Multi-agent Task Assignment Based on the Situation Field and Game Theory

In the striking phase of modern warfare, commanders’ decision-making is of extremely high complexity, which calls for the development of human-machine hybrid decision-making techniques. In this study, we propose a multi-agent task assignment method based on the situation field and game theory for the human-machine hybrid decision-making system. The problem is firstly modeled as a multi-objective game problem after situation assessment. Then the differential evolution algorithm is utilized to obtain the optimal assigning suggestions for commanders to choose from. Simulations and experiments are conducted to demonstrate the real-time performance and practicability of the proposed method. The winning rates of all the strategy suggestions exceed 55% when the two sides are comparable. Due to the innovative incorporation of the situation field and game theory, the proposed method enables a comprehensive understanding of the environmental situation and the adversarial nature of the striking phase. The method also achieves a sensible optimization of multi-agent task assignment for multiple objectives to combine human and machine intelligence.