The Northfield Undergraduate Mathematics Symposium (NUMS) is an annual event sponsored jointly by Carleton and St. Olaf. In 2023, eight students spoke at the symposium on research they did around the country over the summer.
Mandelbrot Breadcrumbs
Murali Meyer (St. Olaf)
In this project we explore rational functions of the form f(z) = z^n + a/(z^d) + c where z is a complex number, a is a fixed real number, n and d are integers greater than 2, and c is a complex-valued parameter. We locate where, for a fixed above a specified lower bound, the c-parameter plane contains n+d homeomorphic copies of the Mandelbrot set. These baby Mandelbrot sets are distributed across (n+d)/gcd(n,d) distinct regions of the parameter plane. Our work generalizes the research done by Devaney; Jang, So, and Marotta; and Boyd, and Mitchell, each of which examined similar but simpler functions. Like this previous research, we take advantage of techniques first introduced by Douady and Hubbard to prove our results.
The Stable Tamari Order: Applications of Graph Theory to Algebraic Combinatorics
Nzingha Joseph (Carleton)
Conducted at the NYC Discrete Math REU in Summer 2023, our research aims to investigate the combinatorial properties of the Stable Tamari Order. The (classical) Tamari Order is a poset that can be used to order many different combinatorial objects, and a limited subset of non-negative integer sequences. The Stable Tamari Order is a new generalization of the Tamari Order which expands the scope of the Classical Tamari Order to all non-negative integer sequences, broadening the scope of combinatorial objects our poset can describe. We use algebraic objects called the h-vector and M-matrix to describe properties of the Stable Tamari Order, and show results and conjectures obtained from the application of graph theory to the interpretation of these M-matrices.
Examining the Effects of Route of Exposure in a Physiologically Based Pharmacokinetic (PBPK) Model of Chloroform in Rodents
Will Asinger (St. Olaf)
This project develops a multi-route Physiological Based Pharmacokinetic (PBPK) model to assess the impact of various exposure routes on chloroform levels in the body over time. Unlike previous models that focused on specific routes, this study considers oral, dermal, inhalation, and intravenous exposures. Simulations were used to compare routes of exposure in silico. Then, the model was fit to experimental data on different routes of exposure in rats and mice. Comparing exposure routes, intravenous administration proved most efficient, causing rapid increases in concentration in tissues. Conversely, dermal exposure resulted in a slower distribution of chloroform. This model accurately replicated the pharmacokinetics of chloroform in rats and mice, improving our knowledge of the toxicology of chloroform, a common pollutant.
Random Multiplicative Functions over Function Fields
Annemily Hoganson (Carleton)
One way to investigate the distribution of prime numbers is to ask probabilistic questions about multiplicative functions, which are determined by their behavior on the primes. It is sometimes possible to obtain stronger results studying these questions for multiplicative functions defined over function fields rather than over integers, and these results in the function field setting may provide insights into the integer setting. We study the natural moments of sums of random multiplicative functions over function fields. Using combinatorial and analytic arguments, we obtain an exact expression for the 4th moment, which is unattainable in the integer setting, and an asymptotic expression for the higher natural moments.
Algorithmic Generation of DNA Self-Assembly Graphs
Grace Bielefeldt (St. Olaf)
With recent advancements in the field of nanotechnology, there has been increasing interest in self-assembling nanostructures. These are constructed through the process of branched junction DNA molecules bonding with each other without external guidance. Using a flexible tile-based model, we represent molecules as vertices of a graph and cohesive ends of DNA strands as complementary half-edges allowing the molecules to bond with each other. Due to the unpredictability of DNA self-assembly in a laboratory setting and the risk of undesirable products being incidentally constructed, predicting what structures can be produced from a given list of components, referred to as a “pot of tiles,” is useful but has been proven NP-hard. Our project introduces an algorithm to computationally generate and visualize at least one valid graph given a pot of tiles. By adjusting a number of construction parameters, we can produce graphs of various orders and proportions of tiles.
Optimal Constructions for DNA Self-Assembly of k-Regular Graphs
Xingyi Zhang (Carleton)
Recent advances in nanotechnology and microbiology have spurred research on self-assembling complex structures, primarily using DNA for its complementary base-pairing. These complexes can be represented using graph theory, where edges are strands of DNA joined at junctions, represented by vertices. Branched DNA molecules, referred to as tiles, have flexible unpaired cohesive ends for bond-edges formation. We consider the minimum number of tiles and bond-edge types to construct a graph G, while avoiding the formation of smaller or non-isomorphic graphs of the same size.
A Weighted Backward Shift that is Super Strongly Hypercyclic but not (Hyper)mixing
Ian Curtis (St. Olaf)
A Weighted Backward Shift that is Super Strongly Hypercyclic but not (Hyper)mixing
Recently in the area of Operator Theory (the study of linear maps between vector spaces), more specifically Linear Dynamics, new notions called strong hypercyclicity and hypermixing were introduced and studied. Whenever a new notion is introduced in Linear Dynamics, it is often tested on the family of weighted backward shifts due to this family’s simple yet rich structure. Examples of hypercyclic weighted backward shifts which are not hypermixing are known, but it was an open question as to whether there exists a strongly hypercyclic weighted backward shift which is not hypermixing. We answer this question by showing there does exist such a weighted backward shift.
An Algorithm for Polarizing Typed Combinators
Mimi Rapoport (Carleton)
Previous work has examined the use of polarity marked CCG combinators to natural logic inferences. To contribute to this work, we present an algorithm to add polarity markings to proper typed combinators. The algorithm takes a combinator in the form of an application tree and uses the polarizing rules of application and withdrawal to polarize the combinator. Only sound rules of application and withdrawal are used, inductively resulting in a sound polarized rule for the typed combinator.