Performance Comparison Analysis And Application Research Of Complex-value Encoding Moth-flame Optimization Algorithm

The Moth-Flame Optimization Algorithm(MFO)simulates the navigation mechanism of the moth's lateral positioning when flying in the moonlight,it's a new optimization method based on spiral flight.And because of its intuitive structure,easy operation and has strong search ability,the MFO has been widely used in many fields by domestic and foregin scholars.However,as the dimension of the search space increases,the algorithm has the disadvantages of being easily trapped into a local optimum,slow convergence speed and low calculation accuracy,which limits its application on range.In this paper,the original algorithm is improved and analyzed from the aspects of complex-valued encoding methods and multiple strategies,and the improved algorithm is applied to solve complex optimization problems.The purpose is to further improve the performance of the moth-flame optimization algorithm,improve the theoretical basis of the algorithm,and expand its application fields.The main work of this paper are as follows:(1)Using the idea of complex-valued encoding diploid to optimize the coding method of MFO,a complex-valued encoding moth-flame optimization algorithm(CMFO)is proposed,which increases the population diversity of the algorithm,enhances the global search ability of the algorithm,and prevent the algorithm prematurely from falling into a local optimal prematurely.(2)The performance of the complex-valued encoding moth-flame optimization algorithm is compared with the meta-heuristic algorithm of eight complex-coded versions that have appeared in recent years.In the same experimental environment,simulation experiments are performed on 29 benchmark functions(CEC2015),and the experimental results show the pros and cons of the performance of the 9 complex-valued encoding meta-heuristic algorithms.(3)The Harris Hawks optimization algorithm energy segmentation strategy was introduced into the moth-flame optimization algorithm,then proposed a moth-flame optimization algorithm based on flame energy segmentation to make the moth individuals have the perceived flame intensity ability,attempt to enhance the exploration and mining capabilities of the algorithm.Applying ESMFO to benchmark function optimization problems and compared experiments with different algorithms.The experimental results show that the complex-value moth-flame optimization algorithm based on flame energy grading is effective and stable to improv the optimization performance of the algorithm,which further improves the performance of the MFO algorithm.(4)According to the model of artificial intelligence to evaluate software cost,a new evaluation method based on ESMFO algorithm is proposed.This paper provides the NASA training data set to the algorithm to predict the value of the constant variable.After completing the training and obtaining the variable values,the obtained values are applied to the 5 test data set to estimate the software cost.The experimental results show that the proposed algorithm solves the software cost estimation problem with lower error tate and better results than other methods.