A guided evolution strategy for discrete sizing optimization of space steel frames

Korucu A., Hasançebi O.

Structural and Multidisciplinary Optimization, cilt.66, sa.8, 2023 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 66 Sayı: 8
  • Basım Tarihi: 2023
  • Doi Numarası: 10.1007/s00158-023-03640-7
  • Dergi Adı: Structural and Multidisciplinary Optimization
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Compendex, INSPEC, zbMATH
  • Anahtar Kelimeler: Design-driven search methods, Discrete sizing optimization, Guided evolution strategy, Metaheuristic search techniques, Space steel frames, Structural optimization
  • Orta Doğu Teknik Üniversitesi Adresli: Evet

Özet

In this paper, a new design-driven hybrid optimization algorithm called guided evolution strategy (GES) is proposed for a reliable and rapid optimum design of space steel frames. The rationale behind the proposed GES algorithm is to improve convergence characteristics of the evolution strategies (ESs) optimization method by guiding search process according to the satisfaction/violation of strength constraints in a previous design. This is referred to as guided mutation, which is introduced as an auxiliary tool to a stochastic mutation scheme for accelerating the convergence speed of the optimization algorithm. The efficiency of the GES algorithm is investigated and quantified using design examples where sizing optimization of two space steel frames are achieved under strength and displacement constraints imposed according to ANSI/AISC 360-10 (Specification for structural steel buildings, ANSI/AISC 360-10, Illinois, 2010) and ASCE/SEI 7-10 (Minimum design loads for buildings and other structures, ASCE/SEI 7-10, Reston, 2010) design specifications. The solutions produced to these design examples with the GES algorithm are compared to those of some selected metaheuristic search techniques in terms of accuracy of the obtained solutions as well as speed of convergence to the optimum designs. It is shown that the GES algorithm has improved search abilities with respect to other employed techniques.