Utilizing Artificial Neural Networks for Predictive KPI Analysis in Bridge Projects

Abstract

KPI Analysis is a very important element in the process of construction project management especially with regard to the predictive of duration of bridge projects, main objective of this study is to develop intelligent prediction models, using Artificial Neural Network (ANN), to forecast earned value indicators in the early stages of the lifecycle of bridge projects in Republic of Iraq. Data used to develop neural network model for estimation of earned value indicators were past bridge contract data from Iraq, where, Iraqi Ministry of Construction, Housing, Municipalities, and Public Works launched bridges projects in 2021 and is expected to continue until 2025. variables affecting on earned value indicators of bridge projects were divided into two main categories; First category called Independent Variables, which include three variables, as the following; Schedule Performance Index (SPI), Cost Performance Index (CPI), and Cost-Schedule Index (CSI). While, second category called Dependent Variables, which include four variables as the following; Planned Cost (PC), Planned Duration (PD), Total Area (TA), and, Total Bridge Length (TBL). Three model was built for the prediction of earned value indectors of bridge projects. It was found that ANN models have the ability to predict SPI, CPI and CSI  with excellent degree of accuracy of the coefficient of correlation (R) 0.8853%, 0.8037%, 0.81094%, and average accuracy percentage of 89.91%, 69.829%, 83.794% respectivily. This indicates that the relationship between the independent and independent variables of the developed models is good and the predicted values from a forecast model fit with the real-life data.