Research on Task Offloading for Wireless Power Supply Edge Computing Based on RNN-SAC Algorithm

In mobile edge computing, local devices unloading computing, processing and storage tasks to edge servers in order to reduce service latency and improve service quality. At the same time, local devices can also use radio frequency energy harvesting technology to collect electromagnetic wave energy from the wireless channel of edge services to resolve energy consumption issues. Considering the complex environment of multi-user and dynamically changing wireless power supply edge computing, aiming at achieving long term rewards and optimizing task computing rate, and comprehensively considering resource constraints such as wireless channel electromagnetic wave, this paper proposes a wireless power supply edge computing task offloading algorithm based on RNNSAC. RNN-SAC algorithm optimizes policies, value functions and entropy through an RNN network, a policy network, two value networks and two target value networks, and has certain advantages in exploratory nature, robustness and stability. Simulation results show that RNN-SAC is superior to common deep reinforcement learning algorithms in convergence, reward value, computing rate and other indicators, and can be well adapted to the edge computing system environment.