基于多智能体粒子群的协同空战目标决策研究

Target Decision in Collaborative Air Combats Using Multi-agent Particle Swarm Optimization

以多机多目标协同空战为研究背景,针对复杂多变的战场态势,结合敌我双方的实战约束条件和威胁评估函数,建立了一种能够体现火力攻击优先级的协同空战目标决策仿真模型。为了能够快速准确的求解决策方案,通过在基本粒子群算法中引入多智能体理论中的交互机制,并分别设计智能体邻域合作算子、变异算子和自学习算子,提出了一种改进的多智能体粒子群算法。仿真结果表明,该算法能够计算得到合理有效的决策方案,并且具有良好的仿真实时性。

Under the research background of collaborative multi-aircraft and multi-target air combats, combined with the actual combat constraint conditions and the threat assessment functions on both sides, a collaborative air combat target decision simulation model is established for complex and changeable battlefield situations, which can reflect the priority of fire attack. To solve the decision scheme quickly and accurately, an improved multi-agent particle swarm optimization algorithm is proposed by introducing the interaction mechanism of the multi-agent theory into particle swarm optimization algorithm; and the neighborhood cooperation operator, mutation operator and self-learning operator for the agent are designed respectively. The simulation results show that the method can work out a reasonable and effective decision scheme, and has a good real-time simulating performance.