Research And Application Of Fusion Equilibrium Optimal Algorithm Based On Mutated Intelligent Chaotic Swarm

At present,the common intelligent optimization algorithms generally have some defects due to the imbalance of search mode leverage.For example,for the chaotic optimization algorithm(COA)with strong global search performance,it fails to make full use of the characteristics of chaotic system,and the contradiction between the global search of chaotic system and the strong convergence strategy has not been effectively mediated,which leads to the poor computational performance of COA For swarm intelligence optimization algorithm,because of its strong convergence strategy and limited global exploration strategy,the possibility of the algorithm falling into local optimum increases.There is also a kind of fusion intelligent optimization algorithm which attempts to combine the global exploration and convergence strategy,but the contradiction between the two strategies can not be well reconciled because of the design problems of the algorithm framework.That makes the purpose of the fusion algorithm to complement each other difficult to achieve or the effect is not significant.The above problems affect the performance of intelligent optimization algorithm in engineering applications.Therefore,it is necessary to choose a reasonable swarm intelligence convergence strategy,from the perspective of search equilibrium,consider to design a fusion algorithm framework based on the improved chaotic optimization algorithm to be compatible with the strong global exploration characteristics of chaos optimization algorithm and the strong convergence strategy of swarm intelligence optimization algorithm.The main contents and conclusions are as follows:(1)Through the research and comparison of chaotic systems,the nonlinear term and "Divided" modular operation are introduced into Lorenz system successively to realize the complexity,discretization and homogenization of Lorenz system,and a discrete chaotic system with uniform distribution,high stability,more controllable disturbance parameters and disturbance tolerance is obtained.Through simulation test,the DMLCS system has better performance than the contrast group chaotic systems,and the system is combined with arcsin-logistic map as the disturbance core module for initial solution generation and iterative disturbance.(2)Based on the designed disturbance core,firstly,the basic recarrie global COA framework is implemented.Based on this framework,an improved variable scale density strategy,search center fine-tuning strategy and solution space oscillation strategy are introduced,and a variable density divided mutated lorenz chaotic Interval Oscillation search algorithm(DMLCOA)is improved and implemented.The algorithm has satisfied search ability in dimensions less than 20,however,the algorithm shows uneven search patterns in higher dimensional tests.(3)In order to achieve more balanced local convergence and global exploration performance,a fusion algorithm is designed based on dmlcoa and swarm intelligence optimization strategy.Aiming at the problem that WOA algorithm has difficult solving some cases,it is proved by inference experiments that its partial swarm intelligence behavior strategy causes zero preference trap;therefore,only the strong convergence performance strategy without preference in WOA is extracted,and the strategy and disturbance core are integrated into a new compatible algorithm framework.The framework transformed the main function of chaotic system,gives full play to the performance of chaotic system in solution disturbance,and effectively reconciles the contradiction between the core of chaotic disturbance and swarm intelligence convergence strategy,and a dicreted chaotic swarm oscillation fusion search algorithm with spiral strategy(DCSOA-S)is obtained.The algorithm shows higher search equilibrium and better global steady convergence performance than the comparison group in the general case test,and it can find the global optimal solutions of all general test functions with higher accuracy without the search preference trap.In addition,the economic load distribution(ELD)problem and Kapur entropy multi threshold image segmentation(IS)problem are selected as engineering application cases.The DCSOA-S algorithm is used to solve the problems and compared with the group of other algorithms.The results show that the algorithm has excellent engineering ability and transformation value.