Multi-spacecraft hunting and pursuit game decision-making method based on group independent learning strategy

The invention discloses a multi-spacecraft hunting and pursuit game decision-making method based on a group independent learning strategy, and the method specifically comprises the following main processes: building a multi-spacecraft hunting and pursuit game optimization mathematical model by taking speed pulse as a basic strategy of pursuit and pursuit parties; an intelligent learning algorithm is designed based on a near-end strategy optimization framework, and on this basis, three decision capabilities of pulse size selection, behavior switching and task allocation are fused; designing a group-independent basic game behavior set, and establishing a reward function model taking a behavior reward as a core; and three auxiliary game mechanisms of acceleration and deceleration, semi-mandatory behavior switching and dynamic task allocation are designed. The algorithm provided by the invention is guided by a bottom layer simple behavior, compared with a traditional intelligent learning strategy based on a termi