基于缓存优化的移动边缘计算资源分配策略

Resource allocation strategy for mobile edge computing based on cache optimization

移动边缘计算研究中,边缘服务器通过缓存任务数据可以有效节约计算资源,但如何分配缓存资源解决边缘服务器的竞争关系,以及能耗和效益问题,达到系统性能最优是一个NP难问题。为此提出基于缓存优化的在线势博弈资源分配策略OPSCO(online potential-game strategy based on cache optimization),采用新的缓存替换策略CASCU(cache allocation strategy based on cache utility),最大化缓存的效用。通过优化边缘服务器的效益指示函数,将缓存替换代价等因素与李雅普诺夫优化、势博弈以及EWA(exponential weighting algorithm)算法结合,对边缘服务器的竞争关系建模,进行势博弈相关证明和分析。仿真结果表明,OPSCO相比于其他资源分配策略,可以明显提升任务完成率和缓存效用,并降低设备能耗和时间开销,解决了移动边缘计算在线缓存场景中的资源分配以及数据缓存问题。

In mobile edge computing research,the edge server can effectively save computing resources by caching task data.But how to allocate cache resources to solve the competitive relationship between edge servers,as well as energy consumption and efficiency issues,and achieve optimal system performance is a NP-hard problem.Therefore,this paper proposed an online potential-game resource allocation strategy OPSCO based on cache optimization,using a new cache replacement strategy CASCU to maximize the utility of the cache.By optimizing the efficiency indicator function of edge servers,combining factors such as cache replacement cost with Lyapunov optimization,potential game,and EWA algorithm,it modeled the competitive relationship of edge servers,and conducted potential game related proofs and analyses.The simulation results show that compared with other resource allocation strategies,OPSCO can significantly improve the task completion rate and cache utility,reduce equipment energy consumption and time overhead,and solve the resource allocation and data cache problems in mobile edge computing online cache scenarios.