分布式卫星系统递归式任务分配机制研究

Research on recursive task allocation mechanism of distributed satellite system

针对分布式卫星系统(DSS)能量约束实际,分析了当前研究中侧重于执行成本、忽视协调成本的问题,结合DSS节点空间分布特点,提出了DSS的网络拓扑模型;以此为基础,对基于MAS的合同网(CNP)任务分配机制进行了扩展,提出了递归式分配机制思路,并对任务分配流程进行了设计和优化。该机制能够避免因单点故障引发的分配故障,同时为任务协调成本和任务执行成本有机结合提供了有力支撑。通过定性和定量相结合的方式证明了递归式机制比合同网机制在协调成本上有较大的优势,为后续任务规划深入研究打下了良好的基础。

In the case of realistic energy constraints of distributed satellite system（ DSS）,this paper presented the shortcomings of current research on task allocation of DSS. Current research more focused on task implement cost and little attention was paid to negotiation cost. This paper proposed the network topology model combined with spatial distribution characters of DSS. And it made improvement on task allocation of contract net protocol（CNP） based on multi-agents system and put forward recursive task allocation idea. Then it designed and optimized process of recursive task allocation in detail. This mechanism can avoid task allocation breakdown by single point failure and provide the powerful support for the foundation of combination task negotiation cost with task implement cost naturally. The recursive mechanism has the obvious advantage over the CNP mechanism in negotiation cost through qualitative analysis and quantitative analysis and lays the helpful foundation of task planning later.