AJEGWO: A HYBRID METAHEURISTIC FOR JOINT TASK OFFLOADING AND RESOURCE MANAGEMENT IN MEC

In mobile edge computing (MEC), efficient task offloading and resource allocation is more challenging and of great concern, particularly for resource-constrained user devices (UDs) with diverse workloads. The underlying joint task offloading and resource allocation problem is known to be NP-hard due to its combinatorial nature and complex coupling of variables. Thus, this paper introduces a novel hybrid metaheuristic algorithm, Adaptive Jaya Embedded GWO (AJEGWO), for jointly optimizing computation decision and resource allocation in MEC environments. The AJEGWO is based on integrating the leadership, and exploration capabilities of Grey Wolf Optimization into the adaptive JAYA algorithm, along with momentum and Lévy flight strategies to enhance global search, enables AJEGWO to achieve improved convergence and diversity. The approach aims to minimize a weighted system cost considering both energy consumption and latency, subject to user and system constraints. Simulation results show that AJEGWO outperforms baseline algorithms, achieving robust convergence regardless of initial conditions, and significantly reducing average latency (by over 95% and 85% compared to Jaya and GWO, respectively) as well as energy consumption (by over 60% versus both benchmarks).