Back to results

George Mason University

Topological Methods for Evolution Equations

Abstract

In this dissertation we develop a framework for rigorously computing the Conley index of isolated invariant sets for flows generated by finite-dimensional systems of ordinary dif- ferential equations x ̇ = f(x), where f : Rn → Rn. Our main contribution in this area is the characterization of isolating blocks in terms of the level sets and superlevel sets of two real- valued functions, u,v : Rn → R. The functions u and v incorporate geometric quantities computed on the boundary of proposed isolating blocks and relate them to local behavior of the vector field f. In order to obtain numerically rigorous results in this area, we have developed a new tool for computing superlevel sets of real-valued functions u : Rn → R that guarantees our superlevel set approximations are homotopy equivalent to the actual superlevel sets we are interested in. This new tool is presented as the first logical half of this document, as it is a significant advance in its own right. Our work makes use of basic differential geometry on piecewise smooth manifolds, ex- ploits the interplay between flows and the topology of the underlying phase space (provided by the Waz ̇ewski theorem), and employs interval arithmetic and automatic differentiation. We provide full details for a collection of algorithms which enable the practitioner to easily apply our framework to a wide variety of problems in the theory of dynamical systems. Several examples are provided showing the relative simplicity of our approach over earlier

Author and committee

dc:creator, dc:contributor.*
Author
  • Stephens, Thomas Dean

Subjects

dc:subject × 7

Identifiers

dc:identifier.*
Identifier
hdl:1920/10461
OAI identifier oai:identifier
oai:MARS:1920/10461

Chain of custody

source
Harvested from
George Mason University
Base URL
mars.gmu.edu/server/oai/request
Last updated
2026-07-27
Source record
OAI-PMH GetRecord
citation

Stephens, Thomas Dean. Topological Methods for Evolution Equations. 2016.