仿真网格中资源分配的进化博弈研究

Resource Allocation Based on Evolutionary Game in Simulation Grid

网格环境中用户的有限理性使得资源分配不能直接实现纳什均衡。针对网格用户理性的局限性,提出了一种基于进化博弈的网格资源分配策略,从动态角度来考察网格用户群体行为的演化过程。该策略先将网格环境中多用户竞争同一计算资源的问题形式化为一个多人博弈;然后建立了用户之间的进化博弈模型,利用复制动态方程求得用户出价博弈的进化稳定策略;最后研究了不同情况下用户群体的进化稳定点和评估函数特性。仿真实验结果表明进化博弈方法能够通过反复博弈使得网格用户不断学习并调整策略,逐步达到进化稳定均衡,从而实现了网格资源的优化分配。

Due to bounded rationality of users, resource allocation could not arrive at Nash equilibrium directly in grid environment. To address this issue, a grid resource allocation strategy based on evolutionary game was proposed to investigate the evolutionary process of grid user colony from the dynamic viewpoint. Firstly, the problem of multiple users competing for a common resource was formulated as a multi-player game. Secondly, an evolutionary game model of grid users was established. Using the method of replicated dynamics, an evolutionary stable strategy was produced to allocate resource. Finally, the evolutionary stable point and evaluation functions characteristics of grid users were discussed under different conditions. The experimental results show that evolutionary game approach makes grid users study and adjust strategy constantly through repeated games to achieve evolutionary stable equilibrium, which leads to an optimal allocation of grid resource.