Genetic Algorithm Based on Classification for the Traveling Salesman Problem

This paper, considering the characteristic of traveling salesman problem (TSP), puts forward genetic algorithm based on classification (GABC). GABC classifies the individuals into different levels by their fitness. Each level with its respective operations generates new solutions in different numbers and accelerates the new solutions in high levels to approach the local minimum by accelerating operator. Furthermore, GABC adopts new selection strategy in each level. The results obtained show that GABC is an effective and robust method. Its performance outperforms some other techniques.