Cuckoo Search Algorithm Based On Change Factor

The optimization problem is to find the best solution quickly and accurately under the restriction of specific realistic environment,and get the best result of practice.In many fields such as engineering design,energy distribution and medical applications,there are optimization problems,and the cuckoo search algorithm(CS)which has the advantages of less parameters,easy implementation,and good effect,becomes an effective solution to the optimization problem.The CS algorithm is a kind of optimization process that simulates the cuckoo's constant search for the living habits of the high-quality host nest.However,the classic cuckoo search algorithm has the problems of uneven distribution of initial population and low quality,global and local search methods can not be reasonably controlled,population diversity is not easy to adjust and the reconstructed bird nest solution quality is not high,affecting the accuracy and convergence speed performance of the algorithm.To solve the above problems,this thesis proposes a novel cuckoo search algorithm based on change factor(CFCS),the algorithm is mainly improved as follows:(1)In order to solve the problem that the initial population of CS algorithm is unevenly distributed and the quality is not high,it is proposed to introduce the quasi-Monte Carlo method to initialize the population at the initial bird nest stage.Firstly,the Hammersley sequence of the quasi-Monte Carlo is used to generate nests in the divided subspaces so that the population can be evenly distributed in the search space;and then the nest positions in the entire space are sorted according to the fitness value.Selecting the optimal bird nest position becomes the initial population of the initial algorithm,which improves the initial population quality.(2)In order to solve the problem that the global and local search methods of the CS algorithm cannot be reasonably controlled,a method of using the change factor in the bird nest update iteration stage is proposed.Firstly,the iterative process of the algorithm is divided into the early,middle and later phases according to the number of iterations;and then in different periods,different factors are used to define a new step size model and an iterative updating model of the bird's nest position,which improves the rationality of the application of the search method and improves the convergence speed of the algorithm.(3)In order to solve the problem that population diversity of CS algorithm is difficult to adjust,a method of using staged adaptive probability in the stage of abandon-rebuild the nest is proposed.Firstly,following the bird nest update iteration stage,the algorithm continues to be divided into the early,middle,and later phases;and then according to the number of iterations,the bird nest abandonment probability is calculated dynamically,and the population diversity in different periods is dynamically adjusted by different probabilities so that the algorithm can flexibly adjust the population diversity and improve the convergence speed of the algorithm.Further following the bird nest update phase,the rationality of the application of the search method of the algorithm is ensured.(4)In order to solve the problem of the CS algorithm rebuilding the bird nest quality is not high,In the abandonment-rebuilding bird's nest phase,the reconstruction of the bird nest model was improved.Adjust the optimal step size by introducing the optimum nest position in the rebuilding bird nest model,the quality of the reconstructed bird nest solution is improved,and the accuracy of the algorithm's optimization value is improved.Finally,the CFCS algorithm is applied to function optimization to verify the effectiveness of the algorithm.In this thesis,common benchmark functions are used to simulate and analyze the algorithm and contrast algorithms.The results show that the CFCS algorithm can converge to the optimal solution at a faster speed,and its optimization accuracy can also be improved.