基于MAS的分布式卫星系统任务规划研究

针对分布式卫星系统的规划与调度问题,提出了一种基于多Agent协商机制的分布式规划与调度方法.首先建立了分布式卫星系统的分层分布式多Agent控制结构,并详细设计了各个层次Agent的功能.其次,给出了卫星Agent之间基于合同网的交互协议,设计了基于优先级的招标任务选择策略以及基于任务动态插入的投标方法,并给出了实现全局优化的评标策略和目标函数.针对观测环境的动态情形,给出了相应的动态任务处理流程.最后通过一个卫星编队实例进行了验证.

Research on MAS-based Mission Planning for Distributed Satellite System

The planning and scheduling problem of distributed satellite system(DSS) was investigated,and a distributed planning and scheduling approach for DSS was proposed based on negotiation of multiple agent system(MAS).A layered distributed architecture of MAS for DSS was desinged,and the functions of agents for different layers was presented elaborately.An contract-net-based coordinate protocol between different satellite agents was proposed.A priority-based strategy was designed to select the bid mission and the bid approach based on dynamic task insertion was proposed.The appropriate evaluation strategy and corresponding objective function were given.Aiming at some dynamic observation situations,corresponding dynamic mission conversion processes were designed.A satellite formation example was presented to demonstrate the validity of the distributed planning and scheduling approach.