Research And Application Of Improved Moth-flame Optimization Algorithm

Optimization problems widely exist in various engineering fields,and historical methods to solve such problems have many disadvantages,such as large amount of computation and long computing time.Because of its intelligence and essential parallelism,the swarm intelligence optimization algorithm shows a powerful advantage in solving optimization problems.The mothflame optimization algorithm is widely used in complicated industrial fields because it has the characteristics of less parameters to be adjusted and low requirements for running environment compared with many similar algorithms.However,with the deepening of the research,the defects of the moth-flame optimization algorithm are gradually exposed.For example,it is often to find a local optimal solution and the rate of finding the result is slow.In this paper,the moth-flame optimization algorithm is studied,and the work includes:(1)Aiming at the disadvantage that the moth-flame optimization algorithm is easy to sink into a local extreme point,the advantage of effective search space can be expanded by using reverse learning strategy,and two-dimensional reverse learning can be carried out for moths;at the same time,the out-of-bounds reset strategy is introduced into the algorithm,and the individuals who exceed the solution space in the iterative process are randomly reset to prevent the decrease of population diversity caused by the moth's out-of-bounds;Chaos theory is applied to the optimization process of the algorithm,and new positions are added to the flame by Tent chaotic mapping,which improves the possibility of the algorithm searching for the global optimum.After combining the three strategies,a moth-flame optimization algorithm based on reverse and mapping strategies is proposed(RMMFO).(2)In view of the slow convergence speed of RMMFO,chaotic initialization is used to generate a more uniform initial population,thereby improving the quality of initial solution;self-adaptive weight is introduced into the moth's position updating formula,which makes the spiral flight mode the main factor guiding the moth's flight path in the early stage of searching,and the flame position in the later stage becomes the main factor guiding the moth's flight path,effectively balancing the global and local searching ability of the algorithm.The two strategies are integrated into the RMMFO algorithm,and a fusion multi-strategy moth-flame optimization algorithm is proposed(FMSMFO).(3)Tests are performed on multiple functions to judge the performance of the FMSMFO algorithm in various aspects.Compared with the other five excellent algorithms,it is found that FMSMFO algorithm has higher accuracy and faster speed when solving different types of functions.(4)The FMSMFO algorithm is applied to the array pattern.By optimizing the array antenna with low sidelobe and deep zero point,it is verified that the FMSMFO algorithm can be practical in the engineering field and has a certain degree of innovation.