КЛАСТЕРНИЙ ПІДХІД ДО МОДЕЛЮВАННЯ ЗОВНІШНЬОТОРГОВЕЛЬНОЇ ЛОГІСТИКИ УКРАЇНИ

Objective. is to develop a scientifically grounded model for managing Ukraine’s external logistics flows based on a multivariate analysis of customs data, taking into account changes in the geopolitical environment, the transformation of international trade routes, and the challenges associated with full-scale military operations. Special attention is paid to the assessment of the geographical structure of foreign trade, the balance of physical and monetary indicators of trade operations, and the identification of typical groups of partner countries by logistic characteristics. The research methodology is based on a combination of multivariate statistics, machine learning, and spatial analysis methods. To classify Ukraine’s trade partners, the k-means clustering method was applied, allowing segmentation of 211 countries worldwide by three key indicators: total import weight, total export weight, and geographical distance from each partner country's capital to Kyiv. The empirical data source used was Ukraine’s official customs statistics for the year 2024. The research results reveal significant structural imbalances in foreign trade: despite a positive balance in physical volumes, the monetary value of exports is considerably lower than imports. This indicates a raw material orientation of Ukrainian exports and a technological dependency on imports. The clustering process identified four groups of partner countries: (1) regional partners with active bilateral trade; (2) distant countries with a strong positive balance; (3) distant, import-dependent markets with a negative balance; and (4) strategic countries with large trade volumes and mixed balances. The scientific novelty of the study lies in the application of the clustering approach to customs data analysis for strategic logistics flow management in foreign economic activity. A classification of partner countries by logistic characteristics has been developed, which makes it possible to uncover non-obvious correlations between trade volumes, geographical remoteness, and the efficiency of trade cooperation. The practical significance of the results lies in the potential use of the clustering model to develop differentiated trade strategies, optimize transport logistics, formulate regionally oriented export support policies, and modernize logistics infrastructure. The proposed analytical tool can be integrated into government information systems to support decision-making in the field of foreign economic relations.