Comparative cost analysis of using high-performance concrete in tall building construction by artificial neural networks

Artificial neural networks are used in this investigation to establish the relationship between the quantities/costs of concrete and form-work required for the structural elements of high-rise commercial buildings (including solid slabs, beams, columns and shear walls, and the entire structure) and the design variables (grid sizes, number of stories, and grades of concrete). Two neural network-based schemes - hierarchical and hybrid predictions on cost estimation - are compared. The fast back-propagation algorithm is used for training the feed-forward network. After training, the neural network models have been proven to be accurate in predicting the costs of using high-performance concrete in wall-frame structures for high-rise building construction. Verifications are also conducted using a separate set of design parameters. The paper concludes with a comprehensive discussion on the prediction results.