Research And Application Of Ant Colony Algorithm On Clustering Analysis

Ant colony algorithm is a kind of combinatorial optimization algorithm based on bionic principles. It has the strongpoint of outer intelligent optimization algorithms and swarm intelligence algorithm. It was proposed by M.Dorigo in 1991, and was successfully applied to such problems:Travelling Salesman Problem, Quadratic Assignment Problem, Vehicle Scheduling Problem, circuit design and network routing design, etc. Ant colony algorithm is used for data mining in recent years, which provide a broader line of thought for clustering analysis and classification algorithms of data mining.This paper describes the basic idea, principles and theories of ant colony algorithm in detail, and then analyzes the advantages and disadvantages in detail, proposes a number of improved methods, and applies the ant colony algorithm to clustering analysis of data mining; make a more in-depth research and analysis on several kinds of ant colony clustering method, and obtain a improved ant colony clustering algorithm. The main contents of this paper are as follows:(1) The overview of ant colony algorithm. Ant colony algorithm is a swarm intelligence algorithm built based on bionic, which has its own unique advantages in solving combinatorial optimization problems. This paper briefly introduces the development status of ant colony algorithm, describes the basic ideas, principles, mathematical models and implementation process in detail, and proposes improved ant colony algorithm after analyzing and comparing the advantages and disadvantages of several common ant colony algorithm.(2) The overview of clustering analysis. Clustering analysis is an important part of data mining and operations research. The target of clustering analysis is to divide a group of objects into several clusters according to certain rules, which need that the similarity between the objects in the cluster is as greatest as possible and that the dissimilarity between the clusters is as largest as possible.(3) The overview of ant colony clustering algorithm. Apply the characteristics of ant colony algorithm to clustering analysis so as to make the results of cluster analysis better and more efficient. In this paper, the principles of several classical ant colony clustering algorithms are briefly introduced, the corresponding improved method is proposed aimed at the shortcomings of several ant colony algorithms, and a new kind of combinatorial clustering algorithm is proposed by combining the characteristics of several ant colony clustering algorithm. At the same time, the model, the specific algorithm and the algorithm steps of combinatorial clustering algorithm is provided. Finally, the paper provides a simulation experiment analysis, which gets the conclusion by comparing the combinatorial clustering algorithm with K-means algorithm, the basic ant colony clustering algorithm and its improved algorithm.