Phil Pollett's Research Pages
[Home]
ARC Funded Projects
Analytical and Computational Methods for Studying Explosive Random Processes (2001-2003)
Summary
Many stochastic models, such as those used in the study of populations, account for explosive behaviour by allowing infinitely many transitions in a finite time. As a result, there can be a range of processes sharing the same parameters. It is therefore important to be able to identify those processes that satisfy some additional criteria. We shall adopt a fresh approach based on approximating the transition probabilities, quantities that govern the evolution of the process. Our methods will provide a computationally effective means for evaluating quasi-stationary distributions, statistical tools that are useful in describing the long-term behaviour of these processes.
If you have any comments on these pages,
feel free
to e-mail me:
pkp@maths.uq.edu.au