Seismic loss optimum design of steel structures using learning-based charged system search

One of the structural engineers' challenges on a regular basis is balancing the expense of initial construction with the cost of future structural loss. At first appearance, employing the optimization method seems to be a viable option. However, since both the structure's analysis and design, as well as the computation of the cost of future loss, are time-consuming and expensive, combining these processes with the costly optimization progression is prohibitively expensive. The purpose of this study is to present a methodology for the risk-based optimum design of steel frame structures, as well as to enhance a previously published metaheuristic algorithm for a better optimization approach. The methodology given here allows structural designers to account for seismic risks in the design optimization process without incurring expensive computing expenditures. This approach may be appealing for practical work since it minimizes time-consuming charges and provides designers with a structural impression. Furthermore, as compared to the standard version, the new optimization algorithm improves performance while decreasing computing costs. Bayesian linear regression is used in conjunction with a parameter identification challenge to derive probabilistic models for estimating structural analysis demand responses. The minimum amount of total initial cost and seismic loss cost is regarded as the objective function of the design of three chosen mid- to high-rise moment frames for the optimization purpose. The results demonstrated enhanced optimization performance as well as a decreased loss cost for employed structures.