In this project methodological research focuses on the development of computationally intensive methods for statistical inference under both the classical and Bayesian paradigms. The work can be split into several key areas: the development of novel simulation algorithms; the development of tools to assess the performance of existing simulation algorithms; and the mathematical study of the performance and properties of new and existing simulation methods.

We are specially interested in the development of reversible jump MCMC methodology which provides a very powerful suite of tools for tackling model discrimination problems. Work in this area ranges from analysing the mathematical properties of these algorithms in terms of their ability to rapidly explore a vast array of high dimensional probability surfaces, to more practical implementational problems such as the developing tools to check that the algorithms have performed well.

1. Ricardo Ehlers, LEG/UFPR
2. Steve Brooks, University of Cambridge
3. Luiz Ledo, UFRJ