Development of prediction model for delay and cost overruns in the bridge construction project by using Artificial Neural Network

Bridge construction projects frequently faces delays and cost overruns due to various uncertain factors. Traditional risk assessment methods are limited in their ability to analyse complex data and make precise predictions. There is necessity of advanced techniques such as Artificial Neural Network to increase the efficiency in the risk forecasting. This study focuses on to develop an Artificial Neural Network model for predicting delay and cost overruns in bridge construction project. The model was developed by using 123 historical dataset of completed bridges from the Western Maharashtra (India) region, from this 70% dataset utilize for training and 30% dataset kept for testing purpose. The Cascade Forward Backpropagation model shows higher accuracy for predicting delays and cost overrun with an R value of 0.96673, MSE of 0.0012624 and RMSE of 0.035530, as compared to Feedforward Backpropagation model. This research utilizes dual model comparative approach implemented exclusively in bridge infrastructure project, along with real-world dataset from Maharashtra Public Work Department, which is rarely used in previous research work. It offers practitioners for taking decision and detecting the risk in the early stage, as well as to optimize the financial losses.