A new algorithm for TSP based on swarm intelligence

Inspired by the behavior of people, a new algorithm for the combinatorial optimization is proposed. This is a heuristic approach based on swarm intelligence, which is firstly introduced as the theoretical background in this paper. It is also a parallel algorithm, in which individuals of the swarm search the state space independently and simultaneously. When one encounters another in the process, they would communicate with each other, and utilize the more valuable experiences to improve their own fitness. A positive feedback mechanism is designed to avoid vibrations. Ten benchmarks of the TSPLIB are tested in the experiments. The results indicate that the algorithm can quickly converge to the optimal solution with quite low cost. Some conclusions about the algorithm are summarized finally