Swarm equilibria in domains with boundaries

Author: 
Date created: 
2018-08-13
Identifier: 
etd19783
Keywords: 
Swarm equilibria
Energy minimizers
Gradient flow
Attractors
Nonsmooth dynamics
Nonlinear diffusion
Numerical methods
Abstract: 

This thesis involves the study of a well-known swarming model with interaction and external potentials in one and two dimensions. We refer to this model as the plain aggregation model and later study the model with nonlinear diffusion, so-called the diffusive model here. Typically set in free space, one of the novelties of this thesis is the study of such swarming models in the presence of a boundary. We consider a no-flux boundary condition enforced in a particle context via a ``slip'' condition. Of particular relevance to the context of this thesis, the swarming model used here can be formulated as an energy gradient flow and thusly, one might expect equilibrium states to be minima of the energy. In this work we demonstrate, through both analytical and numerical investigations, a continuum of equilibria of the plain aggregation model that are not minima of the energy. Furthermore, we show that these non-minimizing equilibria are achieved dynamically from a non-trivial set of initial conditions with a variety of interaction potentials and boundary geometries. Thus we show conclusively a deficiency with the plain aggregation model in domains with boundaries, namely that it appears to evolve into equilibria that are not minima of the energy. Following this we then propose a rectification to this deficiency in way of nonlinear diffusion. This choice of nonlinear diffusion is especially attractive because it preserves compact states of the plain aggregation model. We showcase how the diffusive model approaches, but does not equilibrate at, the non-minimizing equilibria of the plain aggregation model. Furthermore we demonstrate how minimizers of the diffusive model do approach minimizers of the plain aggregation model in the zero diffusion limit.

Document type: 
Thesis
Rights: 
This thesis may be printed or downloaded for non-commercial research and scholarly purposes. Copyright remains with the author.
File(s): 
Senior supervisor: 
Razvan Fetecau
JF Williams
Department: 
Science: Department of Mathematics
Thesis type: 
(Thesis) Ph.D.
Statistics: