MATHEMATICAL MODELING OF CLUSTERIZATION OF THE REGIONAL ECONOMY

Abstract

The study of clustering processes is an urgent task for both scientific research and practical management of the regional economy. Cluster structures play an important role in the development of the modern economy, contributing to the increase in the competitiveness of regions. Despite a significant amount of research in the field of clustering, many aspects remain unresolved. One of the key components of clustering is choosing the right tools for evaluating cluster effectiveness. Mathematical modeling of cluster structures in the regional economy is a key tool for analyzing and forecasting regional economic development. This article examines the specific features of building mathematical models considering the characteristics of regional economic systems. The main factors influencing cluster formation and methods that account for sectoral and geographical specifics are identified. The development of regional economy clustering models is based on the use of modern mathematical methods such as cluster analysis, Bayesian networks, and machine learning algorithms. The primary focus is on adapting existing approaches to the conditions of instability and uneven regional economic development. The application of such models enhances the effectiveness of managerial decisions aimed at supporting clustering and stimulating regional economic growth. An important aspect is the choice of the optimal approach to modeling depending on the specifics of the region. For regions with a high concentration of industrial production, it is advisable to use hierarchical clustering methods, which allow taking into account industry specialization. In regions with a high level of innovation potential, fuzzy clustering methods are effective, which take into account the uncertainty and incompleteness of the initial data. Such approaches allow for more accurate forecasting of cluster development. The article also provides recommendations for improving cluster policy based on the results of modeling.