协同空战任务决策的一个关键是编队内部的协调。基于复杂系统的一般求解方法,结合人工智能领域的MAS技术,提出了一种任务协调的分步实现方法。首先,就复杂的协同作战任务提出了一种分级分解思想,通过6级分解,将一项复杂作战任务分解为单个作战Agent的阶段性战术级任务,实现了协同作战任务的细化和量化,为各Agent协同完成复杂任务奠定了基础。在任务分解的基础上,建立了多Agent任务协调的数学模型,在合同网协议的基础上,提出了一种基于市场行为的编队内部的任务的动态协调机制。各作战Agent通过市场协议动态地选择适合自己的作战任务(包括协作性的和非协作性的),以编队总体效用最大化为目标,高效地实现了任务的动态、分布式的分配。仿真结果表明,提出的基于任务分解的协同任务协调机制效果较好,对作战任务决策起到了基础性的作用。

The task coordination within the aircraft formation is critical in decision-making of cooperative air combat. A multi-stage decision algorithm is present based on the MAS in artificial intelligence and the general algorithm for complex problem solving. First,a method is proposed to decompose the complex tasks into hierarchical tasks with 6 levels,and thus a complex combat task can be decomposed into tactical-level task of single Agent,and the cooperation of the task will be treated easily and obviously. Based on the task decomposition,a multi-Agent task coordination model is built up. A market behavior based task coordination mechanism used within the aircraft formation is proposed by using contract net protocol. Any aircraft in the formation can select its suitable tactical mission (including the cooperative and the non-cooperative mission) dynamically for the purpose of optimizing the total utility under the market protocol. The dynamic and distributed task assignment is thus realized effectively. Simulation result showed that the task coordination method present here is efficient for the decision-making of air combats.