Abstract:Rapid development of mobile communications has led to respectable latency-sensitive and computation-intensive mobile applications. There is a huge contradiction between high resource demands of these applications and limited resource of mobile devices. In this regard, mobile edge computing (MEC) is a promising technology, where computation tasks can be offloaded from mobile devices onto network edges with stronger capability. However, the dependency between tasks leads to high complexity for offloading decision. In this paper, we investigate the optimal offloading problem for completing dependency-aware tasks by minimizing the latency and energy cost. An improved non-dominated sorting genetic algorithm-II (INSGA-II) is proposed to solve this multiobjective problem. Simulation results validate the advantage of the proposed algorithm in terms of the performance of low latency and cost.

A Dependency-Aware Task Offloading Strategy in Mobile Edge Computing Based on Improved NSGA-II.