Reinforcement Learning-based Genetic Algorithm for Computation Offloading Optimization in UAV-assisted MEC Systems

UAV-assisted Mobile Edge Computing (MEC) improves IoT wireless systems but struggles with increasing computational demands, resource limitations, and stringent latency requirements. This work optimizes computation offloading (OLMEC) in UAV-assisted MEC using evolutionary algorithms. UAVs provide offloading services to user equipment (UE), balancing local and offloaded task execution. The goal is to minimize delays while adhering to latency and energy constraints. Two solutions are proposed: OMEA, a fine-tuned evolutionary algorithm, and OMEDR, which integrates evolutionary algorithms with deep reinforcement learning. Experimental results demonstrate the superiority of the proposed methods over existing approaches, highlighting their effectiveness in addressing the OLMEC problem.