Collision Detection Optimization in a Multi-Particle System

Abstract

Collision detection optimization algorithms in an event-driven simulation of a multi-particle system is one of crucial tasks, determining efficiency of simulation. We employ dynamic computational geometry data structures as a tool for collision detection optimization. The data structures under consideration are the dynamic generalized Voronoi diagram, the regular spatial subdivision, the regular spatial tree and the set of segment trees. Methods are studies in a framework of a granular-type materials system. Guidelines for selecting the most appropriate collision detection optimization technique summarize the paper.