This article provides results for a comparison of several methods used to generate importance map and improve performances of Monte Carlo simulations. This comparison is made using a specific model that describes a simplified portal monitor used to control illicit radioactive materials in ports. It compares CADIS and FW-CADIS, methods used in ADVANTG to improve the MCNP5 simulation, to the one implemented in RayXpert®, a 3D CAD-based Monte Carlo software developed by TRAD - Tests & Radiations. It is also compared to the method used in TRIPOLI-4®, another Monte Carlo code. Computations were performed on the same model with known results to ensure the validity of this study. The results show a significant gain in performance for MCNP5 using variance reduction techniques and a gain even more important for the Monte Carlo code RayXpert which uses an importance map generation algorithm based on Placzek theory. • Comparison of performance gain between different importance map generation methods. • Use of importance map to enhance Monte Carlo simulation performances. • Comparison between analog and biased Monte Carlo simulations. • Comparing Monte Carlo computation efficiency using importance maps.
Quema et al. (Thu,) studied this question.