Adaptive parallel tempering algorithm
arXiv:1205.1076
Abstract
Parallel tempering is a generic Markov chain Monte Carlo sampling method which allows good mixing with multimodal target distributions, where conventional Metropolis-Hastings algorithms often fail. The mixing properties of the sampler depend strongly on the choice of tuning parameters, such as the temperature schedule and the proposal distribution used for local exploration. We propose an adaptive algorithm which tunes both the temperature schedule and the parameters of the random-walk Metropolis kernel automatically. We prove the convergence of the adaptation and a strong law of large numbers for the algorithm. We illustrate the performance of our method with examples. Our empirical findings indicate that the algorithm can cope well with different kind of scenarios without prior tuning.
33 pages, 3 figures
References in corpus (7)
- Simulated Tempering: A New Monte Carlo Scheme
- On the ergodicity properties of some adaptive MCMC algorithms
- Quantitative bounds on convergence of time-inhomogeneous Markov chains
- Conditions for rapid mixing of parallel and simulated tempering on multimodal distributions
- Convergence of adaptive and interacting Markov chain Monte Carlo algorithms
- On the ergodicity of the adaptive Metropolis algorithm on unbounded domains
- On the stability and ergodicity of adaptive scaling Metropolis algorithms