Cross-Entropy optimization for sensor selection problems

Abstract

In this paper, we apply the Cross-Entropy optimization (CEO) to the problem of selecting k sensors from a set of m sensors for the purpose of minimizing the error in parameter estimation. The computational complexity of finding an optimal subset through exhaustive search can grow exponentially with the numbers (m and k) of sensors. The CEO is a generalized Monte Carlo technique to solve combinatorial optimization problems. The CEO method updates its parameters from the superior samples at the previous iterations. The performance of proposed CEO-based sensor selection algorithm is better than existing sensor selection algorithm, and its effectiveness is verified through simulation results.