Search-Space Characterization for Real-time Heuristic Search

TL;DR

This paper introduces complexity measures for search spaces and demonstrates their effectiveness in predicting the performance of real-time heuristic search algorithms across various domains, including video games and beyond.

Contribution

It develops and validates algorithm-independent complexity measures that predict real-time heuristic search performance across diverse search spaces.

Findings

Complexity measures significantly predict algorithm performance.

Performance varies across different search space types.

Database-driven analysis extends understanding beyond video games.

Abstract

Recent real-time heuristic search algorithms have demonstrated outstanding performance in video-game pathfinding. However, their applications have been thus far limited to that domain. We proceed with the aim of facilitating wider applications of real-time search by fostering a greater understanding of the performance of recent algorithms. We first introduce eight algorithm-independent complexity measures for search spaces and correlate their values with algorithm performance. The complexity measures are statistically shown to be significant predictors of algorithm performance across a set of commercial video-game maps. We then extend this analysis to a wider variety of search spaces in the first application of database-driven real-time search to domains outside of video-game pathfinding. In doing so, we gain insight into algorithm performance and possible enhancement as well as into search space complexity.