Real-time adaptive A*

  • Authors:

  • Sven Koenig;Maxim Likhachev

  • Affiliations:

  • University of Southern California, Los Angeles, CA;Carnegie Mellon University, Pittsburgh, PA

  • Venue:
  • AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Characters in real-time computer games need to move smoothly and thus need to search in real time. In this paper, we describe a simple but powerful way of speeding up repeated A* searches with the same goal states, namely by updating the heuristics between A* searches. We then use this technique to develop a novel real-time heuristic search method, called Real-Time Adaptive A*, which is able to choose its local search spaces in a fine-grained way. It updates the values of all states in its local search spaces and can do so very quickly. Our experimental results for characters in real-time computer games that need to move to given goal coordinates in unknown terrain demonstrate that this property allows Real-Time Adaptive A* to follow trajectories of smaller cost for given time limits per search episode than a recently proposed real-time heuristic search method [5] that is more difficult to implement.