Research And Application Of Chaos Firefly Optimization Algorithm Based On Rough Set Theory

Among the bionic algorithms there is a kind of algorithm called swarm intelligence optimization algorithm,which is an emerging algorithm that simulates the interaction and attraction between insect communities and particle swarms in the animal kingdom.The optimization algorithm studied in this paper is the Firefly Algorithm(FA),which is the intelligent optimization algorithm first proposed by Professor Yang Xinshe of Cambridge University in the process of mathematical modeling.Since its introduction in 2008,it has received close attention from the academic community and has become one of the hot research topics of scholars at home and abroad.First,this thesis discussed the basic operating principle and corresponding algorithm flow of the standard firefly algorithm,and then proposesd an improved algorithm based on the defects found according to the algorithm analysis process——Chaos Firefly optimization Algorithm based on rough set theory(Chaos Firefly optimization Algorithm based on Rough Set theory.RSCFA).In this paper,the two aspects of the firefly initial population's structure and the movement formula(how the firefly moves during the optimization process have been changed in sequence:First,for the initial population constructed by the random method in the standard firefly algorithm,it is prone to uneven distribution.Therefore,in this paper,the cubic mapping chaos model is used to homogenize the initial population,which effectively improved the defect that the population cannot be uniformly covered,increased the diversity of firefly populations;Second,for the problem that firefly individuals always oscillate around the extreme points during the iteration process,the improved rough set theory is introduced and construct the characteristic coefficient T,which is used to modify the original position update formula,and effectively prevented the occurrence of oscillation problems,and significantly improved the optimization accuracy of the algorithm,at the same time,the convergence speed is also improved,the most important thing is to improve the balance ability in the global and local search.In the verification experiment,through multiple test functions,compared with the other four optimization algoritluns,it is found that the optimization results of RSCFA are better.Finally,the firefly algorithm is combined with the actual problem:In this paper,the K-means algorithm is used as a combination object.Because it is a typical representative of the clustering algorithm,considering its easy convergence in any local optimal solution,the advantages of the firefly algorithm are used,for example,it has good global search ability and fast convergence speed,etc,so we skillfully transformed clustering problems into optimization problems.At the same time,the three algorithms are used for comparison.and four typical datasets in the UCI are selected to experiment with the algorithm.Combined with the advantages of K-means clustering algorithm that can quickly converge and the improved firefly algorithm can achieve the advantage of global optimization,get rid of local optimization and obtain global the optimal solution shows that the chaos firefly algorithm based on rough set theory combined with the K-means clustering algorithm can achieve the best clustering effect.