Issues in the Multiple Try Metropolis mixing

Abstract

The multiple Try Metropolis (MTM) algorithm is an advanced MCMC technique based on drawing and testing several candidates at each iteration of the algorithm. One of them is selected according to certain weights and then it is tested according to a suitable acceptance probability. Clearly, since the computational cost increases as the employed number of tries grows, one expects that the performance of an MTM scheme improves as the number of tries increases, as well. However, there are scenarios where the increase of number of tries does not produce a corresponding enhancement of the performance. In this work, we describe these scenarios and then we introduce possible solutions for solving these issues.

References in corpus (5)

• Generalized Multiple Importance Sampling
• Layered Adaptive Importance Sampling
• Orthogonal parallel MCMC methods for sampling and optimization
• Efficient Multiple Importance Sampling Estimators
• Issues in the Multiple Try Metropolis mixing

Cited by in corpus (7)

• A Survey of Monte Carlo Methods for Parameter Estimation
• A Review of Multiple Try MCMC algorithms for Signal Processing
• Orthogonal parallel MCMC methods for sampling and optimization
• Group Importance Sampling for Particle Filtering and MCMC
• Issues in the Multiple Try Metropolis mixing
• Lattice Gaussian Sampling by Markov Chain Monte Carlo: Bounded Distance Decoding and Trapdoor Sampling
• A multiple-try Metropolis-Hastings algorithm with tailored proposals