Learning to Optimize Energy Efficiency in Energy Harvesting Wireless Sensor Networks

We study wireless power transmission by an energy source to multiple harvesting nodes with the aim maximize efficiency. The transmits using one of available levels in each time slot and transmit information back harvested energy. does not have any channel state it only knows whether a received codeword from given node was successfully decoded or not. With this limited information, has learn optimal level that maximizes efficiency network. model problem as stochastic Multi-Armed Bandits develop Upper Confidence Bound based algorithm, which learns Numerical results validate performance guarantees proposed algorithm show significant gains compared benchmark schemes.