Opportunistic scheduling in two-way wireless communication with energy harvesting

A two-way half-duplex communication model is considered, where two nodes want to exchange a fixed number of bits with each other, and both nodes are powered by energy harvesting (EH) sources. The problem of minimizing the sum of the time required to send the required bits in both the directions is considered. The model also includes the processing cost at each node, that models the power needed for nodes to stay powered on during transmission. In the offline setting, where the EH arrival profile is known non-causally, an iterative algorithm based on alternating maximization is shown to be optimal. In the more realistic setting of causal knowledge of the EH arrival profile, an online algorithm is shown to be optimal in terms of the competitive ratio and the optimal competitive ratio is shown to be 2.