lies on the line connecting the external point to any other point
in the array. A necessary and su fficient condition for determining if a point in the array is weakly viewable by the
external point, as well as the number of points that are weakly visible, is determined. From this, the probability that an
external point can view a randomly chosen point in the array can be calculated.
260. Tug-of-war games with spatial and temporal dependence in noisy environments
Yair Shenfeld Massachusetts Institute of Technology
Advisor(s): Scott Sheffield, Massachusetts Institute of Technology
Zero-sum games called Tug-of-War games are known to give rise to solutions of differential operators such as the ‘infinity Laplacian’ if the movements of the players are deterministic and the p-Laplacian if the movements of the players
are effected by noise. Recent works have investigated the types of PDEs that arise in games where the movements of
the players are deterministic but the sets of possible movements depend on space and time in the domain of the game.
The case where the movements of the players are spatially and temporally dependent and are also effected by noise
which depends on time and space has not be treated in the literature thus far. We address this problem and find the
PDE that arises out of this game. We also provide intuition about the form that the PDE takes and allude to possible
real-world applications that follow from our analysis.
261. Scoring Cardiac Health: A Model of the Relationship between Diet and the Risk of Cardiovascular
Disease
Matthew Miller Southwestern University
Advisor(s): Therese Shelton, Southwestern University
Like many other types of disease, cardiovascular diseases can be treated with a greater success rate the earlier that
they are detected. For this reason medical researchers have long been developing means of detecting cardiovascular
diseases earlier and earlier. Our goal is to develop a “scorecard” to measure cardiac health in the hopes of identifying
cases of serious cardiovascular disease earlier. To this end we compiled the diet and health profiles of patients who
came to an emergency room complaining of chest pains and have begun to study them for relationships between diet
and other factors and the patients’ outcomes. Those who chose to participate were asked to complete a food frequency
questionnaire, and their doctors completed surveys about the health profiles of the patients. From the results of the
questionnaire, we measured the subjects’ diet compositions in terms of five nutrients: protein, carbohydrates, fat, fiber,
and total caloric intake. We will apply regression analysis to measure the relationships between those nutrients and
certain risk factors for cardiovascular disease. We intend for the scorecard to also be used to measure general cardiac
wellness.
262. Knot Depth for Positive Braids
Patricia O'Brien The University of Texas at Austin
Elliot Kaplan Ohio University
Advisor(s): Teena Gerhardt, Michigan State University
Given a projection of a knot, one can alter it at a crossing using what is called the oriented skein relation. This relation
takes this projection, and considers instead the two new knot projections which result by (a) changing which strand in
the crossing is on top, and (b) removing the crossing in a way consistent with the orientation. The depth of a knot is
essentially the minimal number of levels in a resolution tree for that knot using the oriented skein relation. The depth of
Titles, Authors, Advisors and Abstracts
71
a knot is related to the complexity of the knot’s Alexander polynomial and, more generally, any polynomial invariant
which satisfies the skein relation. Every knot can be represented by a word in a group called the braid group. We study
the depth of knots representable by positive braids—a class including all torus knots and links. We have proven that
the depth of a knot represented by a positive braid is ` w C 1, where ` is the length of the braid word and w is the
number of strands in the braid.
263. Quandles of Virtual Knots
Sherilyn Tamagawa Scripps College
Advisor(s): Sam Nelson, Claremont McKenna College
Knot theory is an important branch of mathematics with applications in other branches of science. In this poster, we
explore invariants on a special class of knots, known as virtual knots. We find new invariants by taking quotients
of quandles, and introducing the fundamental Latin Alexander quandle and its Gr¨obner basis. We also demonstrate
examples of computations of these invariants.
264. Bounds for self-intersection numbers of closed curves on the 2-Punctured Plane
Cameron Thieme University of Southern California
Advisor(s): David Crombecque, University of Southern California
The homotopy classes of closed curves on a topological space form a group called the fundamental group. The fundamental group of the 2-punctured plane is free on two generators, where each generator is a simple closed loop around
a puncture. In this project, we address the problem of finding the minimum number of self-intersections of these loops
up to homotopy equivalence. In 2010, M. Chas & A. Phillips approached this same problem and provided a bound
for any element in the group. Although this bound is powerful because of its generality, often it is also very wide. By
constructing some algorithms, we were able to improve these bounds for two types of elements in the fundamental
group.
