合作博弈的自私网格负载均衡的建模和仿真

针对自私计算网格的负载均衡问题,构建了符合其特征的自私网格数学模型和外来作业计算成本模型。利用合作博弈论分析该模型,验证了负载均衡为全局最优策略,并得出了多集群自私网格进行负载均衡的效果。由此提出了基于合作博弈的有界迭代负载均衡算法,解决异构负载集群在自私网格负载均衡中的公平性问题。最后,通过实验仿真验证了方案的有效性与实用性。

Abstract：

Aiming at the problem of load balancing in selfish computational grids, a mathematical model for selfish grids and a cost function for execution of foreign jobs were established. The cooperative game-theoretic framework was used to analyze the model in order to show that load balancing is a global-optimal strategy for grids and it is profitable to collaborate. Then, a bounded integrative load balancing algorithm was proposed which balances the load more equitably in multi-cluster heterogeneous load selfish grids. Finally, the results are assessed experimentally by simulation.