Date of Award

8-2025

Degree Type

Masters Thesis

Degree Name

Master of Science (MS)

School

Mathematics and Natural Sciences

Committee Chair

Dr. James Lambers

Committee Chair School

Mathematics and Natural Sciences

Committee Member 2

Dr. Qingguang Guan

Committee Member 2 School

Mathematics and Natural Sciences

Committee Member 3

Dr. Huiquing Zhu

Committee Member 3 School

Mathematics and Natural Sciences

Abstract

This thesis presents a scalable, high-resolution simulation framework for a reaction diffusion prey-predator model with finite hunting ranges. The model’s integral predation term produces stiff, non-local dynamics that present a challenge to traditional explicit or fully implicit solvers. This study overcome this barrier by incorporating a third-order Krylov Subspace Spectral (KSS) time integrator into a Fourier spectral spatial discretization. KSS, which treats each Fourier mode with a frequency-dependent polynomial, avoids the need for solutions of large linear systems in implicit schemes and maintains the computing simplicity of explicit schemes. It also allows for larger time steps compared to stability-limited Runge-Kutta methods. Benchmarking KSS against LEJA interpolation and Adaptive Krylov-Projection (AKP) exponential integrators on grids up to N = 8000 shows it maintains a constant iteration count of four Fast fourier tranforms (FFTs) per iteration, achieves relative errors below 10−8 and delivers result up to 40 times faster. Large-scale simulations show a variety of spatial and temporal behaviors, including stationary prey refuges, periodic and aperiodic traveling waves, and regime coexistence. These findings support and expand prior analytical predictions for non-local predation.

Available for download on Sunday, August 01, 2027

Share

COinS