On the influence of using initialization functions on genetic algorithms solving combinatorial optimization problems: A first study on the TSP

Abstract

Combinatorial optimization is a widely studied field within artificial intelligence. There are many problems of this type, and many techniques applied to them can be found in the literature. Especially, population techniques have received much attention in this area, being genetic algorithms (GA) the most famous ones. Although throughout history many studies on GAs have been performed, there is still no study like the presented in this work. In this paper, a study on the influence of using heuristic initialization functions in genetic algorithms (GA) applied to combinatorial optimization problems is performed. Being the first phase of this research, the study is conducted using one of the best known problems in combinatorial optimization: the traveling salesman problem. Three different experimentations are carried out, using three different heuristic initialization functions. Additionally, for each experiment four versions of a GA have been developed for the comparison. Each of these variant differs in the initialization phase. The results obtained by each GA are compared to determine the influence of the use of heuristic functions for the initialization of the population.

Optimal opportunistic tamping scheduling for railway track geometry

“…Initialization is one of the most important steps in the GA. Using heuristic initialization increases the efficiency of the GA, although an excessive use of heuristic initialization will decrease the exploration capacity of the algorithm (Osaba et al, 2014). Therefore, a heuristic method should be used for the initialization, which will result in initial logical solutions for the problem while ensuring the diversity of these solutions.…”

Hybrid genetic algorithm for a type-II robust mixed-model assembly line balancing problem with interval task times

“…Many researchers have proposed to seed EAs with good initial solutions, whenever it is possible, to obtain important improvement in the convergence of the algorithm and the quality of the solutions [19]. However, the excessive use of good solutions in the initial population can decrease the exploration capacity of the GA, thereby trapping the population in local optimums quickly [20,21].…”

Selección óptima del portafolio de proyectos utilizando metaheurísticas de población y trayectoria meta-optimizadas

“…En el caso de procedimientos de solución metaheurísticos basados en población, encontramos diversos algoritmos aplicados a problemas de optimización combinatoria tales como: Genetic Algorithm (GA) utilizados en el TSP (Osaba et al, 2014), casos de Inventory Routing Problems (IRPs) con múltiples locaciones y productos perecederos en los que se busca minimizar el costo de gestión total del inventario (Hiassat et al, 2017); Memetic Algorithm (MA) con similitudes estructurales a los GA, utilizados de igual manera en la solución de Flow Schop Scheduling Problems (FSSP) (Deng and Wang, 2017); algoritmos ABC para el Set Covering Problem (SCP) de complejidad NP-Complete, con el que se han conseguido mejores desempeños que los reportados en diversas investigaciones para algoritmos de ACO (Crawford et al, 2014), entre otros.…”