265. Closing Open Knots with Random Equilateral Polygons
Nicole Lopez University of St. Thomas
Madeline Shogren University of St. Thomas
Advisor(s): Eric J. Rawdon, University of St. Thomas
The recent discovery of knotting behavior in proteins has stimulated discussion about how to classify knots in open
chains. Topology is trapped in closed curves, so it is possible to study the knotting mathematically. Therefore, in order
to study and better understand the knotting patterns in open chains, their endpoints must first be connected to create
closed knots and make it possible to identify and classify the knots embedded within them. The purpose of this project
is to analyze and compare two methods used to classify knots in open chains. The first method draws from a previous
project and closes knots by extending rays from the endpoints out to infinity and then connecting them. The second
method closes equilateral open chains using random equilateral arcs. This process utilizes closure arcs that are 1, 2,
and 3 times the length of the missing edges in the open chain. Varying closure lengths more closely resembles real-life
applications. The two methods are compared by applying each to the same set of open chains and analyzing their
classifications. This project develops efficient processes for all methods and compares them to determine the most
precise method for classifying an open knot.
266. Distinguishing Colored Links
Grant Roth University of Wisconsin-Eau Claire
Advisor(s): Christopher Davis, University of Wisconsin-Eau Claire
A link is an array of circles with various numbers of twists in space which never cross each other. To give a better
picture, one should imagine a collection of strings tangled up with their ends fused together. Classically in link theory,
tools were studied which treat all of these strings the same. The recent idea of coloring the components of links has
spawned a collection of new tools which do distinguish the components. Using these tools we construct examples
of links whose components cannot be interchanged, but which are elementary in the sense that if a single string is
removed then the link comes undone. This work represents an ongoing project.
72
Titles, Authors, Advisors and Abstracts
267. Knot Invariants from Spanning Surfaces for a Two-Bridge Knot
Vincent Longo The College of New Jersey
Advisor(s): Cynthia Curtis, The College of New Jersey
The Alexander polynomial of a knot is one of the most well known and useful knot invariants in the field of Knot
Theory. One way of finding the Alexander polynomial of a knot is by constructing a matrix using the unique orientable
surface bounded by the knot, and taking the determinant of that matrix. We present an extension of this notion for nonorientable surfaces bounded by a knot by defining the State polynomials of a two-bridge knot. We then present some of
the properties shared by the State polynomial and Alexander polynomial, along with some notable differences between
the two.
268. Knot Colorings and Dimer Graphs
Maisie Newman Washington College
Advisor(s): Heather Russell, Washington College
A central question in knot theory is whether two knot diagrams represent the same knot. We can answer this question
with invariants: specific quantities assigned to knots that are the same for isotopic knots. In 1956, Fox defined n-Fox
coloring of knot diagrams as a way to get new invariants. We introduce a new—but related—method to color knot
diagrams called face coloring. One can construct a matrix that encodes a knot’s face coloring data. When a knot is
alternating, there is a way to build a planar bipartite graph from this matrix. Kasteleyn theory provides a framework
for obtaining matrix data from a graph. We apply Kasteleyn theory to our specific class of graphs in order to more
efficiently find data about face coloring and more directly relate this data to a knot’s structure.
269. Defining Virtual Rational Tangles
Ellen Lehet SUNY Potsdam
Advisor(s): Laura Person, SUNY Potsdam
The concept of virtual crossings was first introduced in Kauffman’s paper, “Virtual Knot Theory.” These virtual crossings differ from classical over-under crossings because the two strands involved co-exist in the same space. For this
reason, virtual crossings and virtual knots do not exist in physical world. These virtual structures can be studied along
with the classical structures and often have comparable properties. This project discusses the investigation of virtual
crossings when they are considered in rational tangles. Previously, there has not been very much study of virtual rational tangles and so the goal of this research is to define virtual rational tangles along with identify properties of
virtual tangles. The majority of this project has been inspired by Kauffman’s paper about rational tangles, “On the
classification of rational tangles” and involves drawing comparisons between the classical tangles Kauffman describes
and the idea of virtual rational tangles. These properties include an extension of the continued fraction form, methods
and properties for adding and multiplying tangles, and identifying matrix and fraction invariants.
