University of Missouri - Saint Louis
The Graduate School
An oral examination in defense of the dissertation for the degree
Doctor of Philosophy in Physics
Prabhavi Kaushalya Premachandra
B.S. in Physics/Chemistry, University of Peradeniya, Sri Lanka, 2003
M.S. in Physics, University of Missouri at St Louis, 2008
Complex Scaling Behavior in Animal Foraging Patterns
This dissertation consists of two parts, and attempts to answer questions from two different areas of biology, ecology and neuroscience, using physics-based techniques.
The first part of my dissertation work sheds light on a poorly researched area of animal movement ecology, i.e., the movement ecology of deterministic foragers that rely on spatial memory to navigate their environment. I address several issues of deterministic foragers relating to movement ecology by analyzing empirical field data and also using theoretical (agent-based) models.
In Chapter 2, I analyze empirical data (move lengths and turn angles) to determine whether Lévy-walk-like movement best describes the movement data of two species of foli-frugivorous primates and also whether the distribution and abundance of resources used by these monkeys could give rise to Lévy-flight-like movement patterns. The results show that there is no support for Lévy-walk-like behavior and the resource distribution pattern does not give rise to such movement in these foraging monkeys.
In Chapter 3, I use agent-based models to simulate search behavior in different environments (landscapes) to investigate the impact of the resource landscape on the optimal foraging movement patterns of deterministic foragers. The computational model I have developed includes parameters such spatial memory and satiation, which have not received much consideration in the field of movement ecology. The results show that when the food availability is abundant or scarce the optimal foraging pattern of a generalist who can consume various and abundant food types does not reach the Lévy range and hence shows no evidence for Lévy-walk-like behavior.
The second part of my dissertation (Chapter 4) presents an investigation of phase transition behavior in a network of locally coupled self-sustained oscillators. A nearest-neighbor coupled Huber-Braun type neural network is constructed to find a possible phase transition behavior (considering the synchronization index as an order parameter) as the system passes through various bursting states. The results suggest that a phase transition does not occur for this locally coupled neuronal network. However, firm conclusions cannot be made from the results produced since the experiments could be repeated for a finer range of coupling constants. The data analysis in my dissertation adopts a model selection approach and relies on methods based on information theory and maximum likelihood.
|Date: July 31, 2012||
Time: 1 pm to 3 pm
Place: SCB 200A
Sonya Bahar, Ph.D. (Advisor)
|Paul Parris, Ph.D. (Co-Advisor)|
|Alexey Yamilov, Ph.D.||Bob Henson, Ph.D.|
|Bette Loiselle, Ph.D.||
Phil Fraundorf, Ph.D.