摘要
Ant colony system(ACS),a kind of ant colony algorithm,is an effective way of solving shortest path problem,however,it has some defects.In this paper,ACS is improved for avoiding getting stuck in a local minimum,whose defects mainly include the following two aspects:initial pheromone solution and pheromone updating.In order to learn the advantages of improved ant colony system(IACS),experiments are conducted for some times.First,it is applied to 8 traveling salesman problem(TSP)instances,and compared with three self-organizing map(SOM)algorithms.Then the author analyzes the space complexity and convergence of two algorithms and compares them.Simulation results show that IACS has much better performance in solving TSP,and it has certain theoretical reference value and practical significance.
Ant colony system （ACS）, a kind of ant colony algorithm, is an effective way of solving shortest path problem, however, it has some defects. In this paper, ACS is improved for avoiding getting stuck in a local minimum, whose defects mainly include the following two aspects： initial pheromone solution and pheromone updating. In order to learn the advan- tages of improved ant colony system （IACS）, experiments are conducted for some times. First, it is applied to 8 traveling salesman problem （TSP） instances, and compared with three self-organizing map （SOM） algorithms. Then the author ana- lyzes the space complexity and convergence of two algorithms and compares them. Simulation results show that IACS has much better performance in solving TSP, and it has certain theoretical reference value and practical significance.
基金
National Natural Science Foundation of China(No.61275120)
