基于遗传算法的多代理规划优化方法  

The optimized method based on genetic algorithm for multi-agent plan

作　　者：李献业[1] 钟绍春[2] 

机构地区：[1]长春科技大学综合信息矿产预测研究所,吉林长春130026 [2]东北师范大学海信软件开发中心,吉林长春130024

出　　处：《东北师大学报（自然科学版）》2000年第1期1-5,共5页Journal of Northeast Normal University(Natural Science Edition)

基　　金：国土资源部科技司行业基金项目!( 9596 2 6 );国家博士后管理委员会基金项目!(BN90 11- 0 0 1)

摘　　要：给出了多代理 (agent)规划间关系上的分类 ,并在此基础上提出了多个约束条件下多代理规划优化的算法 .算法充分考虑了串行规划的合并和其它方法所不能处理的并行及部分并行规划的合并 ,并能对不同类型的规划 ,在时间期限约束下 。The foundation to work out a reasonable and optimized plan is to optimize the plan which be worked out by agent for various works. The relations among multi-agent plan are classified ,and an optimized algorithm for multi-agent plan within the bounds of many conditions is put forward here. The algorithm consider both the mergence of series plan and the mergence of parallel or part-parallel plan. Using this algorithm , the set of mergence operation which meet quality demand and least cost can be gotten for different type plans within the bounds of time.

关 键 词：多代理系统 遗传算法 多代理规划 人工智能 优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]