Research On Improvement And Application Of Moth-flame Optimization Algorithm

Optimization problems widely exist in the fields of science technology and engineering practice,such as automatic control,mechanical engineering,communication system,bioengineering and so on.Swarm intelligence optimization algorithm is an effective technology to solve complex optimization problems,and it is one of the hotspots of intelligent computing research.Moth-flame optimization(MFO)algorithm is a swarm intelligent optimization algorithm which is proposed to simulate the navigation mechanism of moths in nature.This algorithm has the advantages of simplicity,easy implementation and fast search speed.However,there are some phenomena in the search process,such as poor population diversity and premature convergence.Based on the analysis of the basic principle of MFO algorithm,this paper focuses on the improvement and application of MFO,which has important theoretical and practical significance.The main work of this paper is as follows:Firstly,aiming at the problem of reducing population diversity in the search process of moth-flame optimization algorithm,an improved moth-flame algorithm with dynamic multi-swarm(MIMFO)is proposed.Inspired by the multi-swarms technology in intelligent optimization algorithm,the population of moth is grouped and dynamically reorganized through chaotic grouping mechanism and dynamic reorganization mechanism;spiral search and linear search are carried out for the two sub-swarms of moth;Gaussian mutation is used to generate flame.The effectiveness of MIMFO is verified on the classic test set and Congress on Evolutionary Computation(CEC)2014 test set.Secondly,aiming at the problems of premature convergence,local and global search imbalance of moth-flame optimization algorithm,the MFO algorithm is improved by orthogonal opposite learning strategy,hybrid search mechanism and mutation operation.A moth-flame optimization algorithm with orthogonal opposition-based learning and improved position updating mechanism of moths(OOBLIMFO)is proposed.This algorithm can not only improve the ability of MFO algorithm to jump out of local optimization,but also improve the convergence accuracy and speed of the MFO.Based on the classical test set,CEC 2014 test set and three engineering design problems,the OOBLIMFO is compared with other optimization algorithms.And the results show that OOBLIMFO algorithm has advantages in convergence accuracy and speed.Thirdly,aiming at the problem of single search mechanism leads to low accuracy of moth flame optimization algorithm,the moth-flame optimization algorithm is improved in population initialization,moth position update mechanism and flame generation.A moth-flame optimization algorithm with hybrid search and Gaussian mutation(HSGMMFO)is proposed.Compared with other swarm intelligence optimization algorithms based on the classical test set,the results verify the effectiveness of HSGMMFO.Then,based on the field data,the prediction model of silicon content in hot metal by HSGMMFO optimized fast learning network is established.The results show that compared with other prediction models,the proposed method in this paper can get better results.Fourthly,aiming at the problem of less information exchange between individual dimensions of moth-flame optimization algorithm,considering the interaction between individual dimensions in the moth iteration process,the covariance is used to update the moth position,and Cauchy mutation is used to generate flame.A moth-flame optimization algorithm based on covariance and Cauchy mutation(CCMFO)is proposed.Based on the classical test set,CEC 2014 test set and 57 real-world constraint optimization problems,the effectiveness of CCMFO algorithm is verified.Then,CCMFO is applied to the parameter optimization of the compound controller combining feed forward and feedback control of mold vibration displacement system.And the results show that this method can effectively reduce the difficulty of parameter selection and improve the tracking accuracy.Finally,aiming at the multi-objective optimization problem,Cauchy mutation,dynamic adjustment mechanism and adaptive parameters are used to improve the flame generation,moth position update and parameter adjustment strategy of MFO,and the r dominance relationship to guide the search direction.An improved multi-objective moth-flame optimization algorithm based on r dominance(r IMOMFO)is proposed.The effectiveness of the algorithm is verified by test sets.In addition,r IMOMFO is applied to the control parameter optimization of multi-spacecraft attitude cooperative controller.And the results show that the optimized controller can accelerate the convergence speed and improve the control accuracy.