An Improved Imperialist Competitive Algorithm For Solving Traveling Salesman Problems

In order to better solve traveling salesman problems, we proposed an improved imperialist competitive algorithm (IICA). Firstly, the original continuous algorithm is discretized to be suitable for solving combinatorial optimization problems. Then, four improvements are proposed to enhance the overall performances of the proposed algorithm: introducing immune fitness based on immune mechanism, employing parallel computation strategy, improving the strengthening link of imperialist countries according to the idea of simulated annealing, as well as adopting adaptive mechanism to accelerate algorithm convergence speed. Finally, comparing simulation results by proposed algorithm and other algorithms on typical traveling salesman problems selected from the classic test library demonstrate that IICA can solve these problems on different scales more effectively, and it has better convergence accuracy and speed.