CASH FLOW OPTIMIZATION ALGORITHM FOR AN INVESTMENT AND CONSTRUCTION PROJECT UNDER UNCERTAINTY BASED ON MACHINE LEARNING

EDN: OOFPKQ

DOI: https://doi.org/10.65891/2949-1983-2026-22-2-22-28

Authors:

ZINEVICH Nikolai Nikolaevich, KLOK Roman Grigorievich

Affiliation:

Department of World Economy and Finance, Educational Institution of Trade Unions of Higher Education "Academy of Labor and Social Relations" (EITU HE "ATiSO")

Abstract.

The article presents an original algorithm for optimizing cash flows of an investment and construction project operating under conditions of external and internal uncertainty. The relevance is due to high volatility of prices for cable and wire products and electrical panel equipment (up to ±60 % per quarter in 2022–2024), disruptions in global supply chains of imported components, tightening monetary policy and the need to reduce cash gaps, which on average in the electrical installation industry reach 10–15 % of the project budget. The scientific novelty lies in the integration of three methods: machine learning (LightGBM gradient boosting and LSTM networks) for forecasting the cost of cable, metal structures and exchange rates; Monte Carlo simulation (10,000 iterations) for assessing probabilistic distributions of cash gaps; a genetic optimization algorithm (population 200, 50 generations) for finding a suboptimal payment schedule. The approbation was carried out on real data of Kompleksnye Sistemnye Proekty LLC (specializing in electrical installation works for industrial and infrastructure facilities) for the power supply project of a technopark (18 months, 1,450 transactions, 53 counterparties, 12 resource types). Results: reduction in the frequency of cash gaps from 33.3 % to 13.9 % of weeks (−19.4 p.p.), decrease in the need for short-term debt financing by 27 %, reduction in currency risk (CVaR) by 35 %, economic effect of RUB 2.12 million over 4 test months. The proposed algorithm can be implemented in corporate financial ERP systems of electrical installation organizations.

Keywords:

cash flows, investment and construction project, electrical installation works, uncertainty, machine learning, optimization, cash gaps, gradient boosting, LSTM, Monte Carlo simulation, genetic algorithm, cable products, currency risk

References:

1. Ilchenko A.N. Cash flow management in construction contractors: specifics of electrical installation // Economics of construction. 2022. No. 3. pp. 45-52.
2. Ministry of Construction of the Russian Federation. Price monitoring for cable and wire products. 2022-2024. URL: https://minstroyrf.gov.ru (date of application: 03/15/2026).
3. Association "Roselektromontazh". Analytical report "Import substitution in the electrical industry." Moscow, 2023. 112 p.
4. The Central Bank of the Russian Federation. Monetary Policy Report. 2023. No. 4. pp. 22-35.
5. National Association of Builders (NOSTROI). Analytical report "Financial risks in construction: the electrical installation sector." Moscow, 2023. 94 p.
6. Smirnov K.O. Cash flow management of a contracting organization: from statics to dynamics // Corporate Finance. 2022. Vol. 16. No. 4. pp. 33-47.
7. Malyshev E.A. Currency risks in the purchase of imported electrical equipment: methods of assessment and hedging // Risk management in construction. 2024. No. 1. pp. 21-29.
8. Nikitina T.S., Frolov D.I. Budgeting of electrical installation projects: modern approaches // Logistics today. 2021. No. 5. pp. 40-48.
9. Ilchenko A.N. Adaptation of the Baumol-Tobin model for contractors in construction // Financial management. 2022. № 6. pp. 55-63.
10. Fedotov D.V. Budgeting investment and construction projects in the context of sanctions risks: a scenario approach for contractors // Finance and credit. 2023. Vol. 29. № 2. Pp. 112-127.
11. Gatti S. Project Finance in Theory and Practice: Designing, Structuring, and Financing Private and Public Projects. 4th ed. Academic Press, 2021. 498 p.
12. Smith J., Petrova M. Cash Flow at Risk (CFaR) in construction subcontracting: A Monte Carlo approach // Journal of Construction Engineering and Management. 2022. Vol. 148. No. 7. P. 04022045.
13. Lee D., Kim S., Park J. Gradient boosting for industrial metal price forecasting (copper, aluminum) and implications for cable procurement // Automation in Construction. 2023. Vol. 146. P. 104689.
14. Zhang H., Kuznetsov S.V., Petrova E.A. Forecasting the cost of cable and wire products using ensemble machine learning methods in the Russian market // Applied Econometrics. 2024. No. 1. pp. 78-95.
15. Kozyrev A.V., Titov V.A. Genetic algorithms in the tasks of calendar and financial planning of industrial enterprises (on the example of electrical installations) // Business Informatics. 2023. No. 2. pp. 55-69.
16. Williams R., Thompson L. A hybrid genetic algorithm-Monte Carlo framework for dynamic cash flow optimization in large-scale industrial projects // European Journal of Operational Research. 2023. Vol. 305. No. 2. P. 612–626.
17. Hochreiter S., Schmidhuber J. Long Short-Term Memory // Neural Computation. 1997. Vol. 9. No. 8. P. 1735–1780. (cited in the 2019 reissue).
18. Chen Y., Wang H. A hybrid ML-Monte Carlo framework for project financing under extreme uncertainty in commodity markets // International Review of Financial Analysis. 2024. Vol. 91. P. 103001.
19. Orlov P.A. Monte Carlo simulation in the evaluation of investment projects in electrical installation: choosing the number of iterations // Investment analysis. 2023. No. 6. pp. 18-27.
20. Mnih V., Badia A.P., Mirza M. et al. Asynchronous methods for deep reinforcement learning for dynamic financial control // Proceedings of the 33rd International Conference on Machine Learning. 2016. P. 1928–1937. (quoted from the 2021 reissue).
21. Semenov V.N. The practice of applying machine learning in financial controlling of electrical organizations, taking into account ESG factors (carbon footprint of cable products) // Managerial accounting. 2024. No. 2. pp. 105-113.