Algoritmo Memético Autoadaptativo para Solución de Problemas Combinatorios

Abstract

The problems of combinatorial optimization are problems of high complexity, whose best solutions have been given through the use of metaheuristics. Memetic Algorithms (MA) hybridize a Genetic Algorithm (GA) with a Local Search Algorithm (LS), resulting in a metaheuristic with advantages in global and local exploration. The present investigation proposes an MA that adjusts during the exploration the methods and parameters of Local Search depending of the population diversity, and in that adjustment the algorithm innovates also making explorations for the infactible space of solutions. The Self Adaptive MA is tested using the knapsack problem.