Research On Improved Sparrow Search Algorithm And Its Application

Swarm intelligence algorithm is an optimization algorithm which can share the information of the individuals in the group to carry out subsequent iteration and search.It has achieved good results in many engineering optimization problems.The sparrow search algorithm(SSA)is a new type of swarm intelligence optimization algorithm,which has the advantages of fast convergence speed and easy parameter-adjusted.Once proposed,it has received extensive attention.However,the SSA has the same shortcomings as most swarm intelligent optimization algorithms,that is,it is easy to fall into local optimum,low search accuracy and unbalanced exploration and development capabilities.At the same time,the SSA is a global search algorithm,which can only find a global optimum for optimization problems.Aiming at the above problems,this paper improves the SSA,and proposes a vector encirclement model-based SSA(ISSA)and a multimodal SSA(MMSSA).Further,the ISSA and MMSSA are applied to the trajectory planning and inverse kinematics of the manipulator,respectively.The main contributions of this paper are as follows:(1)An ISSA is proposed.Firstly,a vector encirclement model is proposed and employed to update the position of the scrounger of the SSA,which makes full use of the sparrows with poor fitness value in the sparrow population and greatly improves the convergence speed and optimization accuracy of the SSA.Secondly,based on the idea of genetic algorithm,a producer centralization strategy is proposed,which greatly improves the global search ability of the SSA.Finally,a direction selection strategy is proposed,which is combined with vector encirclement model to enrich the diversity of sparrow population.The experiment on CEC2017 benchmark functions shows that the ISSA has fast convergence speed and high optimization accuracy.In addition,the experimental results on 30 real-world optimization problems demonstrate that the proposed ISSA can solve a wide range of real-world constrained optimization problems successfully with satisfactory performances.(2)A MMSSA is proposed.Firstly,combining the idea of niche and multi-population,a neighborhood search operator is proposed to improve the local search ability of swarm intelligence algorithm.Secondly,by combining the neighborhood search operator with reinforcement learning,a learning neighborhood search operator that could intelligently adjust the size of the neighborhood is proposed.Finally,ISSA is improved into a multimodal optimization algorithm by combining with the proposed learning neighborhood search operator.The experimental results show that MMSSA could handle multimodal optimization problems with satisfactory performances;(3)The application of ISSA in trajectory planning of manipulator is studied.The experiment shows that the method can plan a smooth trajectory satisfying kinematic constraints.Furthermore,the proposed MMSSA is successfully applied to the inverse kinematics problem of the manipulator.the effectiveness of the proposed method is demonstrated by the simulation experiment on the 6DOF manipulator puma560.