基于模糊C-均值聚类的TSP演化算法

Evolutionary Algorithms of TSP Based on Fuzzy C-Means Clustering

提出了一种基于FCM聚类的TSP演化算法。该算法以聚类中心为新的结点组成一个简单的TSP问题,用演化算法寻求其最短路径。在最短路径中,对于每一聚类,可寻求其距前面的聚类和后面的聚类最近的两结点之间的最短距离,若其中的结点较多,则再次演化得到其最短路径,若结点较少,则可用Warshall算法可得到最短路径。通过三个阶段的演化可得到较好的结果。

Evolutionary Algorithms（EA） to TSP based on FCM clustering is proposed.A simple TSP is formed on original complex TSP using the cluster center of fuzzy c-means clustering.The optimization route of the simple TSP is obtained by using evolutionary algorithms.In the route,each point represents a cluster.In each cluster,the optimization route between two points,which are nearest to the front cluster and the behind one respectively. If there are many elements in the cluster,evolutionary algorithms are used again,otherwise warshall algorithms avoiding EA.The better result of CHN144 TSP is obtained through three stages evolution,