Analysis And Application Research Of Crow Search Algorithm And Dwarf Mongoose Optimization Algorithm

Crow Search Algorithm(CSA)and Dwarf Mongoose Optimization(DMO)are two new swarm intelligence optimization algorithms proposed in recent years.However,the two algorithms still have shortcomings such as premature convergence and easy to fall into local optimum.Based on the above problems,this paper carried out research on the improvement and application of the crow search algorithm and the dwarf mongoose optimization algorithm,and achieved certain research results.The research results of this paper are as follows:(1)Propose a new improved crow search algorithm(EICSA)oriented to the memory location of excellent individuals.based on the amount of food stored by the individual,most individuals in the population are classified as ordinary individuals,and a few individuals with a large amount of stored food are classified as for excellent individuals.Excellent individuals only carry out local search activities near their food storage nests;most ordinary individuals are guided by the food storage nests of excellent individuals,and conduct global exploration with a large step size in the early stage of the algorithm to maintain the diversity of the population;Local development with a shorter step size enhances both the global exploration and local development capabilities of the algorithm.The optimized performance of EICSA is verified by numerical experiments and simulations.(2)A dwarf mongoose optimization algorithm(FADMO)for assigning search tasks based on foraging ability is proposed.First,the tent chaos adaptive step size is used to balance the global search and local development;to solve the blindness problem of the alpha group search,optimize its moving direction and moving ability;to solve the misleading problem of the moving direction of the reconnaissance group algorithm,enhance its individual error correction ability to improve Individual foraging ability;improving the babysitters group moving algorithm to optimize and improve the local development ability of the algorithm;finally,a new population foraging strategy is proposed,which balances the calling strategies between algorithms and improves the overall performance of the algorithm.The optimized performance of FADMO is verified by numerical experiments and simulations.(3)Apply FADMO to solve the parameter optimization problem of support vector machine.Through numerical experiments on six benchmark datasets,the results show that FADMO performs well in solving parameter optimization problems of support vector machines.