Generalization ability of swarm intelligence algorithms

Cheng, Shi1; Qin, Quande2; Liu, Bin3; Feng, Jiqiang4; Lei, Xiujuan1; Shi, Yuhui5

2018

SWARM INTELLIGENCE, VOL 1: PRINCIPLES, CURRENT ALGORITHMS AND METHODS

MICHAEL FARADAY HOUSE, STEVENAGE, HERTS SG1 2AY, ENGLAND

INST ENGINEERING TECH-IET

29-53

In this chapter, generalization ability of swarm intelligence algorithms solving problems with different number of dimensions is analyzed and discussed. Three algorithms, brain storm optimization in objective space (BSO-OS), fireworks algorithm (FWA), and particle swarm optimization (PSO) algorithm, are selected as illustrations to explain the definition of algorithm's generalization ability. The performance of BSO-OS, FWA, and PSO algorithm on solving problems with different number of dimensions is analyzed. Based on the experimental results, the algorithm's generalization ability was measured by the results ratio of algorithms with the same settings on problems with different number of dimensions. This generalization ability measurement could be extended to problems with different components. Without analysis on the landscape of problems, this measurement could give a practical illustration of the generalization ability of algorithms for solving problems with different number of dimensions or different components. Based on the analysis on the generalization of algorithms and the hardness of problems, we could have a better understanding of the relationship between problems and algorithms, and therefore design more effective algorithms to solve different problems.

WOS:000479114400003

英语

BKCI-S

其他

Web of Science

1.Shaanxi Normal Univ, Sch Comp Sci, Xian, Shaanxi, Peoples R China;
2.Shenzhen Univ, Dept Management Sci, Shenzhen, Peoples R China;
3.Nanjing Univ Posts andTelecommun, Sch Comp Sci & Technol, Nanjing, Jiangsu, Peoples R China;
4.Shenzhen Univ, Coll Math & Stat, Shenzhen, Peoples R China;
5.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen, Peoples R China

Cheng, Shi,Qin, Quande,Liu, Bin,et al. Generalization ability of swarm intelligence algorithms. MICHAEL FARADAY HOUSE, STEVENAGE, HERTS SG1 2AY, ENGLAND:INST ENGINEERING TECH-IET,2018:29-53.