An adaptive quadratic interpolation and rounding mechanism sine cosine algorithm with application to constrained engineering optimization problems
Xiao Yang,Rui Wang,Dong Zhao,Fanhua Yu,Chunyu Huang,Ali Asghar Heidari,Zhennao Cai,Sami Bourouis,Abeer D. Algarni,Huiling Chen
DOI: https://doi.org/10.1016/j.eswa.2022.119041
IF: 8.5
2023-03-01
Expert Systems with Applications
Abstract:The sine cosine algorithm (SCA) is a well-known meta-heuristic optimization algorithm. SCA has received much attention in various optimization fields due to its simple structure and excellent optimization capabilities. However, the dimension of objective function also increases with the increasing complexity of optimization tasks. This makes the original SCA appear to have insufficient optimization capability and likely to fall into premature convergence. A multi-mechanism acting variant of SCA, called ARSCA, is proposed to address the above deficiencies. ARSCA is an enhanced SCA algorithm based on the adaptive quadratic interpolation mechanism (AQIM) and Rounding mechanism (RM). RM enables a more balanced state between exploration and exploitation of the ARSCA. AQIM enhances local exploitation capabilities. To verify the performance of ARSCA, we compared ARSCA with some advanced traditional optimization algorithms and variants of algorithms for 30 consecutive benchmark functions of IEEE CEC2014. In addition, ARSCA was applied to 6 constrained engineering optimization problems. These six algorithms include the tension–compression spring design problem, the welded beam design problem, the pressure vessel design problem, the I-beam design problem, the speed reducer design problem, and the three-bar design problem. Experimental results show that ARSCA outperforms its competitors in both the solution quality and the ability to jump out of the local optimum. The relevant codes for the paper are publicly available at https://github.com/YangXiao9799/paper_ARSCA.
computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science