Particle number control for direct simulation Monte-Carlo methodology using kernel estimates

Abstract

On the concept and theory of induced drag for viscous and incompressible steady flow

Improving computational efficiency in DSMC simulations of vacuum gas dynamics with a fixed number of particles per cell

“…Besides methods involving particle split and merge, alternative approaches involve replacing the original set of particles with a new set of desired particles. Gorji et al [44] proposed an algorithm based on kernel density estimation to control the number of particles, ensuring the generation of independent statistical samples. This approach involves determining the required number of particles for each cell based on its mass content and sampling particles from the estimated distribution, thereby replacing the existing particles.…”