对抗环境下多Agent协商问题的研究  

RESEARCH ON MULTI-AGENT NEGOTIATION ISSUES IN AN ANTAGONIZED ENVIRONMENT

作　　者：申静[1] 

机构地区：[1]陕西理工学院数学与计算机科学学院,陕西汉中723000

出　　处：《计算机应用与软件》2014年第1期165-169,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61301237);陕西省自然科学基金项目(2011JE012);陕西省教育厅专项科研基金项目(2013JK1195)

摘　　要：利用多Agent系统具有的自治性和实时反应性,探讨对抗环境下的多Agent协商决策问题,提出一种混合式的多Agent结构协商模型,给出以最大团队效益为前提的协商求解策略和协商角色交换算法。通过协商,对抗环境中的Agent成员能够很好地进行动作策略选择和移动,能更好地进行进攻和防守。仿真实验验证了算法的可行性和有效性,结果表明其在一定程度上解决了多Agent系统中实时动态和受限通信对抗环境下的多Agent决策与合作问题。By taking advantage of the intrinsic autonomy and real-time reaction of the Muhi-Agent Systems, the paper discusses the issues of Multi-Agent negotiation and decision making in the antagonized environment and proposes a hybrid multi-agent structure negotiation model. By taking a team＇s maximum benefits as its premise, it proposes a negotiation solving strategy and a negotiation role exchange algorithm. Through negotiation, the Agents within the group in an antagonized environment can perfectly execute action strategy selection and movement to carry out better offensive or defensive actions. Simulation experiments verify the feasibility and effectiveness of the algorithm. Results indi- cate that it solves, to some extent, the Multi-Agent decision making and cooperation problem of the Multi-Agent Systems in a real-time, dy- namic and limited communicative antagonized environment.

关 键 词：对抗 协商 多AGENT系统 策略 仿真 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]