A mixed Nash equilibrium solution for visibility-based pursuit-evasion game with multiple obstacles

Bu, Shaoming; Liang, Li; Wang, Youqing

Sci China Inf Sci, 2024, 67(9): 194201

This study primarily investigates a visibility-based pursuitevasion game with multiple obstacles. The pursuer aims to maintain surveillance over the evader as long as possible, while the evader’s objective is to escape the pursuer’s surveillance as quickly as possible. We use the explicit strategy method to analyze the winning conditions and optimal strategy of the player in the case of corner obstacles and propose a "Present-Future-Past" (PFP) loop structure to extend the strategy to the case of multiple obstacles. The action value functions are designed to evaluate various strategies of pursuer and evader, and the optimal strategy is determined through mixed Nash equilibrium theory. This method has high scalability and can be used to solve decision problems in multi-agent and multi-task scenarios.

PDF