摘要
针对多智能体系统(MAS)任务分配问题中多个任务与MAS两者的分布式特征,将任务分配问题形式化为分布式约束满足问题(DCSP)进行求解,分别建立了以任务为中心和以agent为中心两种MAS任务分配模型,基于改进的DCSP分布式并行求解算法,提出了基于DCSP的MAS任务分配问题求解框架。该方法适合求解agent间通信有随机延迟以及agent间存在多约束的问题,应用实例的求解表明了其实用性与有效性。
The problem of multi-agent system task allocation was studied considering it to be formed into a distributed constraint satisfaction problem.Both multiple tasks and MAS were distributed,they could build task-centered and agent-centered two kinds of constraint network construction model.Utilized the improved ConcDB algorithm,provided a framework of multi-agent task allocation which based on DCSP.This approach adapts to solution the problem when agents are in uncertain environment or there are a lot of constr...
出处
《计算机应用研究》
CSCD
北大核心
2009年第2期515-517,520,共4页
Application Research of Computers
基金
航空科学基金资助项目(20075185013)
关键词
分布式约束满足问题
多智能体系统
任务分配
并行动态回溯算法
求解框架
DCSP（distributed constraint satisfaction problem）
MAS（multi-agent system）
task allocation
ConcDB（concurrent dynamic backtracking）algorithm
solution frame
