Research On Improved Self-Adaptive Cuckoo Search Algorithm And Its Applications

In recent years,with the development of swarm intelligence optimization algorithm,the intelligent optimization algorithm has become the important way to solve complex scheduling problems.It widely used in computer science,engineering and economic problems,and other fields.Due to the large number of local optimal solutions in complex optimization problems,traditional optimization algorithms are difficult to solve these problems.As a swarm intelligence optimization algorithm,cuckoo algorithm is widely used in optimization problems due to its excellent characteristics.In this paper,based on the problem of complex high-dimensional numerical optimization,a solution framework based on swarm intelligence is established,and an improved self-adaptive cuckoo algorithm is proposed.First of all,this paper analyzes the biological and mathematical principles of the cuckoo search algorithm,and explains the distribution of Lévy flight and mittag-leffler.Then,this thesis improves the algorithm of cuckoo bird based on the weakness of global searching ability of cuckoo algorithm.The new algorithm combines levy flight and mittag-leffler distribution,which improves the convergence speed and accuracy of the algorithm.Then,this thesis had adapted the parameters of the algorithm,and adjusts the search space to balance the local search ability and global search capability of the algorithm based on the distance between the current solution and the optimal solution.Finally,the convergence of the new algorithm is analyzed theoretically.The main work of this paper is as follows:(1)By using the protection strategy of cuckoo algorithm,once the global optimal solution is found,the optimal solution will not get out of the optimal solution and get into local optimal.(2)In this thesis,the variation and crossover operation in the differential evolution algorithm is applied to the improved cuckoo algorithm.The search efficiency of the parameter adaptive mechanism of differential evolution depends on the setting of control parameters,and the new control parameters can be determined during the evolution process.(3)The new algorithm simulates 30 functions on the CEC2017 benchmark set,and compares the results with other advanced algorithms.The experimental results show that the global optimal value can be obtained quickly and accurately by combining the adaptive step length and distance with the simulation and theoretical analysis.At the end of the thesis,summarizes research content of the whole algorithm and puts forward the future research direction.