Hybrid genetic algorithm for optimal seismic design

The genetic algorithm (GA) is widely used in the optimal structural design field due to the excellent global exploration capacity and applicability of it. Although GA is very robust and has various advantages, it is very computationally intensive and poor at the local search. To improve this problem, this study presents a hybrid optimization method in which a local search operator based on the resizing design method is embedded in the framework of GA. The resizing method is to redesign the sectional properties of corresponding elements based on the displacement participation factor of elements consisting of buildings without the repetitive structural analysis. This improves the stiffness of the building through efficiently redistributing the structure material. The GA is efficiently used in a global exploration, while the resizing design method is used in a local exploitation. This study uses the NSGA-II, which is a kind of GA, to optimize the multi-objective functions. The efficiency of this hybrid optimization method was investigated using the steel moment frame example. The result showed that the hybrid method had great improvement on the convergence rate than the original NSGA-II.