Multi-Pursuer Multi-Evader Games Based on Dynamic Elimination Priorities of the Dominated Strategies

Multi-agent coordination mechanisms are considered as important tasks planning methods that allow the intelligent collaboration of the agents. In this paper, we propose a flexible task coordination mechanism based on game theoretic principles. The main objective of this approach is to provide an agents’ strategy meeting the expectations of a conditioned equilibrium. In other words, we have based on Iterated Elimination of Weakly Dominated strategy algorithm IEWDS with dynamic elimination priorities to obtain the strategy profile with the highest average reward and lowest dynamism degree. In order to reflect the improvement brought by this method, we applied it to the Pursuit-Evasion Problem in comparison with another game theoretic approach where the players choose to play the strategy provided by IEWDS with a static elimination priority. In the experimental part, we study the average capturing time of the evaders, the development of the reward acquisition during the execution of the tasks, and also the stability degree of the pursuit groups. The simulation results prove that the proposed approach improves the pursuit processing in relation to the integrality of the studied parameters.