Parametric cost estimate of forming and placing of concrete using neural network

Neural Networks have found their way to many applications in construction. One of the most common applications of neural networks in construction is to develop parametric estimates of construction projects or specific construction operations. Due to their versatility and ability to handle fuzziness, they have performed well in estimating specific construction operations for which cost is dependent on specific parameters. This paper presents a back-propagation Neural Network (NN) for the development of a parametric cost-estimating model of concrete forming and placement using a commercial forming system (Steel-Ply ). The main objective is to develop a neural network cost-estimation model and verify its accuracy using actual data. Actual project data, from a local contractor in western Illinois, was used to develop the NN model. The model was developed and optimized on a spreadsheet format. Parameters considered include the season of the operation, the wall thickness and height, the method of placement, and the shape index of the structure. The same data used to develop the NN cost-estimating model is used to perform a linear regression analysis to predict the cost of forming concrete. Outputs of the developed NN model were compared with estimates obtained from multiple linear regression models. The results indicate that the back-propagation NN model can be used satisfactorily to estimate the forming and placing of concrete. Furthermore, practitioners as well as students can use the developed NN model to learn about mechanism of neural networks.