Cooperative Reinforcement Learning based Throughput optimization in Energy Harvesting Wireless Sensor Networks

Energy Harvesting-Wireless Sensor Network (EH-WSN) has got increasing attention in recent years. During its actual deployment, we find that the energy which can be harvested from the environment is always continuous changing and unpredictable. This paper aims to investigate the energy management approach of EH-WSN under such circumstance and propose a corresponding dynamic scheme to optimize the network throughput. Here we adopt a Cooperative Reinforcement Learning (CRL) method to analysis: Firstly we model the external environment status, and then the CRL algorithm based on Q-learning starts regulating the EH-node's duty cycle according to the external energy's variation, meanwhile the feedback reward takes responsibility for the evaluation of CRL's regulation. Different from traditional reinforcement learning, CRL facilitates EH-nodes to share their local knowledge with others periodically. With this information, Ell-node chooses which action to take for the current time slot: (I) idling, (II) sensing, (III) calculating, and (IV) transmitting. Experiment results show that the proposed scheme can make Ell-node working energy-balanceable, and satisfying the network throughput requirement effectively, it also improves the energy utilization efficiency obviously in contrast with existing strategy.