Split to win: near-optimal sensor network synthesis via path-greedy subproblems

Jonathan Bunton, Tzanis Anevlavis, Gunjan Verma, Christina Fragouli, Paulo Tabuada

Abstract: As lightweight networked sensing and computing becomes pervasive in military operations, we are faced with the problem of quickly determining a maximally useful network topology for sensors subject to computation and communication constraints. In this paper, we frame network synthesis as a combinatorial optimization problem. Then, leveraging tools from submodular optimization, we develop a simple greedy algorithm that produces near-optimal sensor placements satisfying these communications and computation constraints.