A Quasi-Deterministic Approximation of the Monte Carlo Importance Function

Publisher Website
Google Scholar
Abstract
The basic quasi-deterministic method provides an approximate importance function in arbitrary user-defined phase-space regions. The approximation is twofold. First, each region is averaged over and becomes a discrete state. Second, Monte Carlo methods estimate transport probabilities and scores between the discrete states. These two approximations lead to a set of linear equations for the state importances that can be deterministically solved. This new method is compared against the standard MCNP importance generator. A generalization of the method provides an importance function in the physical and random number spaces that may be useful for random number biasing techniques.