基于博弈的多用户部分卸载策略研究

Research of Multi-User Partial Offloading Stragety Based on Game

部分卸载通过将任务进行拆分并行运行在不同计算节点以降低时延与能耗。现使用部分卸载的方案,同时将任务部分卸载在两个计算节点。由于每一移动用户计算任务时延与能耗受其他用户卸载决策影响,将系统卸载决策问题描述为一个多玩家博弈。此外证明了上述博弈纳什均衡存在,表明系统可以收敛到一个稳定的状态。为了达到纳什均衡,设计了一种基于博弈的多用户部分卸载方案。最大量仿真表明,在代价和效用两方面上述方案均能获得更好的表现效果。

Partial offloading reduces time delay and energy consumption by spliting tasks and running them on different computing nodes in parallel.This article uses a partial offloading scheme,and partially offloads tasks on two computing nodes at the same time.Since the latency and energy consumption of each mobile user's computing task are affected by the offloading decisions of other users,we described the system offloading decision problem as a multiplayer game.In addition,we proved that the game Nash equilibrium exists,indicating that the system can converge to a stable state.In order to achieve the Nash equilibrium,we designed a game-based multi-user partial offloading scheme.Finally,a large number of simulation experiments show that our scheme can achieve better performance in terms of cost and utility.