Joint Optimization of Latency and Reward for Offloading Dependent Tasks in Mobile Edge Computing
Yanqi Gong, Fei Hao, Yifei Sun, Longjiang Guo · 2021
In the 5G era with the explosive growth of data, offloading tasks to edge servers that are executed closer to it becomes one of the most popular computational paradigms. Different from cloud computing, mobile edge computing (MEC) significantly addresses the problem of latency-sensitive applications execution, such as online gaming and VR/AR applications. In order to improve the quality of experience of end-users, those applications are often divided into multiple tasks with dependencies. Regarding the problem of offloading tasks with dependencies, the existing researches focus on either latency or reward optimization that leads to the practical difficulty of this problem. Towards this end, this paper first introduces a novel metric reward per unit time which integrates the latency and reward for better optimization of tasks offloading strategy. Then, the reward per unit time is viewed as the objective function, and further addressing the above problem is equivalent to finding the optimal offloading strategy to maximize the value of the objective function. The simulation experiments are conducted for demonstrating that the proposed offloading strategy is feasible and effective.