Energy Harvesting Cooperative Multiple Access Channel with Decoding Costs

We consider an energy harvesting cooperative multiple access channel (AC) with decoding costs. In this setting, users cooperate at the physical layer (data cooperation) in order to increase the achievable rates. Data cooperation comes at the expense of decoding costs: each user spends some amount of its harvested energy to decode the message of the other user, before forwarding both messages to the receiver. The decoding power spent is an increasing convex function of the incoming message rate. We characterize the optimal power scheduling policies that achieve the boundary of the maximum departure region subject to energy causality constraints and decoding costs by using a generalized water-filling algorithm.