Energy-Efficient Power Control for Energy Harvesting Cooperative Relay Sensor Network

In this paper, we consider the power control problem in cooperative relay sensor network, in which the nodes harvest energy from environment to transmit data. The goal of this paper is to optimize the energy efficiency of the network subject to the harvested energy and transmitted data. Due to the stochastic properties of channel state information (CSI) and energy harvesting process, we formulate the energy efficiency problem as a Markov decision process, and the model-free reinforcement learning algorithm SARSA is proposed to learning the power control policy. To address the problem with continuous state space, we propose a set of feature functions to approximate action value function. Numerical results verify that the proposed algorithm is outperform than other two algorithms.