自利Agent追捕联盟生成算法  

Self-interested Agent pursuit alliance generation algorithm

作　　者：吕磊[1] 王浩[1] 林航[2] 方宝富[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009 [2]国网安徽省电力公司信息通信分公司,安徽合肥230061

出　　处：《合肥工业大学学报（自然科学版）》2017年第6期769-774,828,共7页Journal of Hefei University of Technology：Natural Science

基　　金：国家高技术研究发展计划(863计划)资助项目(2012AA011005);国家自然科学基金资助项目(61175051;61175033;61203360);安徽省自然科学基金资助项目(1308085QF108);合肥工业大学博士学位人员专项基金资助项目(JZ2014HGBZ0014)

摘　　要：多Agent合作追捕是多Agent系统研究的经典问题,在机器人等领域具有重要的应用前景。文章提出了面向任务的自利Agent联盟生成算法,该算法能同时处理多个不同类型的逃跑Agent的任务分配问题;因为追捕Agent和逃跑Agent速度相等,追捕联盟成员位置的分布十分关键,所以提出了基于贡献度的联盟成员选择策略;同时为了较好地体现追捕Agent的自利性,定义了需求度作为自利性的度量,解决了冲突协商且有利于资源的优化配置。通过与经典拍卖算法的比较表明,该算法显著提高了追捕成功率。Multi-Agent cooperative pursuit is a classic problem of multi-Agent systems, which has important applications in the field of robotics. The alliance generation algorithm of task-oriented self-interested Agent is proposed. The algorithm can handle the task allocation problem of many different types of escape Agents. For the pursuit Agent and escape Agent have equal speed, the position distri- bution of alliance members is critical. Therefore, an alliance member selection strategy based on con- tribution degree is put forward. In order to better reflect the self-interest of pursuit Agent, need de- gree is defined as a measure of self-interest, which can deal with the conflict negotiation and promote the most optimum allocation of resources. Compared with the classic auction algorithm, the proposed algorithm significantly improves the success rate of the pursuit.

关 键 词：追捕-逃跑问题 自利Agent 追捕联盟 贡献度 需求度 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]