Prediction of Blast-Induced Ground Vibration Using Gene Expression Programming (GEP), Artificial Neural Networks (ANNs), and Linear Multivariate Regression (LMR)

摘要

In this paper, an attempt was made to find out two empirical relationships incorporating linear mul-tivariate regression (LMR) and gene expression programming (GEP) for predicting the blast-inducedground vibration (BIGV) at the Sarcheshmeh copper mine in south of Iran. For this purpose, five types ofeffective parameters in the blasting operation including the distance from the blasting block, the burden,the spacing, the specific charge, and the charge per delay were considered as the input data while theoutput parameter was the BIGV. The correlation coefficient and root mean squared error for the LMRwere 0.70 and 3.18 respectively, while the values for the GEP were 0.91 and 2.67 respectively. Also, forevaluating the validation of these two methods, a feed-forward artificial neural network (ANN) witha 5-20-1 structure has been used for predicting the BIGV. Comparisons of these parameters revealed thatboth methods successfully suggested two empirical relationships for predicting the BIGV in the casestudy. However, the GEP was found to be more reliable and more reasonable.