Research On The Improvement And Application Of Salp Swarm Algorithm

Salp Swarm Algorithm(SSA)is a new swarm intelligent optimization algorithm proposed by Mirjalili et al.in 2017 via simulating the swarming behaviours of salps when moving and foraging in deep oceans.The algorithm has the advantages of simple structure,few parameters and easy operation.However,similar to other swarm intelligence algorithms,SSA also has some disadvantages such as slow convergence speed,loss of population diversity in the late iteration,and the disadvantages make it difficult to balance its exploration and exploitation ability.Therefore,this thesis improves the SSA and widens its application scope.The specific work is as follows:(1)An improved salp swarm algorithm is proposed and applied to welding beam problem.In order to improve the solving accuracy and convergence speed of salp swarm algorithm,an improved salp swarm algorithm is proposed.Firstly,elite opposition-based learning strategy is applied to leader individual to balance the exploration and exploitation ability of the algorithm.Then,in order to improve the accuracy of the algorithm,a difference strategy is introduced to update the follower position inspired by the difference evolution.Finally,Gaussian mutation of food location is carried out in the search process to avoid falling into local optima,laying a foundation for the global search for the algorithm.Experiment results on 13 standard test functions and a classical engineering problem show that the search performance of the improved salp swarm algorithm is better than that of the contrast algorithm.(2)Salp swarm algorithm based on laplacian crossover and disruption operator is proposedIn order to enhance the exploration and exploitation ability of salp swarm algorithm,an improved salp swarm algorithm based on Laplace crossover and Disruption operator was proposed.The Laplace crossover operator is introduced into the leader position update to enhance the local exploitation ability of the algorithm.Introduce Disruption operator in follower position update to balance the global exploration and local exploitation ability of the algorithm;Finally,the improved algorithm is compared with other five swarm intelligent algorithms on CEC 2014 benchmark function.The results show that the improved algorithm has high accuracy and fast convergence speed,especially in solving complex multimodal functions.Wilcoxon test and performance index analysis of the algorithm show that improve algorithm of this paper has better performance.(3)Multi-strategy ensemble salp swarm algorithmin for robot path planning is proposedA multi-strategy ensemble salp swarm algorithm is proposed for solving problem of robot path planning.In the algorithm,a new adaptive leader structure is proposed to balance the exploration and exploitation ability of the algorithm;The cascade of chaotic system can improve the Lyapunov exponent of the cascade chaotic system,so the chaotic map of Logistic-Cubic cascade is introduced as the disturbance operator of the food source to avoid the algorithm falling into the local optimum;A disperse foraging strategy based on adaptive parameters is adopted to force a part of followers to explore promising areas.The algorithm in this paper is compared with three improved SSA algorithms and seven advanced swarm intelligence algorithms on CEC 2014 functions.The results show that the comprehensive optimization performance of the algorithm in this paper is better.The proposed algorithm is applied to solve the robot path planning problem,in which the path is smoothed by cubic spline interpolation.Simulation experiments are implemented on computer in the environments where the obstacles are 8,9,13,respectively.The results demonstrate that the proposed algorithm can achieve the best results compared with the given contrast algorithms in given simulation scenarios.