华南理工大学学报(自然科学版) ›› 2011, Vol. 39 ›› Issue (12): 82-89.

• 计算机科学与技术 • 上一篇    下一篇

DRQSHCTAgent 支持的开放环境下的可信协同

周尤明 高济 傅朝阳   

  1. 浙江大学 计算机科学与技术学院,浙江 杭州 310027
  • 收稿日期:2011-06-03 修回日期:2011-09-13 出版日期:2011-12-25 发布日期:2011-11-04
  • 通信作者: 周尤明(1972-) ,男,博士,讲师,主要从事人工智能、多Agent 系统和协同设计等的研究. E-mail:monitorstudent@126.com
  • 作者简介:周尤明(1972-) ,男,博士,讲师,主要从事人工智能、多Agent 系统和协同设计等的研究.
  • 基金资助:

    国家"863”计划项目( 2007AA01Z187) ; 国家自然科学基金资助项目( 60773177)

DRQSHCT Agent-Supported Trust Cooperation in Open Environments

Zhou You-ming  Gao Ji  Fu Zhao-yang   

  1. College of Computer Science and Technology,Zhejiang University,Hangzhou 310027,Zhejiang,China
  • Received:2011-06-03 Revised:2011-09-13 Online:2011-12-25 Published:2011-11-04
  • Contact: 周尤明(1972-) ,男,博士,讲师,主要从事人工智能、多Agent 系统和协同设计等的研究. E-mail:monitorstudent@126.com
  • About author:周尤明(1972-) ,男,博士,讲师,主要从事人工智能、多Agent 系统和协同设计等的研究.
  • Supported by:

    国家"863”计划项目( 2007AA01Z187) ; 国家自然科学基金资助项目( 60773177)

摘要: 为降低开放环境下因Agent 违约带来的风险,提出了DRQSHCTAgent 模型.该模型引入服务合成和处方产生任务的多个求解方案,通过品行信任度和服务信任度定义直接信任,通过定义多个服务信任度克服服务信任度的掩盖性,通过群体推荐克服个体链推荐的不足,通过计算方案信任度来选择并优化求解方案.模型通过合同样本提高了计算信任度的合理性和可操作性,同时由于引入了基于合同的约束,Agent 的自组织协作既是可信的,又具有较高的可靠性.设计实例和实验表明,该模型更适合开放环境下的可信协同问题求解.

关键词: DRQSHCTAgent, 可信协同, Agent 信任, 合同, 多Agent 系统

Abstract:

Proposed in this paper is a DRQSHCTAgent model to reduce the risks caused by agents’contract violations. In this model,multiple schemes of the task are obtained by introducing service compositions and recipes; a direct credit is defined in terms of the moral and the service credits; the concealing of the service credit is uncovered by defining multiple service credits; the shortcomings of individual chain recommendation are overcome by the group recommendation; and the solving schemes are selected and optimized by calculating the scheme credit. The proposed model improves the rationality and maneuverability of the credit calculation with the help of contract samples,and ensures the self-organizing cooperation credibility and the system reliability due to the contract-based constraints. Both the design instance and the experiment indicate that the proposed model is more suitable for the resolution of trust cooperation problem in open environments.

Key words: DRQSHCTAgent, trust cooperation, Agent credit, contract, multi-Agent system