Performance Analysis of an Integrated Digital Twin and Artificial Intelligence Framework for Time and Cost Optimization in Construction Projects Using Real-Time Data

Abstract

Recent advances in Digital Twin and Artificial Intelligence technologies have fundamentally transformed construction project management. This study aims to evaluate the performance of an integrated framework combining Digital Twin and Machine Learning algorithms for optimizing time and cost in construction projects. Real-world data were collected from three large-scale infrastructure projects in Iran, complemented by datasets from international repositories such as the Construction Industry Institute and the European BIM Database. The datasets include real-time information from IoT sensors, BIM models, and project scheduling and cost reports. The proposed framework integrates Digital Twin modeling with Reinforcement Learning and Recurrent Neural Network (RNN)-based prediction to enable real-time simulation, delay forecasting, and corrective decision recommendations. The analysis reveals that adopting this integrated approach results in an average reduction of 18% in project duration and approximately 12% in total cost compared with traditional management practices. Moreover, the model demonstrates strong adaptability to unexpected changes, such as material price fluctuations and delays in equipment delivery. This study contributes to the advancement of intelligent construction management tools by integrating Digital Twin, Machine Learning, and real-time data analytics. The findings can guide project managers and BIM–IoT developers in optimizing performance across large-scale civil engineering projects.