Policy Gradient based Channel Resource Allocation Strategy

Senlei Bao, Shubin Zhang, Kaikai Chi

Abstract: Mobile edge computing (MEC) contributes to the realization of the high transmission rate requirements of the new services of the fifth generation mobile communication (5G). In this paper, we consider a radio frequency powered MEC network with a fully offloaded strategy in the scenario where the channel environment changes in real time, where each computing task of each wireless device (WD) is completely offloaded to the MEC server. Our goal is to obtain an optimal resource allocation strategy to maximize the sum of the transmission rates of the entire network in a single time slot. We need to address the impact of channel environment and resource allocation on data transmission by wireless devices. We propose a resource allocation strategy based on the policy gradient algorithm. To evaluate the performance of the proposed algorithm, we compare it with the iteration optimization. Through the simulation results, the algorithm proposed in this paper can achieve more than ninety percent computation rate with one thousandth processing time, which is suitable for time-varying channel.