Algoritmo recocido simulado para el problema de la programación del tamaño del lote económico bajo el enfoque de ciclo básico

Abstract

The problem considered is that of scheduling the production of several different items over the same machine with restricted capacity and on a repetitive basis. The problem is NP-hard and there exist several methods for the problem. In this research we worked with Bomberger's basic cycle approach, for which there are reported several solution proposals, among which are found so far only genetic algorithms in regard to the implementation of metaheuristic techniques to solve the problem. We solved Bomberger's classical example problem, applying simulated annealing metaheuristic. The contributions of this research consist on how to obtain bounds for search space of the variables, and a strategy to control the search along the solution space performed by the algorithm. Because this is a first implementation, we tested several values of the parameters. The algorithm achieves the same cost in almost all tests and the algorithm performs better where the ratio is high, improving some of the results reported in the literature.