Application Of Swarm Intelligence In Clustering Algorithm

With the rapid development of information technology,artificial intelligence optimization technology is becoming mature.Clustering analysis and Swarm intelligence as hot spots have been paid much attention by researchers from all over the world,and been widely used in many fields.In this paper,the swarm intelligence optimization algorithms have been used to research the traditional clustering analysis problem.The Bat Algorithm and Chicken Swarm Optimization Algorithm were improved and combined with the traditional clustering algorithm to solve the problem of cluster analysis.The main research contents of this paper are as follow:1.To solve the premature convergence problem of Bat Algorithm,a hybrid improved Bat Optimization Algorithm was proposed.First,the Levy flight search strategy was introduced into the algorithm,which can make the bat group explore the global optimization more efficiently.Second,a threshold is used to assist the pulse loudness and emission rate judge the local optimum,the algorithm can jump out of the local optimization easily.Finally,the nonlinear inertia weight was combined to improve the individual position update mode,balancing the global search ability and local optimal.2.In order to solve the limitation of Chicken Swarm Algorithm,a chicken swarm algorithm based on improved simulated annealing and the chicken swarm algorithm based on difference variation were proposed.In the first improvement method,opposition based learning and boundary mutation were introduced into initial population,which made the algorithm can widely search in the solution space;meanwhile on the basis of improving optimization method for the location of hen,the simulated annealing is used to disturbed the current optimal.In the second improvement method,a new improved optimization method for the location of hen is proposed,a threshold is introduced to determine whether the current is trapped in the local optimum,and the global optimum is disturbed by differential evolution.The above improvements make the algorithm can jump out of the local optimization,increasing the ability of searching the optimal solution.3.Finally,the features of swarm intelligence can be improved the efficiency and quality of clustering analysis,two improved clustering analysis algorithm which were based on the improved bat algorithm and the chicken swarm algorithm based on difference variation is proposed.Compared with the traditional clustering analysis algorithm,the clustering efficiency and precision were highly increased.