An improved genetic algorithm for solving Travelling Salesman Problems

An improved genetic algorithm is presented for solving traveling salesman problems. This method introduces unfitness function as a criterion at the beginning of iteration, designs a new crossover operator, applies a hybrid mutation operator, and makes immune operation on individual after mutation operation. In addition, we recompose the population to insure that every individual in it is different. We give a simple proof of the theory. The simulation numerical results show that this algorithm is efficient.