Efficient behavior factor estimation in moment-resisting reinforced concrete frames through gene expression programming

This study presents a novel approach for estimating the behavior factor of moment-resisting reinforced concrete (RC) frames using a gene expression programming (GEP) method, which involves designing and analyzing over three hundred RC frames. A comprehensive database detailing the specifications of moment-resistant RC frames has been established. This database has several influential parameters as the input parameters. The performance of the developed models was evaluated using statistical indicators, and the best model was determined. The chosen model demonstrated values of 0.0061, 0.049, and 0.0037 for root mean squared error (RMSE), mean absolute error (MAE), and mean squared error (MSE), respectively. Additionally, the R2 values for the training and test data were 0.93 and 0.82, respectively. Finally, a highly accurate mathematical equation was obtained to predict the behavior factor of the RC frames using GeneXpro Tools software. After sensitivity analysis of the behavior factor predicted to the investigated parameters, the results indicated that seismic conditions have minimal impact on the behavior factor of moment-resisting RC frames. The number of stories has an inverse relationship with the behavior factor, while the impact of changing the span length ratio to story height on the behavior factor is not uniform. The study's findings indicated that the GEP method effectively predicted the behavior coefficient of RC frames.