Abstract: Aiming at the shortcomings of slow convergence speed and poor calculation accuracy in the optimization process of Salp Swarm Algorithm (SSA), a Novel Salp Swarm Algorithm is proposed (NSSA). Firstly, the deficiencies of SSA in the process of following leaders are analyzed. learning from the idea of grey wolves following the header in grey wolf algorithm, the way of salps following the leader is improved. Performance comparison between NSSA and other algorithms is conducted on 23 benchmark functions, and the algorithm is applied in image matching. All experimental results show that NSSA has better convergence speed, calculation accuracy and robustness.

Key words: salp swarm algorithm, function optimization, image matching, grey wolf algorithm