270. Non-left-orderable surgeries on twisted torus knots
Justin Goluboff Columbia University
Katherine Christianson Columbia University
Srikar Varadaraj Columbia University
Linus Hamann Columbia University
Advisor(s): Jennifer Hom, Columbia University
Boyer, Gordon, and Watson have conjectured that an irreducible rational homology 3-sphere is an L-space if and only
if its fundamental group is not left-orderable. Since large classes of L-spaces can be produced from Dehn surgery on
knots in the 3-sphere, it is natural to ask what conditions on the knot group are sufficient to imply that the quotient associated to Dehn surgery is not left-orderable. Clay and Watson develop a criterion for determining the left-orderability
of this quotient group in and use it to verify the conjecture for surgeries on certain L-space twisted torus knots. We
generalize a recent theorem of Ichihara and Temma to provide another such criterion. We then use this new criterion
to generalize the results of Clay and Watson and to verify the conjecture for a much broader class of L-space twisted
torus knots.
Titles, Authors, Advisors and Abstracts
73
271. A reduced presentation of the virtual singular braid monoid
Andrew De la Pena California State University, Fresno
Sarah McGahan California State University, Fresno
Advisor(s): Carmen Caprau, California State University, Fresno
A braid is a set of n strings passing between two horizontal bars. These strings may interact with one another but must
always travel in the downward direction. If the two horizontal bars are brought together and each pair of string ends
are glued together in order, the resulting structure will be a knot or a link, called the closure of the braid. The focus
of our research has been to study the set of braids whose closure is a virtual singular knot or link. We define the set
of virtual singular braids on n strands as a monoid via generators and relations. The defining set of relations mimic
the isotopies of virtual singular links. Further we prove that the virtual singular braid monoid has another presentation
using fewer generators and relations.
272. The Three-Variable Bracket and Tutte Polynomial
Maggie Miller University of Texas at Austin
Advisor(s): Rolland Trapp, California State University San Bernardino
Kauffman’s three-variable bracket for a knot diagram and the Tutte graph polynomial are highly related. In this poster,
we derive an expression for each coefficient of the three-variable bracket for an alternating diagram in terms of the
Tutte polynomial of the checkerboard graph. We use this relation to recover twist number of two-bridge links and
alternating three-braid links from the three-variable bracket. We also show that we may almost recover the number of
crossings in each twist from the three-variable bracket.
273. Harmonic Nine Crossing Knots
Rebecca Miller Carthage College
Advisor(s): Aaron Trautwein, Carthage College
A knot is an embedding of the unit circle into real three-space. Harmonic knots are defined as trigonometric polynomial
parameterizations of the unit circle. In this poster we construct harmonic nine crossing knots. To construct harmonic
knots we created a program in Mathematica that uses truncated Fourier series to approximate linear functions representing the paths of polygonal knots. Harmonic representations of knots are used to define knot invariants including
the harmonic index. The harmonic index of a knot is the minimum integer n such that there is an nth degree harmonic
parameterization of the same knot type. Previous research found the harmonic parameterizations of knots types up
through eight crossing knots. The parameterizations presented in this poster are used to determine upper bounds for
the harmonic index of each nine crossing knot. These parameterizations also aid in the research of new knot invariants
by providing more examples of harmonic representations.
