Akademik Veri Yönetim Sistemi
NEURAL COMPUTING & APPLICATIONS, cilt.34, sa.16, ss.13825-13848, 2022 (SCI-Expanded)
In this study, to improve the algorithmic performance of standard grasshopper optimization algorithm (GOA) that mimics the foraging characterizations of grasshoppers as in swarms, four modified versions of which are proposed to solve large-scale real-size complex both truss and frame type steel structures. All proposed GOA encoded are supplied reciprocatively data transfer with structural analysis software (SAP2000) to easily get the structural responses via open application programming interface functions. Initially, the algorithmic performances of standard GOA and its proposed modified versions are evaluated on two small-size benchmark engineering design problems, namely pressure vessel and grain train design problems. And then, a 160-bar space steel pyramid, a 693-bar space braced steel barrel vault, and a 455-member spatial braced steel frame are considered as large-scale real-size structural design problems. These structures are optimally designed to reach the best feasible economic structures with minimum design weights while satisfying the structural behavior limitations such as displacement, drift, strength, and stability that are taken from specifications of the American Institute of Steel Construction-Load and Resistance Factor Design. Consequently, the performances of all proposed GO algorithms in finding the optimum designs of large-scale real-size steel structures are compared and evaluated in detail.