Optimization of construction materials logistics and prefabrication to reduce the cycle time of construction and installation works based on digital supply chains and predictive analytics

Abstract

The study is aimed at addressing the problem of reducing the duration of the construction and installation cycle by increasing the predictability and coordination of the supply of materials and prefabricated elements under complex interaction chains between manufacturing plants, transportation, and construction sites. The purpose of the work is to develop and empirically test a comprehensive model for improving material and technical support based on an end-to-end digital supply chain using predictive analytics and machine learning methods. The empirical base consists of data from forty-five large residential and commercial projects implemented in the Central and Ural Federal Districts between 2018 and 2023, divided into a control group with traditional management and a sample of projects with elements of digital logistics. The analysis included data from design and estimate documentation, project schedules, work logs, company resource management systems, telematics complexes, and, for digital projects, data from building information models, cloud platforms, and material identifiers. For the development of predictive models, multiple regression, gradient boosting and random forest algorithms, the Monte Carlo method, and statistical significance tests for differences between samples were applied. It was shown that the transition to digital management leads to a reduction in the average delivery time of key materials by approximately one and a half to two times, while simultaneously significantly reducing delivery time variability and increasing the share of deliveries within the specified time window to values close to full adherence to planned schedules. More than an eighty percent reduction in downtime of tower cranes, concrete pumps, and installation crews was recorded, which is reflected in a decrease in the share of costs for non-productive waiting from about fifteen to around two percent. The integration of digital logistics with prefabrication makes it possible to reduce the full cycle of production, transportation, and assembly of large panel modules by more than sixty percent, mainly due to the direct transfer of data from the information model to production and the transition to just-in-sequence deliveries. The final effect is manifested in the early completion of construction projects and a reduction in the deviation of actual costs from estimated ones due to decreased downtime and material losses. It is concluded that the considered model transforms the material and technical support system into a controlled cyber-physical environment, increases process regularity, and can serve as a foundation for scaling lean construction practices to large projects with a high share of prefabrication.