Tongxin xuebao (Feb 2017)
Image cluster algorithm of hybrid encoding method
Abstract
In the clustering analysis based on swarm intelligence optimization algorithm,the most of encoding method only used single form,and this method might be limit range of search space,the algorithm was easy to fall into local op-timum.In order to solve this problem,image clustering algorithm of hybrid encoding (HEICA) was proposed.Firstly,a hybrid encoding model based on image clustering was established,this method could expand the scope of the search space.Meanwhile,it was combined with two optimization algorithms which improved rain forest algorithm (IRFA) and quantum particle swarm optimization (QPSO),this method could improve the global search capability.In the simulation experiment,it was carried out to illustrate the performance of the proposed method based on four datasets.Compared with results form four measured cluster algorithm.The experimental results show that the algorithm has strong global search capability,high stability and clustering effect.
