Parallel-Driven Edge Computing Task Offloading for Profit Maximization Based on DDPG

Leveraging mobile edge computing (MEC) for task offloading is an effective strategy to address the computational limitations of mobile devices. However, current offloading strategies largely cater to user-centric objectives, neglecting the motives of service providers, unintentionally diminishing their profits. In this work, we propose a novel allocation strategy to improve resource utilization efficiency based on a parallel edge computing system. Specifically, we jointly consider incentives and cross-server resource allocation in parallel-driven MEC. This approach facilitates the distribution of user workloads across multiple edge servers, enhancing system performance and efficiency. In our approach, we leverage the deep deterministic policy gradient (DDPG) algorithm to support task offloading decisions, maximizing the overall profits of service providers. Simulation results show that our approach, compared to non-parallel edge systems, efficiently boosts resource utilization by an average of 13.37%.