Deep Reinforcement Learning based Energy Efficient Resource Allocation for Wireless Powered Edge Computing Network

This paper considers wireless power edge computing networks with time-varying wireless channel environments. A practical non-linear energy harvesting model is introduced, and a joint optimization problem is formulated to minimize the overall cost of energy consumption. To solve the problem, we design a deep deterministic policy gradient (DDPG) algorithm to jointly optimize computation offloading and resource allocation. Numerical results show that the proposed schemes can significantly ensure the timeliness of data and effectively achieve lower energy consumption.