Abstract
Material price fluctuations pose significant challenges for executing construction projects and adhering to budgetary estimates. Existing research studies focused on forecasting construction material prices rather than quantifying and gauging overall project cost risks related to price volatilities, and they relied on traditional time-series forecasting methods that are incapable of offering full probabilistic distributions of price fluctuations and of providing a comprehensive assessment of risk uncertainties associated with material price fluctuations. This paper addresses these gaps by developing an integrated framework to quantify and gauge project risks due to construction material price volatilities. The framework's validity and practicality were demonstrated using real-world projects with various characteristics and different market conditions, including an 11-month bridge replacement project and a 25-month major roadway project. Historical Producer Price Index (PPI) values were collected for four construction materials: steel reinforcement, asphalt, aggregate, and concrete. Three probabilistic deep-learning models - deep autoregressive models, probabilistic feed-forward neural networks, and transformers - were developed to forecast PPI probabilistic distributions. The performance of the developed models was evaluated using probabilistic metrics, and the top-performing models were identified for each material and were compared with a baseline artificial neural network model and a Bayesian prophet model. Finally, stochastic risk models were developed to integrate the predicted distributions into the price escalation clauses of standard construction contracts (i.e., FIDIC) to model risk uncertainties and plot stochastic risk profiles. The findings provided valuable insights about patterns and fluctuations in prices across various construction materials, market volatilities, extreme events, and different types of clauses, including "any-increase escalation clauses "and "threshold escalation clauses." This study contributes to the growing body of knowledge on construction material price escalation by offering a comprehensive approach that provides project parties with data-driven insights that inform strategies to mitigate financial setbacks resulting from price fluctuations and volatilities in their projects.
Recommended Citation
Y. Jezzini et al., "Modeling Framework to Quantify and Gauge Project Cost Risks due to Construction Material Price Volatilities using Predictive Probabilistic Deep-Learning Algorithms and Stochastic Risk Modeling," Journal of Construction Engineering and Management, vol. 151, no. 7, article no. 04025071, American Society of Civil Engineers, Jul 2025.
The definitive version is available at https://doi.org/10.1061/JCEMD4.COENG-16055
Department(s)
Civil, Architectural and Environmental Engineering
Keywords and Phrases
Construction materials; Deep learning; Price escalation clauses; Price fluctuations; Probabilistic forecasting; Stochastic risk modeling
International Standard Serial Number (ISSN)
1943-7862; 0733-9364
Document Type
Article - Journal
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2025 American Society of Civil Engineers, All rights reserved.
Publication Date
01 Jul 2025
