Theory of direct simulation Monte Carlo method

TL;DR

This paper presents a rigorous theoretical framework for the direct simulation Monte Carlo (DSMC) method, modeling it as a Markov process and deriving the associated equations, including the Boltzmann equation, with applications to both homogeneous and inhomogeneous gases.

Contribution

It introduces a formal Markov process approach to DSMC, derives the master equation, and develops new algorithms for inhomogeneous gases, extending the theoretical understanding of DSMC.

Findings

DSMC modeled as a Markov process with a master equation.

Single particle distribution satisfies Boltzmann and Wang Chang-Uhlenbeck equations.

New algorithm for inhomogeneous gases consistent with Boltzmann equation.

Abstract

A treatment of direct simulation Monte Carlo method (DSMC) as a Markov process with a master equation is given and the corresponding master equation is derived. A hierarchy of equations for the reduced probability distributions is derived from the master equation. An equation similar to the Boltzmann equation for single particle probability distribution is derived using assumption of molecular chaos. It is shown that starting from an uncorrelated state, the system remains uncorrelated always in the limit $N\to \infty ,$ where $N$ is the number of particles. Simple applications of the formalism to direct simulation money games are given as examples to the formalism. The formalism is applied to the direct simulation of homogenous gases. It is shown that appropriately normalized single particle probability distribution satisfies the Boltzmann equation for simple gases and Wang Chang-Uhlenbeck equation for a mixture of molecular gases. As a consequence of this development we derive Birds no time counter algorithm. We extend the analysis to the inhomogenous gases and define a new direct simulation algorithm for this case. We show that single particle probability distribution satisfies the Boltzmann equation in our algorithm in the limit $% N\to \infty ,$ $V_{k}\to 0,$ $\Delta t\to 0$ where $% V_{k}$ is the volume $k^{th}$ cell. We also show that that our algorithm and Bird's algorithm approach each other in the limit $N_{k}\to \infty$ where $N_{k}$ is the number of particles in the volume $V_{k}$.