Latency Optimization for Heterogeneous Task Offloading in Cooperative MEC Network

In this paper, we consider a cooperative mobile edge computing (MEC) network where both the user equipments (UEs) and the MEC server can help with the task computing. Multiple heterogeneous tasks of different sizes for each UE are separately offloaded to the nearby UEs and MEC server's central processing unit and graphics processing unit for processing. Tasks of the same type are transmitted sequentially so that the waiting latency is required for offloading transmission and computation. We aim to minimize the maximum latency of UEs for processing all the tasks while ensuring that all tasks are successfully transmitted and processed. To solve the formulated non-convex problem, an iterative algorithm named directed mutation process based on discrete differential evolution is proposed. Simulation results are presented to verify the performance gain provided by the proposed algorithm.