Perfect sampling from independent Metropolis-Hastings chains☆

Abstract “Perfect sampling” enables exact draws from the invariant measure of a Markov chain. We show that the independent Metropolis-Hastings chain has certain stochastic monotonicity properties that enable a perfect sampling algorithm to be implemented, at least when the candidate is overdispersed with respect to the target distribution. We prove that the algorithm has an optimal geometric convergence rate, and applications show that it may converge extreme rapidly.

[1]  J. Møller.  Perfect simulation of conditionally specified models , 1999 .

[2]  L. Tierney.  Markov Chains for Exploring Posterior Distributions , 1994 .

[3]  J. Propp,et al.  Exact sampling with coupled Markov chains and applications to statistical mechanics , 1996 .

[4]  A. A. Borovkov,et al.  STOCHASTICALLY RECURSIVE SEQUENCES AND THEIR GENERALIZATIONS , 1992 .

[5]  R. Tweedie,et al.  Perfect sampling of ergodic Harris chains , 2001 .

[6]  R. Tweedie,et al.  Simulating the Invariant Measures of Markov Chains Using Backward Coupling at Regeneration Times , 1998 .

[7]  S. Foss,et al.  Two ergodicity criteria for stochastically recursive sequences , 1994 .

[8]  Luigi Accardi,et al.  Probability Towards 2000 , 1998 .

[9]  Olle Häggström,et al.  Characterization results and Markov chain Monte Carlo algorithms including exact simulation for some spatial point processes , 1999 .

[10]  R. Tweedie,et al.  Rates of convergence of the Hastings and Metropolis algorithms , 1996 .

[11]  J. A. Fill.  An interruptible algorithm for perfect sampling via Markov chains , 1998 .

[12]  A. A. Borovkov,et al.  Asymptotic Methods in Queueing Theory. , 1986 .

[13]  P. Green,et al.  Exact Sampling from a Continuous State Space , 1998 .

[14]  R. Tweedie,et al.  Perfect simulation and backward coupling , 1998 .

[15]  Wilfrid S. Kendall,et al.  Perfect Simulation for the Area-Interaction Point Process , 1998 .

[16]  Richard L. Tweedie,et al.  Markov Chains and Stochastic Stability , 1993, Communications and Control Engineering Series.