Index
Adams, Richard (175), 50
Alexander, Elaine (116), 35
An, Jing (87), 27
Ander, Erik (115), 35
Anderson, Loren (26), 11
Antonellis, Marcus‘ (253), 68
Anzis, Benjamin (34), 13
Appelbaum, Sarah (250), 68
Ariche, Jeremy (98), 30
Arnold, Andrew (149), 44
Baello, Rob Rexler (198), 55
Ball, Chelsea(114), 35
Banks, Josiah (32), 12
Barker, William (83), 26
Barron, Thomas (9), 7
Bartlette, Kai (48), 16
Bean, Brennan (110), 34
Bekele, Abraham (27), 11
Belt, Andrew (50), 17
Beneish, Lea (197), 55
Berger, Alexander (172), 49
Berlin, Heather (238), 65
Berns-Zieve, Rose (29), 12
Betancourt Velez, Fernando (3), 5
Bishop, Berkeley (176), 50
Bladow, Daniel (69), 22
Bland, Jared (108), 33
Brandao, Mark (78), 25
Brookner, Aaron (5), 6
Brown, Jared (73), 23
Brown, Ryan (158), 46
Brynes, Alyssa (216), 59
Bucknam, Brian (89), 28
Bugg, Caleb Bugg (94), 29
Buhr, Matthew(109), 33
Burdick, Joseph (162), 47
Burris, Christie (171), 49
Bustamante, Jessica (51), 17
Byrnes, Jessie (55), 18
Camacho, Alejandro (86), 27
Campbell, Kevin (212), 58
Cericola, Christopher (255), 69
Chambers, Megan (30), 12
Chan, Jonathan (204), 56
Chandra, Sharat (38), 14
Chen, Linhan (174), 49
Chiros, Heather(164), 47
Chou, Amy (101), 31
74
Christianson, Katherine (270), 72
Chu, Tiffany (80), 25
Clark Jr., Curtis (177), 50
Clarke, Jahlani (55), 18
Clemente, Gabriella (159), 46
Cleveland, Joseph (223), 61
Coleman, Donald (198), 55
Coleman, Mercedes (147), 43
Collazo, Ram´on (207), 57
Conaway, Caitlyn(14), 8
Condon, Daniel (244), 66
Cook, Jared (61), 20
Coons, Jane (112), 34
Cornelus, Cecile(163), 47
Coronado, Mayra (68), 22
Cowie, Lauren (33), 13
Crawford, Jennifer (44), 15
Crowell, Rachel (63), 20
Cyrs, Asha (62), 20
Dao, Mai (140), 42
Dasenbrock-Gammon, Nathan (135), 40
Davies, Samantha (85), 27
Davis, Bryant (201), 56
Davis, Jeffrey (186), 52
De La Cruz Santos, Jessie (148), 44
De la Cruz, Julio(207), 57
De la Pena, Andrew (271), 73
de la Torre, Natalia (68), 22
DeGraaf, Stephanie (166), 48
Deleon, Sabrina (241), 65
DeWolf, Lydia (226), 62
Diallo, Thierno (172), 49
Domagalski, Rachel (1), 5
Dombrowski, Timothy (55), 18
Dona, Reanna (111), 34
Donahoe, Ailaura (114), 35
Donahue, Kathleen (130), 39
Donahue, Kathleen (56), 18
Donatelli, Ross(48), 16
Doughty, Robert (176), 50
Du, Yilun (95), 29
Dunn, Robin (251), 68
Dzugan, Jeffrey (223), 61
Edwards, Ashley (124), 38
Egging, Paula (232), 63
Erickson, Jon (163), 47
Estrada, Emmanuel (80), 25
Evans, Ciaran (249), 67
75
Titles, Authors, Advisors and Abstracts
Felix, Gilbert (160), 46
Ferguson, Kyle (2), 5
Fink, Anne (77), 24
Finney, Brandon (191), 53
Fisher, Lee (58), 19
FitzGerald, Cody (147), 43
Forster, Zach (28), 11
Forsyth, Adam (199), 55
Fountain, Rachael (92), 29
Frazier, William (102), 31
Frechette, Claire (206), 57
Frederick, Leah (62), 20
Freedman, Benjamin(124), 38
Fry, Andy (219), 60
Fry, Dana (29), 12
Fu, Casey (32), 12
Heath, Emily (228), 62
Heil, Nicholas (243), 66
Herman, Elizabeth (136), 41
Hillier, Sarah (32), 12
