Rigorous numerics for a singular perturbation problem

TL;DR

This paper develops a computer-assisted rigorous numerical method combined with topological tools to prove the existence of periodic and homoclinic orbits in fast-slow systems like FitzHugh-Nagumo for small parameter ranges.

Contribution

It introduces a novel rigorous numerical approach for analyzing singular perturbation problems, specifically for proving orbit existence in the near-zero epsilon regime.

Findings

Proves existence of periodic orbits in FitzHugh-Nagumo for explicit epsilon range.

Establishes existence of homoclinic orbits for explicit epsilon values.

Provides a computational framework for rigorous analysis of fast-slow systems.

Abstract

Fast-slow systems are notoriously difficult to analyze with rigorous numerics, since the qualitative properties of the solution space change fundamentally when the so-called small parameter $\epsilon$ is varied from 0 to small non-zero values. In this dissertation I develop a computer-assisted rigorous method which can be used in combination with topological tools for proving the existence of period and connecting orbits in the near-zero parameter regime. As an application, I prove the existence of periodic and homoclinic orbits in the FitzHugh-Nagumo system, for $\epsilon \in (0,\epsilon_0]$ with an explicit $\epsilon_0$. This dissertation was prepared under supervision of prof. Piotr Zgliczy\'nski and submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, awarded at the Jagiellonian University, Department of Mathematics and Computer Science in June 2016. Some of the results of this dissertation have been previously published in A. Czechowski and P. Zgliczy\'nski. Existence of periodic solutions of the FitzHugh-Nagumo equations for an explicit range of the small parameter. arXiv:1502.02451, SIAM J. Appl. Dyn. Syst., 15(3), 1615-1655 and text overlap may occur. The dissertation also contains some (yet) unpublished results, in particular concerning the existence of homoclinic orbits for explicit ranges of $\epsilon$.