Perfect simulation

Markov chain Monte Carlo (MCMC) consists of a set of methods for the design of Markov chains whose stationary distribution is the given target distribution. An approximate sample of the target distribution is then produced by sampling a long run of the Markov chain. Perfect simulation methods are a new generation of MCMC algorithms. They extend conventional MCMC methods by ensuring that the sample is not only approximately, but exactly from the stationary distribution. If you'd like to find out more about perfect simulations why not check out the perfect simulation webpage. Prof. Kendall has a cute animation illustrating perfect simulation for the dead leaves model plus some other useful perfect programs. Below are some of my publications on perfect simulation including some introductory arcticles:

• Perfect Simulation of some point processes for the Impatient User, Advances in Applied Probability 31, pages 69 - 87, 1999.
• (joint with W.S. Kendall): Perfect Simulation in Stochastic Geometry, Pattern Recognition 32, pages 1569 - 1586, 1999.
• A Primer in Perfect Simulation, Statistical Physics and Spatial Statistic, ed. Klaus R. Mecke and D. Stoyan, Springer Lecture Notes in Physics, pages 349 - 378, 2000.
• Generic CFTP in Stochastic Geometry and beyond, Revista de Estatistica, Special Issue, Proceedings of the 23rd European Meeting of Statistician in Funchal, pages 81 - 84, 2001.
• Perfect simulation in spatial statistics. In Spatial Statistics Through Applications, pages 21-40. WIT Press, Southampton, 2002.
• (joint with J. Lund): Perfect simulation and inference for point processes given noisy observations. Computational Statistics, 19, pages 317-336, 2004.