Algoritmo genetico aplicado ao problema do 8-puzzle

Abstract

This study presents the implementation and analysis of a Genetic Algorithm (GA) applied to the solution of the 8-puzzle problem, with the purpose of investigating the efficiency of this method in combinatorial search problems. The main objective was to develop an approach that combined selection, crossover, and mutation operators to identify optimal or near-optimal solutions under different puzzle configurations. The method consisted of representing puzzle states as chromosomes, applying genetic operators, and evaluating the population based on a fitness function. The experiments demonstrated that the GA could find viable solutions within a limited number of moves and showed consistent performance across different mutation, crossover, and elitism rates, highlighting the importance of balancing exploration and population diversity. The results obtained were highly satisfactory when compared to the literature, confirming the relevance of proper parameter tuning. It is concluded that the application of GAs to search problems, such as the 8-puzzle, yields promising results, making it possible to adjust parameters to optimize both efficiency and solution quality.