电信科学 ›› 2011, Vol. 27 ›› Issue (9): 42-47.doi: 10.3969/j.issn.1000-0801.2011.9.010

• 云计算专栏 • 上一篇    下一篇

一种基于间接互惠的计算网格合作激励机制研究

杨虎,张东戈,刘浩,白天松   

  1. 解放军理工大学指挥自动化学院 南京210007
  • 出版日期:2011-09-15 发布日期:2017-06-22
  • 基金资助:
    国家自然科学基金资助项目;江苏省自然科学基金资助项目

A Cooperation Incentive Mechanism for Computational Grid Based on Indirect Reciprocity

Hu Yang,Dongge Zhang,Hao Liu,Tiansong Bai   

  1. Institute of Command Automation,PLA University of Science and Technology,Nanjing 210007,China
  • Online:2011-09-15 Published:2017-06-22

摘要:

基于间接互惠理论,构建了计算网格合作激励机制,建立了三维声誉系统以及声誉管理机制,对计算网格中资源拥有者的“合作行为”进行有效激励,并对“拒绝合作行为”中的“恶意拒绝合作”和“偶发拒绝合作”进行了区分。通过对恶意拒绝合作行为进行严厉的惩罚,提升合作稳定性,同时对偶发的拒绝合作行为采取宽容策略,以避免偶发的拒绝合作陷入最坏的“一报还一报”结果。最后,利用合作博弈理论,分析得出在激励机制下,资源拥有者的最优策略是最大可能地共享闲置资源以及不间断合作,以提升自己的合作诚信度,从而验证了激励机制的有效性。

关键词: 计算网格, 间接互惠, 合作博弈, 激励机制

Abstract:

Taking computational grid as an evolutionary cooperating game network,an incentive mechanism is designed for better cooperation based on indirect reciprocity theory within computational grid. In order to encourage the resources owner to act as a cooperator in the computational grid,a three-dimensional reputation system and a reputation management mechanism are designed which can be used to distinguish “accidental deny of cooperation”from “deliberate deny of cooperation”. The“deliberate deny of cooperation”deed is punished rigorously to stabilize cooperative activity,and at the same time,the tolerant strategy is used to deal with the “accidental deny of cooperation”deed to avoid engaging in the worst “tit-for-tat”strategy. According to cooperative game theory,we find out that under the cooperation incentive mechanism,the dominant strategy of the resources owner is trying its best to share the spare resources and continuously performing as a cooperator to increase its cooperation credit value,which demonstratively shows that the incentive mechanism provides an effective result.

Key words: computational grid, indirect reciprocity, cooperative game, incentive mechanism

No Suggested Reading articles found!