An intelligent method for generating artificial earthquake records based on hybrid PSO–parallel brain emotional learning inspired model

Simplicity and high speed of brain emotional model made it an effective computational method, which is used in various applications. In this study, a modified model of brain emotional learning is used for generating artificial earthquake records. In fact, in earthquake engineering, strong ground motions are valuable information, which are recorded in each earthquake. These records could be used for linear and nonlinear time-history analysis of structures. Unfortunately, the numbers of recorded strong ground motions are not enough for most areas of the world. Therefore, many seismic codes permit to use artificial records, which contain specific characteristics. Because of the advantages of emotional models, a hybrid PSO–parallel brain emotional learning model is used for generating artificial records based on a dataset of real records in this research. PSO algorithm is combined with the model for finding the best values of learning parameters. In addition, wavelet packet transform is used for decomposing the earthquake signal to use as the suitable output of network. Despite of original brain emotional model, the proposed modified parallel model is applicable on multiinput–output data. Numerical examples show that the proposed model in this research could be successfully used for generating artificial records with acceptable error for statistical properties of the required pseudo spectral acceleration.