提出了一种新型的自适应蚂蚁聚类算法.该算法将每个待聚类模式看作一只蚂蚁,采用蚂蚁移动模型实现模式的聚类.为了改善蚂蚁移动的随机性,提高运行效率,提出了一种局部最近邻运动原则来指导蚂蚁的移动;并且提出了一种自适应调整蚂蚁移动阈值的方法以简化参数的选取.通过数据的聚类对该算法和已有算法进行了比较.结果表明,该算法具有运行效率高、参数选取简单及自适应性等优点.
To reduce the randomness of ant's movement and improve the efficiency, a principle moving to the nearest neighbor in the local environment was suggested to guide the ant's movement. Moreover, a method to self-adaptively adjust the threshold of ant's movement was presented, which simplified the parameter's selection. The algorithm was compared with others through the application of data clustering. The results show the proposed algorithm has high efficiency, simple parameters and self adaptivity.