Análise Comparativa de Estratégias de Controle Populacional em Algoritmos Genéticos Aplicados ao Problema das N-Rainhas

Abstract

This work compares four population control strategies in genetic algorithms applied to the n-queens problem. Using a unified stopping criterion based solely on the maximum number of fitness evaluations, the study isolates the effect of each substitution mechanism on convergence, solution quality, and success rate. Results show that population control strongly influences evolutionary behavior: age-based control offers the best balance between exploration and efficiency, elitism excels in larger instances, and full or partial replacement degrades as diversity decreases.