Exploring Non-Submodular Scheduling for Large-Scale Sensor Networks

This letter addresses the intricate issue of non-submodular sensor scheduling within the context of large-scale linear time-varying dynamics. The problem involves optimizing the configuration of sensors, a task that is inherently combinatorial, non-convex, and NP-hard. We delve into the utility of a simple greedy algorithm for problem resolution. We provide evidence of the algorithm's effectiveness by presenting an approximation bound for its solutions based on the submodularity and curvature concepts. It is shown that the proposed approximation bound outperforms the competitors in the literature through a discussion on a simple setup. The letter culminates in a comprehensive set of simulation results, which validate the theoretical underpinnings.