Improved heuristics for optimal pathfinding on game maps

作者: Yngvi Björnsson , Kári Halldórsson

DOI:

关键词:

摘要: As computer game worlds get more elaborate the visible pathfinding performance bottlenecks become. The heuristic functions typically used for guiding A*-based are too simplistic to provide search with necessary guidance in such large and complex worlds. This may result A*-search exploring entire map order find a path between two distant locations. This article presents effective heuristics estimating distances locations maps. former, dead-end heuristic, eliminates from areas that provably irrelevant current query, whereas second uses so-called gateways improve its estimates. Empirical evaluation on actual maps shows both reduce exploration time complexity of A* significantly over standard octile distance metric.

参考文章(3)
Robert C. Holte, R. M. Zimmer, A. J. MacDonald, M. B. Perez, Hierarchical A *: searching abstraction hierarchies efficiently national conference on artificial intelligence. pp. 530- 535 ,(1996)
Peter Hart, Nils Nilsson, Bertram Raphael, A Formal Basis for the Heuristic Determination of Minimum Cost Paths IEEE Transactions on Systems Science and Cybernetics. ,vol. 4, pp. 100- 107 ,(1968) , 10.1109/TSSC.1968.300136
Martin Müller, Adi Botea, Jonathan Schaeffer, Near Optimal Hierarchical Path-Finding. J. Game Dev.. ,vol. 1, pp. 1- 30 ,(2004)