Multi-flow scheduling in multi-hop underwater acoustic networks

We study maximum multi-flow scheduling in an underwater acoustic sensor network. The network consists of underwater sensor nodes and surface gateways. Each sensor node seeks to send data to any of the gateways. The objective is to maximize the sum of the flows that can be finished within a fixed time duration. To address challenges posed by large sound propagation latency, we convert the original network to a time-expanded network, and develop a linear programming (LP) formulation for the maximum multi-flow scheduling. The solution to the LP formulation provides an upper bound of the maximal achievable flow. Through introducing a stronger constraint into the LP formulation, a feasible multi-flow transmission schedule is proposed, which guarantees a small approximation factor. Simulations are conducted to verify the theoretical results.