基于博弈的云计算任务分解研究

Study of Cloud Computing Task Factoring Based on Game Theory

云用户在同时拥有本地资源和云计算资源时,如何在二者之间进行任务分解,以最短化任务完成时间是云用户的一个优化决策问题。研究云用户的最优任务分解决策。针对云计算资源的共享特征,用户之间的决策会影响其余用户的任务完成时间;因此通过求解纳什均衡的方法来分析用户的策略行为,从而决定用户的最优决策。求解了大任务和小任务两类用户时的纳什均衡,并通过仿真给出了在不同大小任务量比、云资源本地资源速度比和大任务用户概率时云用户的最优决策。

Task factoring to shortest of task completion time among local resources and cloud computing resources is the cloud users optimized decision problem, if cloud users can use the aforementioned two resources at the same time, and this paper studies cloud users optimal task factoring decision. Because of the sharing characters of cloud computing resources, users decision-making will affect the task completion time of the rest users, and so solving the Nash equilibrium strategy to analyze the users’ behavior and to determine the optimal decision-making are proposed. Nash equilibrium is solved when the two types of users with big tasks and small tasks are adopted, and the cloud users’ optimal decision is proposed through simulation when different ratio of big and small tasks’ workload, different ratio of local and cloud execution speed, and different probability of big task users are given respectively.