TY - JOUR
T1 - Optimum design of steel building structures using migration-based vibrating particles system
AU - Talatahari, Siamak
AU - Jalili, Shahin
AU - Azizi, Mahdi
N1 - Acknowledgment
This research is supported by a research grant of the University of Tabriz (Number: 1615). We sincerely express our gratitude to Assoc. Prof. Saeid Kazemzadeh Azad for providing the required data for the design examples.
PY - 2021/10/1
Y1 - 2021/10/1
N2 - Vibrating Particles System (VPS) is developed based on some principles of physics in which the free vibration of a system with single degree of freedom including viscous damping is concerned. In this algorithm, each possible solution or vibrating particle seeks its equilibrium position in the search space. Despite of a relatively good exploration ability of the VPS algorithm, it is poor at exploitation and the convergence speed of this algorithm is also an issue in some cases. In this paper, the VPS algorithm is hybridized with the Migration-Based Local Search (MBLS) mechanism of the Biogeography-Based Optimization (BBO) algorithm with the strong local search capability to concentrate the search process around promising vibrating particles and locate the optimum solution more precisely. Three hybrid algorithms are developed based on how to use the VPS and MBLS methods as parallel, series and mixed series–parallel schemes. In order to evaluate the capability of the proposed hybrid methods in dealing with difficult structural optimization problems, a 24-story benchmark frame problem, a 10-story steel structure with 1026 structural members alongside a 20-story steel structure with 3860 members are optimized using presented algorithms. The findings affirm the robustness and efficiency of the proposed hybrid methods over the standard existing relevant approaches for optimum design of steel building structures.
AB - Vibrating Particles System (VPS) is developed based on some principles of physics in which the free vibration of a system with single degree of freedom including viscous damping is concerned. In this algorithm, each possible solution or vibrating particle seeks its equilibrium position in the search space. Despite of a relatively good exploration ability of the VPS algorithm, it is poor at exploitation and the convergence speed of this algorithm is also an issue in some cases. In this paper, the VPS algorithm is hybridized with the Migration-Based Local Search (MBLS) mechanism of the Biogeography-Based Optimization (BBO) algorithm with the strong local search capability to concentrate the search process around promising vibrating particles and locate the optimum solution more precisely. Three hybrid algorithms are developed based on how to use the VPS and MBLS methods as parallel, series and mixed series–parallel schemes. In order to evaluate the capability of the proposed hybrid methods in dealing with difficult structural optimization problems, a 24-story benchmark frame problem, a 10-story steel structure with 1026 structural members alongside a 20-story steel structure with 3860 members are optimized using presented algorithms. The findings affirm the robustness and efficiency of the proposed hybrid methods over the standard existing relevant approaches for optimum design of steel building structures.
KW - Optimum design
KW - Steel structure
KW - Vibrating particles system
KW - Migration-based local search
KW - Migration-based vibrating particles system
U2 - 10.1016/j.istruc.2021.05.028
DO - 10.1016/j.istruc.2021.05.028
M3 - Article
VL - 33
SP - 1394
EP - 1413
JO - Structures
JF - Structures
SN - 2352-0124
ER -