Optimization of steel roof trusses using machine learning-assisted differential evolution

  • Nguyen Tran Hieu Faculty of Buildings and Industrial Constructions, Hanoi University of Civil Engineering, 55 Giai Phong road, Hai Ba Trung district, Hanoi, Vietnam
  • Nguyen Quoc Cuong Faculty of Buildings and Industrial Constructions, Hanoi University of Civil Engineering, 55 Giai Phong road, Hai Ba Trung district, Hanoi, Vietnam
  • Vu Anh Tuan Faculty of Buildings and Industrial Constructions, Hanoi University of Civil Engineering, 55 Giai Phong road, Hai Ba Trung district, Hanoi, Vietnam
Keywords: structural optimization, Differential Evolution, machine learning, AdaBoost, steel roof truss

Abstract

A steel truss is a preferred solution in large-span roof structures due to its good attributes such as lightweight, durability. However, designing steel trusses is a challenging task for engineers due to a large number of design variables. Recently, optimization-based design approaches have demonstrated the great potential to effectively support structural engineers in finding the optimal designs of truss structures. This paper aims to use the AdaBoost-DE algorithm for optimizing steel roof trusses. The AdaBoost-DE employed in this study is a hybrid algorithm in which the AdaBoost classification technique is used to enhance the performance of the Differential Evolution algorithm by skipping unnecessary fitness evaluations during the optimization process. An example of a duo-pitch steel roof truss with a span of 24 meters is carried out. The result shows that the AdaBoost-DE achieves the same optimal design as the original DE algorithm, but reduces the computational cost by approximately 36%.

Downloads

Download data is not yet available.
Published
31-10-2021
How to Cite
Hieu, N. T., Cuong, N. Q., & Tuan, V. A. (2021). Optimization of steel roof trusses using machine learning-assisted differential evolution. Journal of Science and Technology in Civil Engineering (STCE) - HUCE, 15(4), 99-110. https://doi.org/10.31814/stce.huce(nuce)2021-15(4)-09
Section
Research Papers