Multi-objective optimization for sensor placement: An integrated combinatorial approach with reduced order model and Gaussian process

We develop a novel sensor placement method that maximizes monitoring performance while minimizing deployment cost. Our integrates reduced order model and multi-objective combinatorial optimization. first decompose the spatio-temporal state field to be monitored by proper orthogonal decomposition (POD), we use Gaussian Process uncertainty in each POD mode. Next, lazy greedy (LG)-∊-constraint optimization derive Pareto-optimal configurations. further design branch bound algorithm calculate global optimum validate correctness of select configurations on LG-derived Pareto frontier. evaluate benchmark our computational experiments based temperature dataset Berkeley Intel Lab. The results demonstrate places sensors at locations large magnitude modes, achieves better estimation accuracy smaller reconstruction errors compared with alternative methods.