1 A Combinatorial Approach for Optimal Sensor Selection Yifat Douek-Pinkovich, Irad Ben-Gal, Tal Raviv Department of Industrial Engineering, Tel-Aviv University, Ramat-Aviv, Tel-Aviv 69978, Israel E-mail: yifatdouek@gmail.com, bengal@tauex.tau.ac.il, talraviv@eng.tau.ac.il Abstract In complex systems, such as automobiles, aircraft, and even smart cities, it is possible to install a large number of sensors that can be used to monitor their states and make decisions regarding their operation and maintenance. However, because the sensors bear their costs and increase the complexity of the monitoring system, it is desirable to select the most inexpensive set of sensors that is sufficient to provide accurate and reliable information. In this paper, we introduce the sensor selection problem and formulate it as an integer linear program. The model is an extension of the minimum test collection problem with respect to several constraints, such as the sensors’ cost which is not necessarily equal. We present an effective exact solution method that uses the special structure of the integer-programming model to reduce its dimension so it can easily be solved. This solution method was implemented and demonstrated to be superior in comparison to a state-of-the-art integer programming solver. Using publicly available datasets from the UCI Machine learning repository, we demonstrate how the Optimization-based Sensor Selection model can be used as an effective feature selection method as well. Keywords: Sensor selection, feature selection, integer linear programming, dimension reduction