Research On Firefly Algorithm For Multimodal Optimization Problems

In 2008,the University of Cambridge scholar Yang proposed the Firefly Algorithm(FA)in 2008.The firefly optimization algorithm is also a group intelligent optimization algorithm.Many scholars have proved that it has many unique optimization problems and application fields with its unique internal evolution mechanism,better local search ability,simple mathematical model and so on.The optimized performance is no less than other cluster intelligent optimization algorithms.However,there are not many related researches on firefly algorithm on some complex optimization problems,such as multi-model function optimization.There are still some problems,such as low peak discovery rate,insufficient optimization precision,slow convergence rate and so on.In view of this,this thesis has made research on the firefly algorithm,there are three main points:(1)The firefly algorithm design based on species-based mechanism improves the ability of the algorithm to solve multi-modal optimization problems.In this paper,two mechanisms for generating multiple species are proposed.The first one is a species-based mechanism based on fixed radius.The best individual of each sub-species is the seed,and all fireflies within a fixed radius around the seed belong to the same sub-species.The second type is a multi-modal mechanism based on the neighborhood number.All firefly individuals are numbered in a ring topology,and the firefly individuals with the serial number within the radius of the neighborhood of each seed belong to the same sub-species.Ten standard test function experiments show that the firefly algorithm based on species has a better improvement in accuracy,convergence speed and success rate than the standard firefly algorithm.(2)A species-based firefly algorithm design that incorporates a local search mechanism.For the lack of precision of firefly algorithm,three local search mechanisms are introduced.The first is the hill climbing search strategy,the second is simulated annealing search strategy,and the third is adaptive search strategy.Three local search mechanisms are added to the species-based firefly algorithm.Standard test function experiments show that the species-based firefly algorithm with local search mechanism solves the multi-modal optimization problem more effectively.(3)Applying the improved firefly algorithm to a simple PCB board to identify resistance problems,design experiments,using template matching techniques to find and locate the feature areas and target components we need in the target image.Thereby verifying the effectiveness of the firefly optimization algorithm and its improvement.Reflecting the advantages of the intelligent optimization algorithm,it also proves the feasibility of the improved firefly optimization algorithm in dealing with practical problems.