Improvement And Application Of Self-adapting Spherical Search Algorithm

With the progress of society,the development of science and technology,and the problem of combination optimization has become a hot spot in many disciplines.Researchers have been inspired by human intelligence,social or natural phenomena of biological groups,and invented many intelligent optimization algorithms to solve complex combination optimization problems.SPHERical Search Algorithm(SS)is a new cluster-based metaphysical algorithm.It converges fast and can maintain the diversity of population while finding the optimal solution.The adaptive spherical search algorithm(SASS)is an adaptive strategy based on successful historical parameter adaptive strategies to enhance the performance of the algorithm to the SSS algorithm.Although the SASS algorithm can find the optimal solution to the problem more effectively,it still faces the lack of exploration ability is relatively weak during the search process.Based on the above problems,this article proposes two improved adaptive spherical search algorithms,and applies the improved algorithm to solve the actual engineering problem.The main research content is as follows:(1)An Enhanced Self-Adapting Spherical Search Algorithm(ESASS)proposed a new search strategy.In order to prevent adaptive spherical search algorithms from precocious convergence in the late iteration,the algorithm is divided into two stages.In the first stage,the adaptive spherical search algorithm is used.In the second stage,half of the individual uses adaptive spherical search algorithms to search in the second stage.Search for search strategies.The new search strategy has strong exploration ability.Therefore,the ESASS algorithm can better balance the exploration and mining in the search process.The results of numerical experiments and actual engineering problems show that the algorithm proposed has good ability to find excellence.(2)For adaptive spherical search algorithms,the exploration ability is relatively weak.In the later period of iteration,it is easy to fall into partial optimal.The problem of precocious and convergence occurs.With Differential Evolution for Global Optimization(SSDE).Add reverse learning strategies during the initialization stage of the population to enhance the quality of the initial solution of the population.The iterative process is divided into three stages,and different iterative formulas are adopted at each stage.Add the parameter adjustment mechanism and change the random parameter in the algorithm to adaptive parameters.Use mutations and selection strategies in differential evolution algorithms to enhance the quality of solutions obtained.The results of numerical experiments and actual engineering issues show that the SSDE algorithm has better ability to find better compared with other algorithms.