NSCSO: a novel multi-objective non-dominated sorting chicken swarm optimization algorithm

• Published in: Scientific reports Vol. 14; no. 1; pp. 4310 - 38
• Main Authors: Huang, Huajuan, Zheng, Baofeng, Wei, Xiuxi, Zhou, Yongquan, Zhang, Yuedong
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 21.02.2024; Nature Publishing Group; Nature Portfolio
• Subjects: 639/705/1042; 639/705/1046; 639/705/117; 639/705/258; Algorithms; Chicken swarm optimization algorithm; Fast non-dominated sorting; Humanities and Social Sciences; Meta-heuristic; Multi-objective engineering design problems; Multi-objective optimization; multidisciplinary; Objective function; Optimization algorithms; Science; Science (multidisciplinary); Multi-objective engineering design problems; Multi-objective optimization; Meta-heuristic; Chicken swarm optimization algorithm; Fast non-dominated sorting
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-024-54991-0

Addressing the challenge of efficiently solving multi-objective optimization problems (MOP) and attaining satisfactory optimal solutions has always posed a formidable task. In this paper, based on the chicken swarm optimization algorithm, proposes the non-dominated sorting chicken swarm optimization (NSCSO) algorithm. The proposed approach involves assigning ranks to individuals in the chicken swarm through fast non-dominance sorting and utilizing the crowding distance strategy to sort particles within the same rank. The MOP is tackled based on these two strategies, with the integration of an elite opposition-based learning strategy to facilitate the exploration of optimal solution directions by individual roosters. NSCSO and 6 other excellent algorithms were tested in 15 different benchmark functions for experiments. By comprehensive comparison of the test function results and Friedman test results, the results obtained by using the NSCSO algorithm to solve the MOP problem have better performance. Compares the NSCSO algorithm with other multi-objective optimization algorithms in six different engineering design problems. The results show that NSCSO not only performs well in multi-objective function tests, but also obtains realistic solutions in multi-objective engineering example problems.