Model of forecasting of material resources and estimated cost at early stages of life cycle of construction objects

Introduction. Digitalization of construction involves the possibility of forecasting material resources with a given degree of accuracy at the early stages of the life cycle of the construction object, which will make it possible to form cost indicators, as well as the volume of material resources and equipment — one of the key elements of management and planning of various stages of the life cycle of the object. The possibility of creating a forecasting tool based on artificial intelligence (AI) and machine learning (ML) technologies for these parameters seems to be a promising development direction, allowing to achieve a high level of accuracy in budget planning and project duration at the pre-project stage of an investment and construction project.

Materials and methods. Design documentation for 37 multi-apartment residential buildings with allocated and normalized parameters: technical and economic indicators, material and technical resources, cost and others. To prepare and train AI models based on Python, the methods of the scikit-learn library were selected to compare the following mathematical models: decision trees, regressions and algorithms based on boosting.

Results. The training and research were conducted using the automated machine learning (AutoML) method. Based on a comparison of the coefficient of determination R2 and the standard deviation (RMSE), ensembles of models were selected that form a forecast for the volume of material resources and equipment, as well as for the estimated cost with an error range of ±8 %. The input values of the models were 11 quantitative and qualitative parameters describing the characteristics of the planned object, the formation of which is possible at the early stages of the life cycle of the object without the development of design documentation.

Conclusions. The results of the study demonstrate the possibility of obtaining actual design data already at the pre-design stage with accuracy corresponding to the stage of development of working documentation for the construction object. Thus, the accuracy of forecasts of the total estimated cost is significantly increased, and it also becomes possible to predict with a given accuracy the volumes of materials and equipment at the early stages of the life cycle of the construction object in order to optimize the entire construction process.

1. Sekushin S.V., Kurbatskaya E.P., Petrovsky A.I. Current issues of price formation and cost estimating in Russian construction industry. Bulletin of the Scientific Council of JSC “IERT”. 2022; 7:40-45. EDN CFMOZV. (rus.).

2. Sorokin I., Nastychuk A. Analysis of the experience of applying machine learning methods in the russian construction industry. Construction and Architecture. 2023; 11(1):18. DOI: 10.29039/2308-0191-2022-11-1-18-18. EDN WEOBBP. (rus.).

3. Petrovskiy A.I., Kapustina N.V. Relationship between construction cost estimation accuracy and cost effectiveness of investment and construction projects. State and Municipal Management. Scholar Notes. 2021; 1:165-169. DOI: 10.22394/2079-1690-2021-1-1-165-169. EDN FVZTTT. (rus.).

4. Petrochenko M.V., Nedviga P.N., Kukina A.A., Sherstyuk V.V. Classification of information models in BIM using artificial intelligence algorithms. Vestnik MGSU [Monthly Journal on Construction and Architecture]. 2022; 17(11):1537-1550. DOI: 10.22227/1997-0935.2022.11.1537-1550. EDN JFYSSO. (rus.).

5. Kirianova L.V., Suleimanov N.F. Comparison of machine learning regression models of housing costs in Hungary. Journal of Economy and Entrepreneurship. 2023; 5(154):636-641. DOI: 10.34925/EIP.2023.154.5.123. EDN HSQACR. (rus.).

6. Mikhailova E.V., Savina V.V., Savin I.M. Artificial intelligence as a tool for optimizing resource schedules. Construction Production. 2022; 2:52-56. DOI: 10.54950/26585340_2022_2_52. EDN AODBGS. (rus.).

7. Kolchin V.N. The specifics of the use of “artificial intelligence” technology in construction. Innovations and Investments. 2022; 3:250-253. EDN JJLECU. (rus.).

8. Blanco J.L., Fuchs S., Parsons M., Ribeirinho M.J. Artificial intelligence: Construction technology’s next frontier. Capital projects & infrastructure. April 2018.

9. Makarov A.N., Gureev M.V. Determining the parameters of a model for forecasting material resources for the construction of residential buildings at the investment feasibility assessment stage. Construction Production. 2023; 4:97-104. DOI: 10.54950/26585340_2023_4_97. EDN XELYIL. (rus.).

10. Nizamov I.V., Elovenko D.A. Overview of AutoML technology, available tools, application example and comparative analysis with the classic solution. Young Researchers’ Journal of ISTU. 2022; 12(3):468-475. EDN TMBATS. (rus.).

11. Dobashin A.S., Stepanov G.S., Gogolin I.A., Gadzhiev D.M. Using mathematical statistics and programming to solve practical machine learning problems. Actual Studies of Higher Education 2023: collection of articles of the International Scientific and Practical Conference. 2023; 229-242. DOI: 10.46916/03052023-2-978-5-00174-965-3. EDN NKBEGI. (rus.).

12. Saranin Z.A. Review of the practice of applying machine learning methods in energy economics. Youth Science Week of the Institute of Industrial Management, Economics and Trade : proceedings of the All-Russian Student Scientific and educational conference. 2023; 256-258. EDN OHSVAR. (rus.).

13. Entezari A., Aslani A., Zahedi R., Noorollahi Y. Artificial intelligence and machine learning in energy systems : a bibliographic perspective. Energy Strategy Reviews. 2023; 45:101017. DOI: 10.1016/j.esr.2022.101017

14. Ahmad T., Zhu H., Zhang D., Tariq R., Bassam A., Ullah F. et al. Energetics Systems and artificial intelligence: applications of industry 4.0. Energy Reports. 2022; 8:334-361. DOI: 10.1016/j.egyr.2021.11.256

15. Sukhov A.A., Sukhankin A.A. Comparative analysis of the best gradient boosting algorithms: CATBOOST, XGBOOST, LIGHT GBM. Recommendations for their application. Challenges of modernity and strategies for the development of society in a new reality : collection of materials of the XXIII International Scientific and Practical Conference. 2024; 287-293. EDN AKOAVR. (rus.).

16. Machado M.R., Karray S., de Sousa I.T. LightGBM: an effective decision tree gradient boosting method to predict customer loyalty in the finance industry. 2019 14th International Conference on Computer Science & Education (ICCSE). 2019; 1111-1116. DOI: 10.1109/ICCSE.2019.8845529

17. Chen T., Guestrin C. XGBoost. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2016; 785-794. DOI: 10.1145/2939672.2939785

18. Liu X., Wang T. Application of XGBOOST model on potential 5G mobile users forecast. Lecture Notes in Electrical Engineering. 2023; 1492-1500. DOI: 10.1007/978-981-19-3387-5_177

19. Ostrikova A.L., Selyutin V.V. Innovative technologies for mass valuation of residential real estate. Ecology. Economy. Informatics. System Analysis and Mathematical Modeling of Ecological and Economic Systems. 2023; 1(8):147-154. DOI: 10.23885/2500-395X-2023-1-8-147-154. EDN PYQMGE. (rus.).

20. Faisal A., Yigitcanlar T., Kamruzzaman M., Paz A. Mapping two decades of autonomous vehicle research : a systematic scientometric analysis. Journal of Urban Technology. 2020; 28(3-4):45-74. DOI: 10.1080/10630732.2020.1780868

21. Sari R., Voyvoda E., Lacey-Barnacle M., Karababa E., Topal C., Islambay D. Energy Justice: A Social Sciences and Humanities Cross-Cutting Theme Report. Cambridge, 2017. DOI: 10.13140/RG.2.2.33861.35043