Research On Improved Moth-flame Optimization Algorithm

The function optimization problem exists widely in various engineering fields.Therefore,the research on optimization algorithm has important theoretical and practical significance.The group intelligent optimization algorithm is the research hotspot in the current optimization algorithm.The Moth-Flame Optimization(MFO)algorithm,as a new group intelligent optimization algorithm,has gradually attracted the attention of scholars.In this thesis,a problem of premature convergence and inability to quickly converge in the MFO algorithm is proposed.A Moth-flame optimization algorithm fused on refraction principle and opposite-based learning(Moth-flame optimization algorithm)is proposed.ROBL-MFO).Firstly,The algorithm introduces the historical optimal flame average in the moth update formula,so that the information between the flames can communicate with each other,the historical optimal flame average is used to ensure the flame quality,and the quality of the moth update is also ensured,thereby improving the convergence speed and the optimization ability of the algorithm;secondly,the random opposite-based Learning strategy is introduced.The random opposite-based Learning strategy can be used to expand the advantage of the algorithm search space,and the random solution of the flame solution can be searched for a better flame solution and avoid the premature convergence of the algorithm;finally,using the refraction principle to refract the obtained inverse solution,it could find other possible possible solutions to improve the diversity of the population and help the algorithm to jump out of the local optimum.In this thesis,Using simulation experiments on multiple benchmark functions to evaluate the performance of ROBL-MFO algorithm.Compared to other comparison algorithms,The experimental results show that the ROBL-MFO algorithm has better convergence speed and optimization precision on the type of test function,and the ROBL-MFO can effectively jump out of local optimum and avoid the premature convergence.Finally,this thesis summarizes the overall research work,and shows that in the next stage,we will continue to study the moth-flame optimization algorithm and expand its application field to improve its application value.There are 16 figures,9 tables and 57 references in this thesis.