Uso de Algoritmo Cultural com uma Nova Abordagem Memética por meio do Simulated Annealing para o Problema do Caixeiro Viajante

Abstract

The paper proposes a hybrid algorithm combining Cultural Algorithms (CAs) and Genetic Algorithms (GAs), with local search via simulated annealing and 2-opt/3-opt heuristics, to solve the Traveling Salesman Problem (TSP). The aim is to overcome premature convergence and local optima traps common in traditional metaheuristics. Experiments with 442 and 532-city instances showed that the approach achieved solutions close to the known optimum, outperforming other methods in the literature.