Assessing the suitability of imperialist competitive algorithm for the predicting aims: an engineering case

Abstract

In the surface and underground mines as well as civil projects, the blasting operation is widely performed for rock breakage. Flyrock is considered as an undesirable environmental impact induced by blasting. Hence, precise prediction of flyrock is a necessary work for safety issue. This research is carried out to evaluate the acceptability of imperialist competitive algorithm (ICA) to approximate the blast-induced flyrock with respect to input parameters including burden, spacing, stemming, weight charge and rock mass rating. In total, 78 blasting operation were investigated and the mentioned parameters, as well as the flyrock values, were measured. In this research work, three ICA-based models, i.e., linear, power, and quadratic models, are introduced. The artificial neural network (ANN) has also been developed by the same data sets and the same input parameters which we used in ICA. The results of the predictors are then evaluated using statistical indicators such as coefficient of determination (R2). Finally, it was proved that the ICA–linear yields a better prediction in comparison with three other models, so that R2 was obtained as 0.954, while the amount of R2 for the ICA–power form, ICA–quadratic form, and ANN models were 0.928, 0.952, and 0.841, respectively.