The Student Honors Papers collection represents exemplary work in mathematics at Illinois Wesleyan University. The Ames Library is proud to archive these and other honors projects in Digital Commons @ IWU, the University's online archive of student, faculty and staff scholarship and creative activity.

Testing Irreducibility of Trinomials over GF(2)*by Steven Hayman*

The focus of this paper is testing the irreducibility of polynomials over finite fields. In particular there is an emphasis on testing trinomials over the finite field GF(2). We also prove a the probability of a trinomial satisfying Swan's theorem is asymptotically 5/8 as n goes to infinity.

Some NP-Complete Problems in Linear Algebra*by Santhosh Sastry '90*

This research project is aimed at studying the theory of NP-Completeness and determining the complexity of certain problems in linear algebra. The first chapter introduces the reader to Complexity theory and defines NP-Completeness. It is supported by Appendices 1 and 2. Appendix 3 lists some known NP-Complete problems.

On the Construction of Prime Desert n-Tuplets*by Derek M. Marusarz '92*

If we were able to show that there is an infinite number of prime desert n-tuplets of length k, for any appropriate odd k, then we would have solved the twin prime conjecture because then we would know that there would be an infinite number of prime desert 1-tuplets of length k =1, i.e., an infinite number of twin primes.

The Recurrence Relation of B-Wavelets*by Rumi Kumazawa '94*

Our goal is to construct smooth wavelet functions. In constructing such wavelet functions, we need a smooth scaling function to begin with. B-spline functions are suitable as our scaling function because they are piecewise polynomials with compact supports and are relatively smooth. B-wavelet functions are just dilations and translations of these B-spline functions. In addtion, we can find a recurrence relation of the B-spline functions with different order. Hence B-wavelets of any order can be constructed successively from the lower order ones.

Macroelements and Orthogonal Multiresolutional Analysis*by Jonathan M. Corbett '96*

Orthogonal multiresolutional wavelet analysis in a two dimension setting furnishes a basis for wavelet analysis. Bernstein-Bezier polynomials over simplexes provide elegant expressions of the necessary and sufficient conditions for a shift invariant space generating an orthogonal multiresolution analysis.

The B-spline Wavelet Recurrence Relation and B-spline Wavelet Interpolation*by Patrick J. Crowley '96*

In most signal processing applications, a given range of data is best described by a set of local characteristics as opposed to a single global characteristic. In image processing, for example, a region of an image that contains numerous edges is best described as a region whose pixel color values change abruptly, i.e., they are not continuous values of color. A region of constant color, or gradually changing color, is best described as a region whose pixel values are constant, or whose values increase linearly by some factor. It is advantageous to represent this data with signals capable of adapting to these types of local characteristics, as opposed to choosing the best global characteristic. Here, the B-spline wavelet recurrence relation is presented. The B-spline wavelet recurrence relation allows a wavelet of order n +1 to be constructed from a wavelet of order n. This recurrence relation provides a mathematical tool capable of locally varying its degree of continuity. The order of differentiability of a B-spline wavelet increases as the order of the B-spline wavelet increases, and the values range from discontinuous to an arbitrary degree of continuity. A brief discussion of interpolation for splines and B-spline wavelets is introduced as a step toward a future application.

Steiner Trees Over Generalized Checkerboards*by Meta M. Voelker '97*

To minimize the length of a planar network, we can build a Steiner minimal tree that is, a tree consisting of the original network points, as well as additional, strategically-placed (Steiner) points. Chung, Gardner and Graham [2] investigated building Steiner trees over grids of unit squares. We generalize their ideas to grids of rhombuses, and show that two near-optimal Steiner trees exist for each grid, one built from Steiner trees over rhombuses and one built from Steiner trees over isosceles triangles. Further, we conjecture that for grids with an odd number of layers, only the small angle of the rhombus drives which tree is shorter; for grids with an even number of layers, the small angle is the most important factor in determining which scheme to use.

LaSalle's Invariance Principle on Measure Chains*by Anders Floor '00*

It was in 1892 that Lyapunov published his paper giving his "second method". The basic guiding principle was that we might be able to know something about the stability of the system from the form of the equations describing it. Specifically, the idea was that it would not be necessary to know the solutions of the equations involved. This is of course very useful since in most cases solutions are extremely difficult or even impossible to find. Lyapunov's insight was that if a function could be found with, among other properties, a negative rate of change along the solution of the system except in the equilibrium case, then disturbances from the equilibrium solution would return to that solution. (In the equilibrium case, the solution is constant.) The kind of function involved is called a Lyapunov function, and it is defined in such a way that it mimics the energy function. In fact, it was the energy function which originally inspired these ideas. There is an intuitive physical appeal about the assertion that systems that lose energy "fall" to an equilibrium state. And in many cases, the expression for energy ends up being our choice for Lyapunov function. The historical data above can be found in [5].

Applications of the Wavelet Transform to Signal Analysis*by Jie Chen '93*

Like the Fourier Transform, the Wavelet Transform decomposes signals as a superposition of simple units from which the original signals can be reconstructed. The Fourier Transform decomposes signals into sine and cosine functions of different frequencies, while the Wavelet Transform decomposes signals into wavelets. Since the Fourier Transform is a global integration transform and there is no time factor in it, it cannot effectively analyze nonstationary signals whose statistical properties change with time. In order to analyze nonstationary signals, we need to decompose signals into units that are localized in both the time and frequency domains. Using the Wavelet Transform with the B-wavelet, we wrote a program package in Mathematica to implement the decomposition and reconstruction algorithms for signal processing. A data acquisition system developed in another project is used to acquire both the synthesized signals and real voice signals. Application of the Wavelet Transform on these signals will be presented.

Multisurface Method of Pattern Separation*by Jennifer L. Jancik '93*

The recognition and separation of patterns is becoming increasingly important in modern applications. For example, it is currently being used at the University of Wisconsin Hospitals to aid in the diagnosis of breast cancer.