Application of variance reduction techniques for tau-leaping systems to particle filters

We propose several techniques for simulation of the prediction step of particle filters for application to a class of estimation problems. These techniques are based on variance reduction techniques previously proposed for stochastic simulation of tau-leaping processes. Proof of almost sure weak convergence is given for an approach in one dimension, and numerical examples pertaining to building population dynamics are presented in both single and multidimensional tau-leaping cases.