Computation Resource Management in Mobile Edge Computing for Healthcare Using Lyapunov-Deep Deterministic Policy Gradient

Qiang He, Yang Xia, Zheng Feng, Lianbo Ma, Yingjie Lv, Keping Yu, Ammar Hawbani, Kaifa Zheng, Li Xu

Abstract: In the mobile edge computing healthcare (MECH) system, the integration of MECH servers and wearable medical sensors can achieve real-time monitoring and analysis of user health. However, the system still faces key challenges such as network security risks and high energy consumption. To address these issues, this paper proposes a dual pronged solution. First, a new mechanism integrating smart contracts and asymmetric encryption is designed to achieve secure and efficient user authentication. Then, a method called Lyapunov Deep Deterministic Policy Gradient (L-DDPG) has been proposed to solve the resource optimization of the system. L-DDPG utilizes the Lyapunov optimization framework to transform the original long-term average constraint optimization problem into an instant optimization problem for each time slot, and solves the system optimization variables using the Deep Deterministic Policy Gradient algorithm. Through this design, L-DDPG effectively combines the advantages of Lyapunov optimization in ensuring system stability, as well as the decision modeling ability of deep reinforcement learning in complex state spaces, thereby simultaneously improving the resource utilization efficiency and safety of the system. The experimental results show that compared with existing baseline methods, L-DDPG significantly reduces the average energy consumption of equipment while effectively reducing task response delay, demonstrating better overall performance.

External IDs:doi:10.1109/tmc.2025.3608771