基于MATLAB的自适应蚁群聚类算法研究与仿真.PDF

SOFTWARE DOI: 10.3969/j.issn.1003-6970.2012.07.035 ZHOU Teng (South-Central University for Nationalities, College of computer science, Wuhan 430074, China) Abstract Introduce the ant colony algorithm and cluster analysis, because of major shortcomings of parameters setting which affects the result about traditional ant colony clustering algorithm, proposed an adaptive ant colony clustering algorithm. The original ant colony clustering algorithm has been mainly improved by these methods: Simplify parameters of the traditional ant colony algorithm, increasing the computation of local similarity, make the ants move with a certain direction, thereby reduce the randomness of the ants move, and analysis of the principal component analysis of the clustering pattern and then cluster, simplify the clustering process. Finally, in the MATLAB simulation and experimental results show that the use of the algorithm can get better clustering results, which demonstrate that the algorithm is feasible. Key words Ant colony algorithm; Cluster analysis; Adaptive; MATLAB ?f (Oi) ? 1 n Oj ? [1 ? L(Oi, r) d (Oi, Oj) ] ? ir i r ii ni / Oi ?f (Oi) ? 1 n d (Oi, Oj) Oi ? L(Oi, r) Oi Oi Oi Oi Oi O d (Oi,Ok) ? min?d (Oi,Oj) ?,Oj ? L(Oi, r) Oi / /