Min-Max Worst-Case Design for Computation Offloading in Multi-user MEC System

Mobile-edge computing (MEC) has been recognized as a promising technique to provide wireless user equipments (UEs) with enhanced computation capability. In this paper, we consider an MEC cellular system, which consists of a number of base stations (BSs) and UEs. We suppose that each BS is equipped with an MEC server which offers computation offloading service for UEs. Considering the fairness among UEs in task execution, we formulate the computation offloading problem as a min-max worst-case design problem that minimizes the maximal task execution time among all the UEs. To solve the optimization problem, we consider both single UE case and multiple UEs case. For single UE case, we design the analytical task partition and computation offloading strategy and also solve the optimization problem via convex optimization tools. For multiple UEs case, to tackle the coupling among multiple UEs, we propose a heuristic computation offloading algorithm which designs the computation offloading strategy for UEs sequentially. Numerical results demonstrate the effectiveness of the proposed algorithm.