THEORETICAL AND METHODOLOGICAL PRINCIPLES OF APPLICATION OF AGENT-ORIENTED APPROACH TO MODELING PROCESSES OF LOCAL HROMADAS

This article presents a theoretical and methodological framework for applying a multi-agent systems (MAS) approach to modeling the budgeting processes of local hromadas. MAS-based modeling enables consideration of agents’ heterogeneity, bounded rationality, and dynamic interaction structures, which traditional models typically overlook. These features allow for a more realistic reproduction of economic processes at the hromada level. The study explores the potential of MAS in solving practical problems such as resource allocation optimization, strategic planning for sustainable development, and strengthening resilience to crises. Particular attention is paid to the development of a budget auction model based on game theory, where economic actors - including local government, enterprises, and residents - engage in a structured competition for resource distribution. The auction mechanism identifies production functions and utility values for each actor, enabling iterative optimization based on predefined efficiency criteria. A proposed simulation framework combines the auction logic with reinforcement learning methods (MARL), where agents improve their strategies through interaction and feedback. The article also outlines a multi-step algorithm for modeling decision-making and coordination processes within the decentralized budget structure. Ultimately, the approach offers tools for more transparent, participatory, and efficiency-driven public finance management. Its implementation can support the digital transformation of local governance, aligning resource distribution with sustainable development goals. The proposed methodology is particularly relevant in the context of Ukraine’s decentralization reform and post-war recovery challenges, and it may serve as a foundation for the development of intelligent public finance systems in local hromadas.