Pursuit-evasion strategies for teams of multiple agents with incomplete information

In this paper, we investigate search strategies for multi-player pursuit-evasion games, in which a team of pursuers try to detect and capture multiple intelligent evaders. These are games with incomplete information for which there is no complete theory of existence of optimal solutions. We present a number of heuristic pursuit strategies and compare their performance using capture time as the performance metric. We demonstrate that reducing sensing overlap between pursuers and avoiding over-assignment of pursuers to target locations improve the performance of pursuit policies, and then derive some conclusions for their further development.