Distributed Task Offloading in Mobile Edge Computing using Metaheuristics
Ala’ A. Samarneh, Abdallah Y. Alma’aitah · 2024
Mobile Edge Computing (MEC) is emerging as a prominent technology that enables different sophisticated mobile applications. With the high transmission rates provided by 5G and beyond and the MEC technology, satisfying applications' strict delay requirements is possible. By providing high computation power closer to the users at the network edge, various applications such as smart cities, context awareness, and content distribution can provide users with a better Quality of Experience (QoE) using MEC. However, offloading all tasks to the MEC is inefficient, as transmitting tasks to the MEC drains the UE battery and might create congestion at the network edge. Therefore, the UE must determine the subset of tasks to offload to the MEC to minimize energy consumption and execution delays. In this work, we investigate a UE-driven offloading approach utilizing three metaheuristic algorithms: Hybrid Improved Whale Optimization Algorithm (HI-WOA), Tug of War Optimization (TWO), and Adaptive Equilibrium Optimization (AEO), and priority queues. We formulate the offloading decision as a constrained multi-objective optimization problem, where the objectives are to minimize the energy consumption and execution delay at the UEs. The HI-WOA, AEO, and TWO demonstrated robust performance through extensive simulations with an average Qualified Tasks Ratio (QTR) of 99.75%, 99%, and 94% respectively, and a significant reduction in the execution delays and energy consumptions under different workloads with varying requirements.