HYBRID MODELS OF OPTIMIZATION OF AN ENTERPRISE'S INVESTMENT PORTFOLIO IN CONDITIONS OF ECONOMIC UNCERTAINTY AND ENTROPY

Abstract

The digital transformation of financial management changes the conditions for constructing an enterprise investment portfolio. Portfolio decisions increasingly rely on open financial data, automated updates of exchange rates, interest rates and instrument yields, forecasting models, and scenario-based risk assessment. For an enterprise, portfolio choice differs from the logic of a private investor or an institutional fund: it must account not only for expected return and risk but also for liquidity, payment schedules, the currency structure of liabilities, instrument maturity, operational solvency, and regulatory constraints. In a financial environment marked by regime shifts, a static portfolio model loses practical reliability unless it is supplemented with forecasting, scenario, and control modules. Purpose. The purpose of the article is to substantiate the methodological foundations of optimizing an enterprise investment portfolio based on hybrid approaches and to develop an applied architecture that combines forecasting models, portfolio optimization, scenario-based risk assessment, enterprise-specific constraints, and model risk control. Materials and methods. The methodology includes systematization of sources, comparative analysis of portfolio optimization models, structural and functional modelling, scenario calculation, and separation of the forecasting signal from the financial decision. Results. The article systematizes the evolution of portfolio optimization from the mean-variance approach to Black–Litterman, CVaR, robust methods, and models enhanced with artificial intelligence tools. It is substantiated that in corporate financial management hybrid optimization should include not only the algorithmic combination of models but also the alignment of forecasts with liquidity, currency position, payment maturity, and internal risk limits. Open data from the National Bank of Ukraine are used to demonstrate the regime heterogeneity of the Ukrainian financial environment in 2020–2026, including changes in exchange rate and interest rate regimes. The article proposes an architecture for optimizing an enterprise portfolio based on hybrid approaches, combining data, forecasting, risk assessment, optimization, enterprise constraints, and model control blocks. A scenario calculation for conservative, balanced, and adaptive portfolios shows that an increase in expected return is accompanied by a disproportionate rise in risk; therefore, an enterprise portfolio should be assessed through operational suitability rather than formal return alone. Discussion. Further research should focus on empirical testing of the proposed architecture using daily financial data and on comparing mean-variance, Black–Litterman, CVaR, robust optimization, and portfolio models enhanced with AI-based forecasting modules. Particular attention should be paid to transaction costs, domestic government bond liquidity, tax implications, the currency structure of enterprise liabilities, maximum portfolio drawdown, and the stability of portfolio weights under changing exchange rate and interest rate regimes.