Rose-Hulman Institute of Technology
NSF-REU Site in Mathematics

NSF renewal funding pending

a Research Experience for Undergraduates

Rose-Hulman Mathematics REU History

The Rose-Hulman REU in Mathematics started with Gary Sherman and group theory. This research defined the new mathematical structure of cwatsets and started an investigation of their properties. The program enlarged to include more faculty to broaden the mathematical topics and has included since its inception in 1988 Hyperbolic Geometry and Tilings with Allen Broughton, Number Theory and Permutation with John Rickert, more Cwatsets with Tom Langley, and the present investigations of Inverse Problems, Geometric Analysis, Computational Number Theory, and Systems Biology.

Cwatsets (Gary Sherman, 1998, 2000): A cwatset is a subset of binary n-space that is "nearly" a subgroup.  For example, F = {000,110,101} is a cwatset in binary 3-space.  Notice that while the coset F + 110 = {110,000,011} is not F, it is nearly F in the sense that applying the transposition (1,2) to the components of the elements of F + 110 yields F.  (Take a minute and find the appropriate permutations for the cosets F + 000 and F + 101.  Are your permutations unique?)  So,  F is not closed under addition (i.e., is not a subgroup of binary 3-space), but it is closed with a twist (cwat). Cwatsets have their roots in statistics, grow in groups and combinatorics,  and were conceived by Rose-Hulman undergraduates, Rose-Hulman NSF-REU participants, and Gary Sherman. Here are two references which should be readily available on your campus:

  1. Confidence Intervals from Groups, Mathematics Magazine 65 (1992) 118-122 by Gary Sherman.  This paper provides the motivation for the definition and study of cwatsets.
  2. Introducing . . . Cwatsets!, Mathematics Magazine 67 (1994) 109-117 by Gary Sherman and Martin Wattenberg.  This paper develops the basic combinatorial and algebraic properties of cwatsets.
Starter problems on cwatsets include (but are not restricted to) determining all cwatsets of "small" cardinality, providing alternate characterizations of cwatsets, inventing algorithms for constructing cwatsets, pursuing parallels between groups and cwatsets, and determining cwatsets which are "statistically good."   The computer algebra system Magma, in conjunction with the existing theory of cwatsets, will prove useful in constructing and exploring examples.  But --- solving our problems will depend on your mathematical energy and creativity!

For more information see the Cwatset Research Home Page.

Gary J. Sherman
(1988-1996 - director, 1998, 2000-senior investigator)
Gary J. ShermanProfessor Gary Sherman received his B.Sc. (1963) and M.A. (1968) degrees at Bowling Green State University and his Ph.D. (1971) from Indiana University. Since then he has taught at Rose-Hulman with a sabbatical stint at Milliken Textile. After stepping down from a six year term as department head he started the Rose-Hulman NSF-REU site and was its Program Director for eight years. During that time he successfully, shepherded 48 students through their first research experience, conceived of and obtained funding for the Theorodrome, the Mathematics Department's computer laboratory/classroom, started up the Department's Technical Report Series (about 2/3 of which are technical reports and preprints arising from the REU) and developed and refined the program philosophy and traditions (e.g., the Turkey Run canoe trip). Professor Sherman's research interests are in finite group theory and discrete mathematics and has a particular interest in using computer calculation, using Magma, as a vehicle for undergraduate students to make discoveries about mathematics. After a hiatus in the summer 1997,  Professor Sherman rejoined the REU faculty in the summer 1998  with a fresh set of  research problems in Cwatsets, and again in the summer of 2000. Professor Sherman is now devoting attention to making his work with undergraduates on cwatsets better known. (see the Cwatsets Page). You can find out more about Professor Sherman through his Home Page.

Hyperbolic Tilings of Riemann Surfaces (Allen Broughton, 1998, 1999, 2000, 2001, 2002, 2003): Three examples of tiled surfaces are the icosahedral tiling of the sphere and two different tilings of the torus by equilateral triangles and by isosceles triangles (click to view). Tilings of the hyperbolic plane induced by tilings of the higher genus surfaces may also be constructed: T245, T255, T355, and T433. However, the corresponding surfaces cannot be easily visualized as in the case of the sphere and torus, so group theoretic methods are needed to determine properties of the geometry. All tilings considered generate large symmetry groups of the surface, and the geometrical and combinatorial properties of the tiling are strongly reflected in the structure of the group and the group's action on the geometry. The geometric and combinatorial problems need to be solved by massive computations in the symmetry groups. Motivated by these problems, participants will perform group theoretic experiments, make discoveries and formulate conjectures by carrying out computer calculations using the software package Magma. The end goal will be to discover and prove theorems about the geometry and combinatorics of the surfaces and, of course, anything about groups or computations in groups discovered along the way.

