面向无人集群博弈对抗的多智能体分层决策框架OA

To address the issues of insufficient multi-agent collaborative decision-making capability and low decision-making credibility of unmanned swarm,a hierarchical decision-making framework is proposed.This framework integrates a high-level behavior decision tree with middle and low-level deep reinforcement learning models to optimize task sequence generation and combat action adjustment strate-gies.By facilitating coordinated planning and multi-objective collaborative control among multiple agents,it enhances the capabilities of global coordination and local adaptation.This method provides theoretical support for the practical application of command decision-making technologies for unmanned swarm.