A New Hybrid Ant Colony Algorithm for Clustering

Focusing on the problem that the ant colony clustering algorithm may be convergence slowly and easily fall into local optimal drawbacks.Proposed a new hybrid algorithm by adds GA to Ant Colony clustering algorithm's every generation.Making use of GA's advantage of whole quick convergence,Ant Colony clustering algorithm's convergence speed was improved.Meanwhile,the operation of crossover and mutation improved the ability of Ant Colony clustering algorithm to avoid being premature.This algorithm has been implemented and tested on several simulated datasets and UCI machine learning datasets.The authors' experiments reveal very encouraging results in terms of the quality of solution found and the processing time required.