This will be the fourth summer of tiling research at Rose-Hulman and we expect to add to the ever growing knowledge of tilings of surfaces. A sample problem, on which progress was made last year, is to find all the tilings surfaces of low genus. Another is to determine the number of intersection points of two curves in the tiling. They are always exactly two on the sphere. Check it out in the icosahedral tiling of the sphere.  A third project which has now been completed is to determine all the quadrilateral tilings of the hyperbolic plane which can be subdivided into a tiling by triangles.

Although obviously helpful, no previous knowledge of the geometry of hyperbolic surfaces will be assumed. This will be learned on an as needed basis through the perspective of the group theory. The experimentation and testing of conjectures will be done by group theoretic calculations using Magma. The participants will learn how to use Magma on an as needed basis to solve problems. Therefore, prior knowledge of Magma is not assumed, though participants should have some programming experience so they will not be starting from scratch.  Additionally, Maple will be used for geometric calculation.

For more information see the Tilings Research Home Page.

S. Allen Broughton
(1996 - consultant, 1997 - 2003 director)
S. Allen BroughtonProfessor Allen Broughton received his B.Sc. (1976) degree from the University of Windsor and his M.Sc. (1978) and Ph.D. (1982) degrees from Queen's University at Kingston. Since then has taught at Memorial University of Newfoundland, (2 years) the University of Wisconsin at Madison, (3 years) and Cleveland State University (8 years) before coming to Rose-Hulman Institute of Technology as head of the Mathematics Department in September, 1994. His interests are in automorphisms of Riemann surfaces, singularity theory, and Lie Groups and Lie Algebra's. You can find out more about Professor Broughton through his Home Page. Professor Broughton was a faculty consultant for the Rose-Hulman REU during summer of 1996, the Program Director for 1997 program, and a co-PI and PI for the summers 1998-2003. Professors Broughton's REU interest is tilings of surfaces. You can read about the student accomplishments on the tilings project site .

Number Theory (John Rickert, 1999, 2001):  Professor Rickert's REU project is on the partitions of integers.  A partition of the integer n is an increasing sequence of integers whose sum is n. For example we write the 5 partitions of  4 as 4 = 1+1+1+1, 4 = 1+1+2, 4= 1+3,  4=2+2, 4 = 4. Partitions play an significant role in many parts of algebra, especially the study of permutations and the symmetric group. Indeed, the conjugacy classes of the symmetric group on  n symbols are in 1-1 correspondence to the partitions of the integers.  Especially interesting questions about partitions concern the number of partitions  pn  of the integer  n.  Here is a table of the first few:

p1 = 1
 p2 = 2
  p3 = 3
 p4 = 5
p5 = 7
p6 = 11
p7 = 15
p8 = 22
p9= 30
p10 =42
p11 = 56
p12 = 77
p13 = 101
p14 = 135
p15 = 176
p16 = 231
p17 = 297
p18 = 385
p19 = 490
p20 = 627
p21 = 792
p22 = 1002
p23 = 1255
p24 = 1575
p25 = 1958

Looking at the fourth column should suggest a theorem. This is one of three congruence relations observed by Ramanujan. The sequence  { p1 , p2 , p3 , . . . }, the associated q-series p1 q +  p2 q2 +  p3 q3. . . ,  and other related partition functions and q-series lead to many interesting number theoretic questions about  partition numbers.  Rediscovery of the work of Ramanujan, such as the theorem alluded to above,  has led to an explosion of  research on partition functions and q-series impacting areas as  diverse as combinatorics and particle physics. We will be considering identities and related series, building on the results of researchers such as George Andrews. A good understanding of series is very useful here.

John H. Rickert
(1999, 2001-senior investigator)
John H. RickertProfessor John Rickert received his B.Sc. at University of Wisconsin, Madison in 1984 and his Ph.D. at the University of Michigan, Ann Arbor in 1990. He came to Rose-Hulman in 1990, taking a one year sabbatical to Penn State in 1998.  His interests are are in Number theory and  Diophantine approximations with a special interest in simultaneous linear approximations and norm form Diophantine equations. He is also very interested in mathematics competitions and problem solving. Professor Rickert was a faculty mentor in the summers of 1999 and 2001. You can find out more about Professor Rickert through his Home Page.

