Multi-policy Aware Offloading with Per-task Delay for Mobile Edge Computing Networks

Abstract

Mobile Edge Computing (MEC) is a promising new technology that offers new opportunities for energy consumption optimization, privacy preservation, and network traffic bottlenecks” reduction. Besides, MEC-based computation tasks offloading can achieve lower latencies and energy consumption. However, with the multi-task multi-user setting, the offloading decisions become hard and critical. Indeed, the communication and processing resources as well as the resulting processing delays and the consumed energies have to be carefully considered. In this paper, we consider a multi-policy offloading scenario where each mobile device holds a list of heavy tasks. Each task is further characterized by its proper processing deadline. Therefore, we designed the corresponding optimization problem that minimizes a weighted-sum function that jointly considers energy consumption, processing delays, and the unsatisfied tasks' workloads. Due to the short decision time constraint in the studied system and the NP-hardness of the obtained problem, we decomposed it using two sub-problems. Then, we proposed a solution to each sub-problem. With the aim of evaluating these solutions., we performed a set of simulation experiments to compare their performance with relevant state of the art method. Finally., the obtained execution times are very satisfactory for moderate number of tasks.