Optimal Sensor Placement and Scheduling with the Value of Information

Abstract

The management of civil infrastructure involves accounting for uncertain or variable factors which influence the performance of a system. These factors can vary in space over the domain of the system and in time as the system changes. Effective decision-making for system management should be guided by models which account for uncertainty in these influencing factors as well as information gathered about the system to reduce this uncertainty. Value of information provides a rational metric for quantifying the benefits of information gathering efforts to support system management decision-making. However, the computation of this metric in spatially and temporally extensive systems presents a practical impediment to its implementation. In this paper, we investigate a special case of system topology, termed as a temporally decomposable system with uncontrolled evolution, in which the computational complexity of value of information evaluation grows at a manageable rate with respect to the problem time horizon. We demonstrate the evaluation of the value of information to support the design of a structural health monitoring scheme, using data collected for the Scott Hall building at Carnegie Mellon University. We also investigate the relative benefits of online sensor placement, i.e., of having the ability to revise sensor placements and/or schedules over time as information is gathered within a system.