Application of a perceptron artificial neural network for building the stability of a mining process

Abstract

The paper describes a method of ensuring the stability of the selected mining process (loading and haulage of copper ore) taking place under variable environmental conditions. Four models of a multilayer perceptron neural network were built for this purpose. Travel times and the condition of transport roads were adopted as input parameters. The output of the network is the cycle time of the analysed process. On the basis of an analysis of learning errors, a model with two hidden layers was selected. A series of experiments was conducted on the selected model. An assessment was also performed to determine at which values of input parameters the stability of the analysed process could be ensured. © 2012 Springer-Verlag.