Improving SMC sampler estimate by recycling all past simulated particles

Abstract

Sequential Monte Carlo (SMC) methods are not only a popular tool in the analysis of state-space models, but offer a powerful alternative to Markov chain Monte Carlo (MCMC) in situations where static Bayesian inference must be performed via simulation. In this paper, we propose a recycling scheme of all past simulated particles in the SMC sampler in order to reduce the variance of the final estimator. We demonstrate how the proposed approach outperforms the classical strategy in two challenging models.

Preconditioned Monte Carlo for Gradient-Free Bayesian Inference in the Physical Sciences

“…In other words, the importance weights of the particles of the t th iteration with respect to the current t th iteration are adjusted by the normalized ESS given by Equation (11). Equations ( 8)-( 12) are related to the recycling scheme proposed by [17,18] and applied at the end of the run. The main difference is that PR is applied to every iteration of SMC, instead of just the last one.…”