Predicting the performance of design-bid-build projects: a neural-network based approach

Abstract

Several studies had shown that many project managers are facing difficulties in predicting the performance of Design- bid-build (DBB) projects. This is due to the fact that there are many factors that affect DBB project success. This research is carried out to identify these factors. In addition, a model to predict the performance of DBB project was developed based on time. Through literature research, a total of forty-four factors that affect DBB project success had been established. The degree of importance for these factors had been determined through questionnaire survey. Eight out of forty-four factors that affecting project performance were found to be the most important factors ITom the viewpoint of project managers and contractors in the Malaysia construction industry. The outcome of the survey formed a basis for the model development. Artificial neural network (ANN) technique is used to construct the models to predict construction project performance based on time. The best performance model was the multiplayer back-propagation neural network model, which consisted of eight input nodes, five hidden nodes and three output nodes. These models were tested by using data ITom nine new projects. The results indicated that the developed model can give a good prediction. In this study, it was concluded that the ANN prediction model can be an efficient tool for predicting the performance ofDBB project from the time aspect.