Novel Research And Implementation Of Grey Wolf Algorithm

In the past 20 years,metaheuristic optimization algorithm has been widely used in all fields of social economy.In 2014,Mirjalili et al.proposed a new metaheuristic optimization algorithm named Grey Wolf Optimization(GWO)algorithm,which has excellent performance in finding the optimal solution.However,there are still some shortcomings that the algorithm converges to the local optimum.However,the improvement of this algorithm on this issue has not yet been fully explored by domestic and foreign scholars.In this paper,aiming at the defect that gray wolf algorithm is easy to fall into local optimum,which leads to the low search accuracy,an improved GWO algorithm is proposed.First of all,in the late evolution of GWO algorithm,as all gray wolf individuals in a population approach the estimated optimal individual region,the entire population tends to fall into the local optimal solution,resulting in the lack of diversity in the population.In this paper,the control factors combine with randomness,and as the search progresses,the local search is more sufficient to improve the performance of the algorithm.Secondly,this paper introduces the Levy flight strategy with better randomness,Levy flight moves a series of small steps and suddenly moves a bigger step in the later stage of evolution,all the gray wolves move to the estimated global optimal solution,and the gray wolves optimize The convergence speed of the algorithm is obviously slowed down or stopped,which is an inherent defect of the group intelligent optimization algorithm of population search.In order to avoid the gray wolf optimization algorithm into a local optimum,this paper carries out the diversity variation on the current optimal wolf's position.The simulation results show that compared with PSO algorithm,GSA algorithm,FEP algorithm and DE algorithm,the improved Gray wolf algorithm can get rid of the local convergence effectively.In search accuracy,algorithm stability and convergence rate has obvious advantages.