Opportunistic relay selection for cooperative energy harvesting communication networks

We consider cooperative energy harvesting communication networks in which a set of source-to-destination pairs competes for a limit number of relay nodes with energy harvesting ability. The performance of each source has been affected by two interactions: the interaction between the sources and relay nodes and the interaction among sources. We model the first interaction as a college admission market and then fits this market into a stochastic environment. We formulate an interactive partially observable Markov decision process (I-POMDP) to study the second interaction. We derive the optimal policy for the sources to sequentially optimize their decisions. Numerical results show that our proposed policy significantly improves the performance of sources.