Logistics supply simulation based on multi-agent cooperation

To achieve automatic logistics supply for the front combat troops in a combat simulation, a multi-agent cooperative simulation model is constructed, and a multi-agent cooperative reinforcement learning algorithm constrained by a common rule is proposed. The algorithm take the multi-agent cooperative learning process as discrete stage games, force all the logistics agents in the system selecting their movements in the principle of maximizing logistics supply efficiency, and solve the problem of equilibrium inconsistent in the game so that the whole army maximizing combat effectiveness. Moreover, the proposed algorithm measures the contribution of each agent to the whole system based on the common rule to solve the credit assignment problem. The experimental results of simulation show the effectiveness of the proposed algorithm.