Multi-goal path-finding for autonomous agents in virtual worlds

Abstract

We are interested in complex path-planning, where several goals have to be satisfied by an autonomous agent. The agent has to pass through several destination points in a virtual environment, minimizing some criterion (e.g., the length of the overall trajectory), we call this problem multi-goal pathfinding. The problem appears for artificial creatures in virtual environments and for computer-controlled characters in 2D or 3D games, which have to exhibit so-called opportunistic behaviors. We present a simple framework and an effective algorithm based on goal reordering using local search techniques. © 2003 by Springer Science+Business Media New York.