An A3C-based Joint Optimization Offloading and Migration Algorithm for SD-WBANs

The problems of insufficient power consumption and delay sensitivity are challenging issues of Software Defined-Wireless Body Area Networks (SD-WBANs) for smart health monitoring. The migration and offloading problems in mobile edge computing (MEC) system complicate the resource allocation on low delay and energy consumption. In this paper, we propose an Asynchronous Advantage Actor-Critic (A3C)-based joint optimization offloading and migration algorithm to address this issue. Based on Actor-Critic network, the proposed algorithm optimizes the neural network with asynchronous gradient descent to maximize the long-term benefits of SD-WBAN. Moreover, we define the total system cost as the combination of SD-WBAN migration cost, offloading cost and quality of Service. Simulation results show that this algorithm can effectively get better long-term benefits of different priority tasks and highly improve the service quality of the users.