Cwatsets (Tom Langley, 2004,2005): A cwatset is a subset of binary space that is an additive subgroup wannabe.  For example

F = {000,110,101}

is a subset of binary 3-space which isn't an additive subgroup because, for example,

110 +101 = 011

is not an element of F.  But F is nearly a subgroup in the sense that

  • F + 000 = F,
  • F + 110 = {110,000,011} is just F with the first two components of each word transposed,
  • F + 101 = {101,011,000} is just F with the first and last components of each word transposed.
That is, for each element f of F there exists a permutation, pi, of three symbols such that the coset F + f is just F with pi applied to the components of each element of F.  In other words, while F is not closed under addition, it is closed (c) with (w) a (a) twist (t); i.e., F is a cwatset. 

Cwatsets were first discovered at Rose-Hulman by Gary Sherman and his students in the late 1980s and have their origins in statistics. The algebraic theory of cwatsets loosely parallels the theory of groups (cyclic cwatsets, subcwatsets, extensions, morphisms), supplying a rich inventory of questions. Just about any question on finite groups spawns a similar question on cwatsets. Connections to graph theory (each simple graph has an associated cwatset which completely describes the graph's isomorphism class) and algebraic coding theory (any cwatset is a nonlinear code) also provide fertile avenues of investigation and have sparked much interest within the mathematical community. Research topics for the summer of 2005 will explore the relations of cwatsets to graphs and coding theory, as well as the construction of "perfect" cwatsets and the study of isomorphism classes of cwatsets.

Tom Langley   (2004-2005 - senior investigator)
Tom LangleyProfessor Tom Langley received a B.S. (1989) in electrical engineering from Rice University, then went to work at the Jet Propulsion Lab in Pasadena, during which time he also obtained an M.S. in electrical engineering from USC. He then moved to San Diego, where he obtained an MA in mathematics from San Diego State and a Ph.D. (2001) in mathematics from UCSD. He is interested in algebraic combinatorics, especially the theory and applications of symmetric functions, as well as graph theory and the mathematics of the internet.

Geometric Analysis (David Finn, 2006, 2007, 2008, 2010, 2011, 2012)

David Finn's research concerns the applications of partial differential equations and the calculus of variations to problems in differential geometry, i.e. geometric analysis. The goal of the project to be investigated is modelling the shape of a drop sugar cookie. The basic heurestic model to be investigated is that the shape of a sugar cookie (homogeneous cookie dough) is given as a surface with prescribed mean curvature. The reasoning behind the heurestic model is that the cookie dough becomes a liquid when it is heated and attains its equilibrium shape as a liquid before it solidifies later in the baking process.

Drop of Cookie Dough
(becoming liquid)

Cookies Baking
(attaining shape)

Baked Cookie
(cooling to final shape)

After the cookie is baked, a natural question is: Can one recover the shape of the cookie from knowing the wetted domain (the region on the cookie sheet the cookie is in contract with) and properties of the cookie dough and the cookie drop (stiffness, density, volume)?

Some questions to be investigated are:

  • How does the shape of the cookie depend on the shape of the wetted domain?,
  • How does the shape of the cookie depend on the physical parameters of the dough and the drop?
  • Sometimes when baking cookies, the cookie dough runs together forming a double bubble cookie with a wetted domain as showm below. The cookie shape is then no longer smooth, as in the picture below. How does the shape of the cookie vary in this situation?

A lot of the investigation will be done by numerical computation of solutions to generate conjuctures, No prior knowledge of partial differential equations, differential geometry, and numerical analysis is necessary. Some exposure to ordinary differential equations and/or basic analysis (advanced calculus/vector calculus) is extremely helpful. For more information see the Shape of a Cookie Page

David Finn
(2006 - senior investigator, 2007-2012 co-director)
David FinnProfessor David Finn received his B.S. (1989) from Stevens Institute of Technology and his Ph.D. (1995) from Northeastern University. He taught at both Merrimack College in Massachusetts and Goucher College in Maryland before coming to Rose-Hulman in 1999. He is interested in nonlinear partial differential equations, especially applications to differential geometry, mathematical physics, and image processing, as well as geometric modelling and the mathematics of (bicycle) tire tracks. Professor Finn has run an NSF-funded project, "Motivating Geometry through Computation and Visualization" during the summers of 2002-2005, to develop a course in Geometric Modelling, and mentor undergrads who are developing interactive web-based materials for such a course. You can find out more about Professor Finn through his Home Page.

This document was last modified: 12/01/12
Questions and Comments about the website to: