Improvement And Application Of Estimation Of Distribution Algorithms And Chaos Algorithm

Optimization problem is often encountered in scientific research and engineering practice.The purpose of optimization is to select the optimal scheme from many schemes for the practical problems.With the progress of human society and the acceleration of the pace of exploration and transformation of the world,all kinds of optimization problems emerge in endlessly and become increasingly complex.In order to solve these optimization problems,the development of new intelligent optimization algorithms and the improvement of algorithm optimization performance have become an important research direction in the field of science and engineering.With the rapid development of information technology and artificial intelligence,the research of intelligent optimization algorithm has made a series of new achievements in the design of new algorithm,the improvement of algorithm performance,the expansion of algorithm application and the perfection of algorithm theory system.The main contributions of this paper as follows:Estimation of distribution algorithm has a strong global search capability,but its partial refinement capacity is weak.To solve the problem,the chaos search with strong partial ergodic property is introduced to improve the estimation of distribution algorithm,and then propose a hybrid MIMIC algorithm combined with chaos search algorithm.The simulation results show that the proposed algotithm can converge the best optimum of unconstrained and has great improvement in the partial refinement ability in the high dimensional test.Therefore,it is an effective optimization method.Then,based on the CS-MIMIC algorithm,mutative scale optimizing search is introduced to reduce the searching space until the best result has been found.The simulation results show that the proposed algorithm can converge the best optimum of unconstrained and has great improvement of accuracy and stability in the high dimensional test.Minmax algorithm is introduced to transform the constrained optimization problems into unconstrained optimization problems.Then the CS-MIMIC algorithm is applied to solve the unconstained optimization problems.Thesimulation results show that the proposed algorithm can converge the best optimum of constrained optimization problems.Therefore,it is an effective optimization method.The chaos algorithm has the characteristics of fast search and high ergodicity.On this basis,the greedy mode algorithm is introduced,which is called HCGSM algorithm.The aim is to search for the optimal arrangement of a variable so that the probability distribution of the selected optimal solution set is the closest to the defined probability distribution,and the optimal solution can be obtained,thus enhancing the local search ability of the chaos algorithm.The algorithm is applied to unconstrained optimization problems.Numerical experiments show that the algorithm is fast and efficient,and is an effective optimization algorithm.Penalty function method is introduced into HCGSM algorithm,which makes the algorithm applicable to nonlinear constrained optimization problems.Numerical experiments show that the algorithm is fast and efficient,and is an effective optimization algorithm.