Traveling Salesman Problem Solving Based on an Improved Genetic Simulated Annealing Algorithm

Rapid convergence in the global optimal solution is a focus of genetic algorithm.Based on the study of genetic algorithm and simulated annealing algorithm,the paper analyses the major merits and shortcomings of the two algorithms,and gives an improved genetic simulated annealing algorithm,which combines the major merits of the two algorithms.Especially,it gives a parallel searching structure of multi layers,and gives a new criterion for judging the premature convergence in this improved algorithm.At last,the algorithm is applied in the traveling salesman problem(TSP),and it is simulated with both of 10-city TSP and 30-city TSP to prove the algorithm's feasibility and efficiency.The results of the simulation indicate that the improved algorithm has a rapid convergence in the global optimal solution.