Proportional Controller from Control Theory for Tuning cooling scheme of Simulated Annealing algorithm

Abstract

To solve difficult problems of combinatorial optimization, heuristic methods such as Simulated Annealing (SA) and Threshold algorithm (TA) have been proposed. Both SA and TA should have adequate characteristics to explore efficiently the space of solutions. The process to determine these characteristics is known as parameter tuning problem. There is a strong interest in developing techniques to adjust parameters correctly, since the heuristic algorithms have a great applicability in industrial problems. Given the quantitative nature of multiple parameters it is possible to propose the selection of the best parameters of a heuristic as a combinatorial optimization problem. In this paper, a new tuning method based on the proportional controller derived from classical control theory for tuning the cooling scheme in real time is presented. Experimentation shows that this method has a better performance than the classical one.