Master of Science / Department of Mathematics / Pietro Poggi-Corradini / This thesis contains three chapters split into two parts. In the first chapter, the discrete
p-modulus of families of walks is introduced and discussed from various perspectives.
Initially, we prove many properties by mimicking the theory from the continuous case and
use Arne Beurling's criterion for extremality to build insight and intuition regarding the
modulus. After building an intuitive understanding of the p-modulus, we proceed to switch
perspectives to that of convex analysis. From here, uniqueness and existence of extremal
densities is shown and a better understanding of Beurling's criterion is developed before
describing an algorithm that approximates the value of the p-modulus arbitrarily well.
In the second chapter, an exclusively edge-based approach to the discrete transboundary
modulus is described. Then an interesting application is discussed with some preliminary
numerical results.
The final chapter describes four different takes of the Susceptible-Infected (SI) epidemic
model on graphs and shows them to be equivalent. After developing a deep understanding
of the SI model, the epidemic hitting time is compared to a variety of different graph
centralities to indicate successful alternative methods in identifying important agents in
epidemic spreading. Numerical results from simulations on many real-world graphs are
presented. They indicate the effective resistance, which coincides with the 2-modulus for
connecting families, is the most closely correlated indicator of importance to that of the
epidemic hitting time. In large part, this is suspected to be due to the global nature of both
the effective resistance and the epidemic hitting time. Thanks to the equivalence between
the epidemic hitting time and the expected distance on an randomly exponentially weighted
graph, we uncover a deeper connection- the effective resistance is also a lower bound for the
epidemic hitting time, showing an even deeper connection.
| Identifer | oai:union.ndltd.org:KSU/oai:krex.k-state.edu:2097/20364 |
| Date | January 1900 |
| Creators | Goering, Max |
| Publisher | Kansas State University |
| Source Sets | K-State Research Exchange |
| Language | en_US |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.071 seconds