Milan, Italy 28-10-2026
Portfolio management of construction projects is an intricate, data-intensive activity in which an organization selects, ranks, and manages multiple synchronous projects over time. Classic approaches are generally inadequate for modeling the dynamic nature of construction projects, including variable costs, resource constraints, and external risks. This paper investigates the use of deep learning models to support improved predictive analytics and decision-making in construction projects. We also introduce a framework based the deep learning methods, such as multilayer perceptron (MLP) networks, recurrent neural networks (RNN) networks, and long short-term memory (LSTM) networks, to make a project success bid, optimize resource allocation, forecast project duration, and estimate risk factors in portfolios.
The research uses historical data from a wide range of construction projects, including relevant attributes (e.g., project cost, duration, risk, resource allocation, and critical milestones). Results show that deep learning models outperform conventional approaches in addressing cost overruns, delays, and resource inefficiencies, and provide better accuracy and greater explanatory power for portfolio dynamics. On the other hand, the proposed models offer real-time decision support, enabling portfolio managers to proactively defend against risk, manage resource allocation and make more informed, evidence based decisions. This work demonstrates the ability of artificial intelligence to change construction project portfolio management and, therefore, pave the way toward cheaper, smarter, and better project results.
Keywords - Deep Learning, Predictive Analytics, Project Portfolio, Decision Support, Resource Allocation
This research investigates the comparative effectiveness of AutoCAD, Revit, and Artificial Intelligence (Kreo Software) in quantity surveying within construction projects, focusing on their accuracy, efficiency, and user-friendliness. Quantity surveying, which is essential for managing project costs from estimation to completion, has traditionally relied on AutoCAD. Still, Revit, considered a more advanced tool, and AI is suspected to revolutionize the field. This study employs a mixed-methods approach, integrating qualitative data from a case study with quantitative results from using different software for compression. The findings indicate that, while AutoCAD remains reliable for basic tasks, Revit’s BIM integration enhances the accuracy and efficiency of project management. AI, although still in its developmental stage, shows promise in automating calculations and predictive analytics, offering reduced human error and increased productivity. However, its accuracy remains a limitation for its full-scale adoption. The research concluded that the selection of tools is highly dependent on the project's complexity, budget, and specific surveying needs. It offers recommendations for practitioners regarding the most suitable technologies and suggests areas for future research, particularly the further integration of AI in quantity surveying, highlighting the evolving nature of the profession and the importance of embracing technological advancements to optimize construction outcomes.
Keywords - Auto CAD, Revit, BIM, construction, Artificial Intelligence, quantity survey