借鉴蚁群算法中的信息素机制，并利用粒子群算法操作简单、易于实现、计算量小的特点，给出一种 新的求解TSP问题方法。对基本粒子群算法进行了改进，针对多样性下降导致的局部最优问题，设计了一种 自动调节机制。根据群体适应度的差异计算多样性，并在群体多样性下降到一定程度时，随机退化部分适应 值较高的粒子，增强群体的多样性。通过对旅行商问题的对比实验验证了该方法的有效性。

Using pheromone of ant coloney system as reference, a novel method of solving TSP problem is proposed. That is using particle swarm optimization ( PSO) . PSO is used because of its simple operation, easy implementation and faster speed. In order to improve the popularity of the particle swarm,make the particle swam not to homogeneous too fast and decrease the possibility of local constrain, the algorithm decides the number of degenerated particles based on a designated popularity function.Experiment results and comparison studies have demonstrated that our work is useful.

成伟明（1981－），男，江苏兴化市人，南京理工大学计算机学院博士生