Hohman, Frederick (138), 41
Holmes, Amber (147), 43
Hong, Fangyuan (128), 39
Hong, Su Ji (64), 21
Hoover, Joseph (179), 51
Hopkins, Joseph (152), 45
Houser, Jennifer (125), 38
Huang, Brice (181), 51
Hulsey, Samantha (257), 69
Hund, Parker (249), 67
Hunt, Emily (119), 36
Hyde, Nicolas (42), 15
Gaetz, Christian (132), 40
Gallagher, John (188), 53
Gallagher, Ryan (148), 44
Gandhi, Kavish (167), 48
Garrett, Shontrice (193), 54
Gates, Dante (35), 13
Getsos, Julia (32), 12
Gibson, Thomas (138), 41
Gillings, Johnny (29), 12
Gills, Joseph (157), 46
Goettler, Kenneth (121), 37
Goetz, Hanah (163), 47
Golbek, Megan (78), 25
Goluboff, Justin (270), 72
Gomez, Justin (190), 53
Gonda, Jessica (176), 50
Gonzales, Ilia (160), 46
Gonzalez, Oscar (211), 58
Gonzalez, Osvaldo (160), 46
Goodrich, Timothy (180), 51
Goudreau, Jacob (75), 24
Gowravaram, Niket (237), 64
Graft, William (41), 15
Green, Arman (103), 32
Grez, Armando (164), 47
Gulbrandsen, Daniel (242), 66
Gupta, Vineet (10), 7
Gurev, Jack(199), 55
Gusman, Anthony (52), 17
Guzman, Jonathan (184), 52
Ifill, Jassiem (144), 43
Hadaidi, Sarah(148), 44
Hagen, August(40), 14
Hallahan, William (17), 9
Hamann, Linus (270), 72
Handagama, Winode (227), 62
Hanson, Angela (91), 28
Harper, Rebecca (90), 28
Hartland, Tucker (78), 25
Hartmann, Kelly (194), 54
Hauschild, Jessica (151), 44
Hauser, Paloma (251), 68
Hawekotte, Keenan (81), 26
Jansen, Caroline (118), 36
Javier, Kayla (205), 57
Johnson, Eric (97), 30
Johnson, Talon (127), 38
Jones, Edna (203), 56
Kaashoek, Justin (101), 31
Kandi, Dorian (55), 18
Kaplan, Elliot (262), 70
Kawana, Jessica (169), 48
Kelley, Victoria (110), 34
Kelling, Claire (63), 20
Kemp, Mary (161), 47
Kemp, Mary (229), 63
Kennedy, Matthew (152), 45
Kennedy, Sidney (154), 45
Kernodle, Krista (126), 38
Keys, Janae (175), 50
Kiers, Joshua (255), 69
Kilmer, Shelby (12), 7
Kim, Bohyun (134), 40
Kim, Jongwon (174), 49
Kim, Jung Yoon (185), 52
Kinney, Mitchell (238), 65
Kipp, Andrew (91), 28
Kispert, Connor (245), 66
Klarmann, Joshua (186), 52
Knopf, Michael (196), 55
Konjura, Kolin (117), 36
Kotsonis, Rebecca (201), 56
Krause, Cassidy (103), 32
Krishna, Nitin (107), 33
Kristensen, Hanna (225), 61
Kumerow, John (258), 69
Kurlej, Arthur (137), 41
Kwon, Soonho (204), 56
Lacey, Dana (248), 67
Lai, Eric (104), 32
Lancto, Zachery (173), 49
Latham, Andrew (46), 16
76
Laudone, Robert (232), 63
Lebron, Dayanara (122), 37
Lee, Hye Rin Lin (86), 27
Lehet, Ellen (269), 72
Lemons, Mitchell (25), 10
Lenz, Nicole (258), 69
Levin, Derek (236), 64
Lewis, Kyle (246), 67
Li, Hannah (63), 20
Li, Weiwei (21), 9
Linz, William (164), 47
Liu, Hengzhou (28), 11
Liu, Xuanhe(88), 27
Liu, Yuting (88), 27
Locus, Madeline (206), 57
Logan, Brooke (131), 39
Loh, Isaac (43), 15
Long, Mary (192), 54
Longo, Vincent (267), 72
Lopez, Nicole (265), 71
Luo, Jerry (129), 39
Mackey, Crystal(208), 58
MacMillan, Samantha (189), 53
Magos, Lucia (40), 14
Magyar, Christopher (215), 59
Mailloux, Ashley (195), 54
