Abstract: In order to improve the solution accuracy and global search capability of Salp swarm algorithm(SSA),an improved Salp swarm algorithm based on normal process search and differential evolution algorithm is proposed,called two types of leaders salp swarm algorithm(TTLSSA).Two types of leaders and two following groups are set up in the algorithm.Among them,the leader performing normal process search needs to carry out normal process migration,crossover,selection and other operations,which are mainly used for global exploration.Under the influence of the gap parameter,which varies with the number of iterations,the leader near the current optimal solution combines both global search and local development functions.Eighteen different types of standard test functions are used to test the performance of the proposed algorithm,compared with DE,SSA,sines and cosines algorithm (SCA),grey wolf optimizer (GWO),and whale optimization algorithm (WOA).TTLSSA ranks the first or joint first in the average precision of 16 test functions,the second in the average precision of 2 test functions,and the second in the ave-rage time of 6 algorithms,indicating that TTLSSA significantly improves the optimization ability without increasing the time cost of SSA.

Key words: Differential evolution, Normal process, Salp swarm algorithm, Test functions