Optimal design of steel frames under seismic loading using two meta-heuristic algorithms

In this paper, optimal design of steel frames is performed under seismic loading. The variables of the problem are taken as the cross-sectional areas of the members. These variables are considered as discrete, and are selected from a list of existing cross sections. Here, the charged system search and improved harmony search algorithms are utilized for optimization. For optimal design of steel frames in the first phase a time history analysis with the relative lateral displacement constraints is performed, and in the second phase a simultaneous dynamic–static analysis with the relative displacement and stress constraints is utilized using two meta-heuristic algorithms. Moment frames and their shear frame counterparts are considered, and their performances are compared for optimal design. In the case of moment frames, apart from the columns, the cross sections of the beams are also considered as design variables. The results indicate a good performance of the optimized moment frame and show that considering the effect of both drift and stress constraints, instead of only drift constraints, one obtains a better design. These results also show the suitability of the charged system search algorithm for optimal design of frames under seismic loading, as an extremely nonlinear problem.