Malachi, Meghan(195), 54
Mann, Dayna (111), 34
Marek, Robert (96), 30
Marquis, Jordan (100), 31
Martin, Bryan (238), 65
Masihimirzakhanian, Talin (66), 21
Maslak, Susan (161), 47
Maslak, Susan (229), 63
Massaro, Marina (231), 63
Mastrianni, Michelle (132), 40
Mathews, Heather (29), 12
Matys, Elizabeth (239), 65
Matzke, Ryan (217), 60
Mauretour, Christie (73), 23
Maxwell III, John (62), 20
Mazurowski, Liam (226), 62
McDaniel, Margaret (227), 62
McGahan, Sarah (271), 73
McGuirl, Melissa (79), 25
McMahon, Elise (7), 6
McMahon, Savanah(116), 35
Mehrle, David (218), 60
Menter, Jordan (65), 21
Metangmo, Baudry (36), 13
Meyers, Wade (66), 21
Meyerson, Olivia (235), 64
Meza, Jeremy (235), 64
Michel, Alexandra (214), 59
Miller, Caleb (53), 18
Miller, Maggie (272), 73
Miller, Matthew (261), 70
Miller, Rebecca (273), 73
Milzman, Jesse (196), 55
Titles, Authors, Advisors and Abstracts
Mingyang, Zhang (88), 27
Mitchell, Kelsey (226), 62
Mitnick, Alice (202), 56
Monarrez, Omar (145), 43
Moose, Aaron (46), 16
Mora, Juan (142), 42
Morales, Adriana (176), 50
Morfe, Peter (54), 18
Morgenstern, Ilan (60), 20
Morrison, Theresa(138), 41
Morse, Anthony (223), 61
Murphy, Jason (220), 60
Myers, Nicholas (61), 20
Naden, Samuel (81), 26
Narayanan, Shyam (210), 58
Nasif, Hassan (153), 45
Negassi, Frehiwet (230), 63
Negron, Raul (211), 58
Nejame, Matt (89), 28
Newman, Maisie (268), 72
Ng, Nathan (50), 17
Nguyen, Anh (248), 67
Nguyen, Vy (124), 38
Nicholson, Lillian (256), 69
Nicholson, Lillian (257), 69
Nicosia, Catherine (146), 43
Ning, Nina(61), 20
Novitzky, Sophia (81), 26
Nu, Chrono (45), 16
Nugent, Peter (236), 64
O’Brien, Patricia (262), 70
O’Connor, Brian (84), 26
O’Driscoll, Mary (47), 16
Oakley, Bryan (136), 41
Olafson, Hailey (13), 8
Ordog, Erika (225), 61
Ortiz, Jazmin (151), 44
Ortiz, Luis (4), 5
Owens, Amanda(232), 63
Pane, Gianluca (209), 58
Paoletti, Christopher (57), 19
Park, Sung-Min (72), 23
Pennington, Hannah (107), 33
Perez, Jessica (114), 35
Petersen, Ethan (254), 69
Pham, Kim (21), 9
Phillips, Anya (238), 65
Porche, Jeremy (14), 8
Potocki, Alex (82), 26
Quan, Hadrian (37), 14
Quijano, Alex John (99), 31
Rabbani, Tahseen (72), 23
Rachan, Rebecca (259), 70
Raghavendran, Surya (155), 45
Rahmani, John (252), 68
77
Titles, Authors, Advisors and Abstracts
Ramirez, Daniel (207), 57
Randolph, Michelle (130), 39
Randolph, Michelle (56), 18
Read, Catherine (48), 16
Rebrovich, Jackson (15), 8
Rebrovich, Jackson (208), 58
Reckner, Tyler (240), 65
Reiswig, Josiah (176), 50
Reith, Thomas (141), 42
Rennie, Joseph (214), 59
Resh, Jill (224), 61
Reyes, Tiffany (109), 33
Richardson, Juliana (110), 34
Richburg, Aquia (182), 51
Riedl, Austin (25), 10
Riedl, Austin (28), 11
Roberts, Michael (105), 32
Roberts, Nicholas (116), 35
Robertson, Shelby (14), 8
Robichaux, Colleen (132), 40
Robinson, Todderrick (193), 54
Rojina, Nicholas (208), 58
Rojina, Nico (15), 8
Rolling, Tim (226), 62
Roman, Fernando (37), 14
Rosado, Michelle (6), 6
Rosenberg, Katie(27), 11
