基于合并与分裂机制的云任务联盟博弈调度  

Cloud tasks coalitional game scheduling based on merge and split mechanism

作　　者：薛凡 吴志健[1] 

机构地区：[1]武汉大学计算机学院软件工程国家重点实验室,湖北武汉430072 [2]黄淮学院国际教育学院,河南驻马店463000

出　　处：《计算机工程与设计》2018年第1期72-78,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61364025;61402481);武汉大学软件工程国家重点实验室开放基金项目(SKLSE2014-10-4)

摘　　要：为降低任务执行代价,提高云资源提供者收益,提出一种云计算环境下基于合并与分裂机制的任务调度联盟博弈优化算法。建立基于资源联盟的任务执行模型,定义资源提供者选择建立联盟的依据;给出联盟完成云任务后总收益在其联盟成员间的分配方式,重点阐述如何通过联盟合并与分裂机制产生稳定的联盟结构,分析结盟算法的性能。通过仿真实验,与同类型算法进行性能比较,比较结果表明,联盟博弈算法可以产生稳定的联盟结构,为联盟中的个体成员带来最大化收益。To reduce the cott of tasks execution and increase the profit of cloud resource providers , a resource provision algorithm based on coalitional game md merge and split mechanism in cloud computing environment was proposed. The tasks executionmodel based on resource coalition was established , the basis of forming coalition for resource providtion method of total profit among coalitional members was provided after the coalition completed all tasks , how to generate stable coalitional structure through coalitional merge and split mechanism was focused and the performance rithm was analyzed. The proposed algorithm was compared with the same type of algorithms by simulation experiments. The re-sults show that the proposed coalition game algorithm can not only generate stable coalition structuhest profit for coalitional members.

关 键 词：云计算 资源提供 任务调度 联盟博弈 合并与分裂 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]