Dynamic Energy Trading for Energy Harvesting Communication Networks: A Stochastic Energy Trading Game

This paper studies energy-harvesting communication systems in which different energy-harvesting devices (EHDs) can harvest different amounts of energy and transmit different numbers of data packets in different time slots. We introduce a dynamic energy trading framework that allows the EHDs to transfer and trade their harvested energy with each other. The EHDs are divided into two groups: seller EHDs that can harvest more energy than they can use, and buyer EHDs that cannot harvest sufficient energy to support their required communication services. In the proposed framework, the role of each EHD as a seller EHD or a buyer EHD as well as the amount of energy that each EHD can buy or sell to others change over time. Each EHD cannot observe complete information regarding the harvested energy or the number of data packets transmitted by other EHDs. We introduce a simple energy trading scheduling protocol for the EHDs to discover their nearby EHDs and establish energy trading links with each other. We formulate a new game theoretic model called stochastic energy trading game to analyze the dynamic energy trading among EHDs in a stochastic environment. We derive an optimal energy-trading policy for each EHD to sequentially optimize its decisions. We prove that the proposed policy can achieve a stable and optimal sequence of matchings between buyer and seller EHDs. We present numerical results to compare our proposed energy trading policy with an existing transmit packet scheduling approach, under various network settings and conditions.