Roth, Grant (266), 71
Roy, Uma (237), 64
Roy, Uma(10), 7
Rudy, Kaitlin (190), 53
Rummerfield, Wendy (133), 40
Ryan, Jack (67), 22
Rychnovsky, Mark (178), 50
Sakurai, Cameron (82), 26
Salazar Cardona, Aura (11), 7
Sanchez, Camille (62), 20
Sapp, Mahalia (81), 26
Sargent, Kelsie (258), 69
Sass, Julian (73), 23
Saucedo, Antonio (160), 46
Schadler, Lindsey (238), 65
Schlenk, Taylor (223), 61
Schmachtenberger, Benjamin (40), 14
Schonsheck, Nikolas (184), 52
Schreiner-McGraw, Jeff (31), 12
Schulte, Abraham(19), 9
Schwarze, Hannah (65), 21
Schwein, David (132), 40
Scott, Emily (16), 8
Seibold, Tom (251), 68
Senkevich, Julia (71), 23
Serafin, Luke (244), 66
Setra, Rafael (168), 48
Shankar, Ravi (78), 25
Shapiro, Kayla (129), 39
Shehadi, Eric (93), 29
Shenfeld, Yair (260), 70
Sheng, Yan (200), 56
Sheng, Yan (78), 25
Shi, Hao-Jun (129), 39
Shogren, Madeline (265), 71
Shrima, Shakthi (199), 55
Shuai, Ziyue (73), 23
Sichak, Ashley (106), 33
Siegler, Kimberly (39), 14
Simon, Samuel (118), 36
Slyman, Katherine (176), 50
Smith, Emily (147), 43
Spaw, Jonathon (195), 54
Spencer, Elizabeth(54), 18
Spitzer, Joseph (143), 42
Sprague, Samantha (233), 63
Spychalla, Claire (213), 59
Stafford, Cody (247), 67
Stanley, Sarah (120), 36
Stanley, Zofia Stanley (221), 61
Stenslie, Solvei(46), 16
Stockdale, Cody (244), 66
Stoner, David (183), 51
Stoops, Sarah (156), 45
Stotts, Ontario (3), 5
Stover, Sara (8), 6
Stubbs, Kevin (31), 12
Stubbs, Kimberly (70), 22
Suarez, Crisel (139), 42
Suh, Hong (1), 5
Sulyok, Cara (71), 23
Sumdani, Hasan (109), 33
Sutter, Matthew (252), 68
Takaki, Nicholas (87), 27
Talis, Emma (110), 34
Tamagawa, Sherilyn (263), 71
Tate, Reuben (22), 10
Tatum, Elizabeth(118), 36
Thieme, Cameron (264), 71
Thompson, Albany (150), 44
Tlachac, ML (236), 64
Torres, Josh (100), 31
Tran, Christina (194), 54
Tran, Thao (28), 11
Trebbien, Joshua (41), 15
Tucker, Cody (74), 24
Updike, Joshua (76), 24
Van Alstine, James (13), 8
Van Peski, Roger (10), 7
VanSchalkwyk, Samantha (19), 9
Varadaraj, Srikar (270), 72
Varley, Erics (249), 67
Veconi, Dominic (226), 62
Velazquez, Josean (113), 34
Walter, Zach (234), 64
Wang, Sirui (249), 67
Wang, Yi (222), 61
Ward, Abigail (235), 64
78
Washington, Michole(37), 14
Webb, Gautam (22), 10
Welle, Thomas (77), 24
Wells, Conor (110), 34
Wifvat, Katie (258), 69
Williams, John (123), 37
Wilson, Alexander (18), 9
Withers, James (57), 19
Witz, Evan W (63), 20
Wolfe, David (49), 17
Worsnup, Dalton (8), 6
Wright, Benjamin (27), 11
Yalla, Gopal (85), 27
Yang, Adela (19), 9
Yang, Elizabeth (170), 48
Yang, Joyce(3), 5
Yang, Qi (129), 39
Yanit, Kathreen (54), 18
Young, Jessica (20), 9
Yu, Alexandria (23), 10
Yuan, Amber (216), 59
Zellmer, Jason (59), 19
Zhang, Jesse (187), 52
Zhang, Xingyu(1), 5
Zhang, Yiguang (165), 47
Zheng, Bruce (24), 10
Zhou, Lingxin (60), 20
Zhu, Dantong (196), 55
Zhu, Kan (129), 39
Zirlin, Dara (196), 55
Titles, Authors, Advisors and Abstracts

Similar documents

×

Report this document