云联盟中收益最优化算法-云计算论坛.pdfVIP

优先出版 计 算 机 应 用 研 究 第33 卷 云联盟中收益最优化算法* 1, 2 1 3 许利军 ，王付强 ，张小庆 (1 ．新乡学院 计算机与信息工程学院，河南 新乡 453003 ；2 ．武汉理工大学 计算机学院，武汉 430063 ；3．武汉轻 工大学 数学与计算机学院，武汉 430023) 摘 要：云联盟环境下，资源提供者旨在协作地向用户提供资源，以获取更多的经济收益。通过联盟博弈研究了云资源 提供者的群体合作行为，基于任务执行时间和执行代价最优化问题，定义了资源提供者组建资源联盟的依据，并设计了 两种任务调度算法PMTC 和TMPC 。同时，对联盟收益进行内部分配时，提出了基于Shapley 值的联盟收益分配算法， 该算法充分考虑了联盟成员对联盟总收益的贡献度，使得联盟总收益在其个体成员间的分配更加均衡与合理。算例分析 结果表明，云资源联盟不仅可以改进任务执行效率和资源提供者收益，而且Shapley 值可以实现收益的公平分配。 关键词：云计算；联盟博弈；任务调度；收益最优化 中图分类号：TP301.6 Profit optimization algorithm in cloud coalition 1, 2 1 3 Xu Lijun , Wang Fuqiang , Zhang Xiaoqing (1. School of Computer Information Engineering, Xinxiang University, Xinxiang Henan 453003, China; 2 School of Computer, Wuhan University of Technology, Wuhan 430063, China; 3. School of Mathematics Computer Science, Wuhan Polytechnic University, Wuhan 430023, China) Abstract: Cloud federation has been emerged as a new paradigm in which group of Cloud Resource Providers (CRPs) cooperate to provide resources for users, to gain more economic advantage. In this paper, we study the cooperative behavior of a group of CRPs by coalitional game. Based on the optimization of execution time and payoff, we define the basis of forming resource coalition. After building the coalition, for gaining optimal resource mappings, two algorithms PMTC and TMPC are proposed. Finally, for the profit sharing in coalitions, a sharing strategy of coalitional profit based on Shapley value is proposed, which could fully consider th