“PLANIFICACIÓN ÓPTIMA DISCRETA PARA PROBLEMAS NP-COMPLETOS MEDIANTE ALGORITMOS INMUNOLÓGICOS ARTIFICIALES”

Abstract

Abstract The purpose of this thesis work is to present a new algorithm that helps in solving combinatorial optimization problems of the NP-complete order which is based on the human immune system, being its abstraction the Artificial Immune Systems. Relevantly this work proposes a different view to the current trends of these algorithms by introducing the concept of vaccine-induced immunity and the application of this to generate solutions. We present three algorithms called: generation of vaccines by Random Selector, by Elitist Selector and Expansion of vaccinated elements. We performed a comparative study of the collection of data from three experiments designed with the aim of assessing the quality of the solutions, the points of convergence and execution times of the algorithms. We show that the proposed algorithms favorably influence the reduction of computing time and the improvement of the quality of the solution