Priority-Based SAC and Experience Replay for Adaptive Multi-Server Task Offloading in Mobile Edge Computing

With the rapid growth of mobile application demands, Mobile Edge Computing (MEC) has emerged as a key technology to address the resource constraints of mobile devices. To tackle the challenge of priority-aware resource allocation in multi-server environments, this paper innovatively proposes a Priority Experience Replay (PER)-enhanced Soft Actor-Critic (SAC) reinforcement learning algorithm, PER-SAC. This approach comprehensively considers task priorities, computational resources, and network conditions to intelligently optimize task offloading decisions, significantly improving system efficiency. Experimental results demonstrate that the PER-SAC algorithm outperforms conventional methods in terms of convergence speed, stability, and adaptability, achieving a 40% increase in cumulative rewards and reducing overall costs by 18%-23%. Particularly suitable for priority-sensitive MEC scenarios, this work provides a novel solution for building efficient and reliable edge computing systems.