Resumen
Optimization methods in discrete-event simulation have become widespread in numerous applications. However, the methods´ performance falls sharply in terms of computational time when more than one decision variable is handled. Current assay develops an adaptive genetic algorithm for the simulation optimization capable of achieving satisfactory results in time efficiency and response quality when compared to optimization software packages on the market. A series of experiments was elaborated to define the algorithm?s most significant parameters and to propose adaptations. According to the results, the most significant parameters are population size and number of generations. Further, adaptive strategies were proposed for these parameters which enabled the algorithm to obtain good results in response quality and time necessary to converge when compared to a commercial software package.