Аlgorithmization of managing the distribution of limited financial resources in the regional social fund based on ART-MAP neural networks

In a context of persistently limited budgetary resources, exacerbated by the growing social burden on regional budgets, the problem of finding effective mechanisms for distributing state social funds is of paramount importance. The social well-being of millions of citizens and the stability of social relations directly depend on how rationally and fairly resources are distributed. A key element in building such an effective system is a clear, scientifically sound, and, crucially, prioritized classification of recipient groups of social assistance. This classification allows for a shift from a egalitarian support approach to a targeted approach, focusing efforts and resources on the most vulnerable groups. This article proposes an innovative approach to algorithmizing this complex process. The proposed method is based on integrating the developed hierarchical classification of recipients with modern neural network technologies, specifically the ART-MAP family of architectures. The use of neural network data allows for the creation of a flexible, adaptive system capable of learning in real time, taking into account the dynamics of changes in the social environment, and ensuring not only accurate but also completely transparent, understandable, and justified dispersion (redistribution) of financial flows, which is critical for upholding the principles of social justice.

1. Shabunova A.A., Kroshilin S.V., Yarasheva A.V., Medvedeva E.I. Socio-Economic Indicators of Russia's National Development Goals: Trends and Forecast. Economic and Social Changes: Facts, Trends, Forecast. 2024;17(5):40–54. (In Russ.). https://doi.org/10.15838/esc.2024.5.95.2

2. Khalturin R., Sudorgin R., Akinshin N. Theoretical justification of the model of search for optimal solutions in complex resource management systems. Transportation and Information Technologies in Russia. 2025;15(1):214–233. (In Russ.). https://doi.org/10.12731/2227-930X-2025-15-1-356

3. Terentyev A.V., Yevtukov S.S., Karelina E.A. Development of zoning method for solving economic problems of optimal resource allocation to objects of various importance in context of incomplete information. In: Proceedings of the International Scientific Conference "Far East Con" (ISCFEC 2020), 01–04 October 2019, Vladivostok, Russia. Atlantis Press; 2020. P. 765–772. https://doi.org/10.2991/aebmr.k.200312.108

4. Averyanov V.V. Adaptive management of technical systems using neural network technologies. Young scientist. 2026;(11):65–67. (In Russ.).

5. Petukhova A.V., Kovalenko A.V. Decision Support Systems (DSS) based on intelligent technologies. Architecture, design and usage of DSS in various sectors. Applied Mathematics and Control Sciences. 2025;(1):47–58. (In Russ.).

6. Carpenter G.A., Grossberg S. Adaptive Resonance Theory. In: Encyclopedia of Machine Learning. New York: Springer; 2010. P. 22–35. https://doi.org/10.1007/978-0-387-30164-8_11

7. Lvovich Ya.E., Lvovich I.Ya., Choporov O.N., et al. Optimization of digital management in organizational systems. Voronezh: Nauchnaya kniga; 2021. 191 p. (In Russ.).

8. Lvovich I.Ya. Decision-making based on optimization models and expert information. Voronezh: Nauchnaya kniga; 2023. 231 p. (In Russ.).

9. Granger E., Prieur D., Connolly J.-F. Evolving ARTMAP Neural Networks Using Multi-Objective Particle Swarm Optimization. In: IEEE Congress on Evolutionary Computation, 18–23 July 2010, Barcelona, Spain. IEEE; 2010. P. 1–8. https://doi.org/10.1109/CEC.2010.5585953

10. Zhilov R.A. Application of neural networks for data clustering. News of the Kabardino-Balkarian Scientific Center of RAS. 2021;(1):15–19. (In Russ.). https://doi.org/10.35330/1991-6639-2021-1-99-15-19