基于群集智能的协同多目标攻击空战决策

Air Combat Decision Making for Coordinated Multiple Target Attack Using Collective Intelligence

为使各武器平台利用数据链(TADIL)交换和共享信息,自主协同计算实现多目标攻击决策,本文基于NASA/Ames研究中心提出的群集智能(COIN)理论,构建以作战单元为智能体(Agent)的协同多目标攻击决策模型,并通过定义Agent的贡献度扩展群集智能架构,用于解决分布式异构Agent造成的系统难以准确收敛问题。为提高收敛速度,减少求解过程中的盲目搜索和提高已形成解的收敛,研究加入了构造性启发和改进性启发两类启发方法。实验表明所提算法同传统算法相比,具有更好的稳定性和扩展性,减少了计算量,收敛速度也有一定的提高。

In modern air combat, it is an important yet difficult task to coordinate multi-fighter to make multitarget attack decisions by using the tactical digital information links （TADILs） to share and exchange combat information. To address the problem, a novel distributed heterogeneous multi-target air combat decision-making model is proposed based on the collective intelligence （COIN） theory put forward by the NASA/Ames Research Center. In this model, each missile is defined as an agent and the conventional COIN framework is expanded by agent contribution rate to realize accurate convergence under a distributed heterogeneous agent environment. In order to improve the convergence rate, two classes of heuristic information is added to the algorithm, i. e. , constructive heuristics and improvement heuristics, to reduce blind search in intermediate results and enhance convergence rate in ultimate results. Simulation results show that as compared with conventional methods, the proposed method is able to converge to the global optimum with more stability and higher expandability, decrease the amount of computation and improve convergence rate.