To address the problems of the standard salp swarm algorithm that it is prone to local optimum with slow convergence in the process of finding the best,the hybrid multi-strategy improved salp swarm algorithm(ISSA)was proposed.Initial popula-tions were generated using the good point set strategy,such that individuals were uniformly distributed in the search space.The idea of opposition-based learning was incorporated into leader position updates to improve the search accuracy of the algorithm.The adaptive t distribution was added,and the number of iterations iter was employed as the degree of freedom parameter to boost exploration development capability of the algorithm.Elite opposition-based learning was used to choose better populations and avoid falling into local optimum.The performance of the improved algorithm was tested by a set of benchmark functions and Wilcoxin rank sum test.Experimental results show that the exploration ability and optimization accuracy of the improved algo-rithm are significantly improved and there are significant differences between the algorithms.The effectiveness of ISSA is further verified bv a real mechanical design case.
关键词
佳点集/反向学习/自适应t分布/精英反向学习/樽海鞘群算法/基准函数/弹簧设计问题
Key words
good point set/opposition-based learning/adaptive t distribution/elite opposition-based learning/slap swarm algo-rithm/benchmark functions/spring design problem
维普
万方数据