Mats V. Informational technologies to support financial decisions in conditions of risk and uncertainty

The thesis is devoted to solving a relevant scientific and technical problem related to the development of new and improvement of existing methods of computer implementation of models of controlled Markov processes using as the example the analysis and construction of long-term investment strategies based on the personal context of an individual investor. The object of the study is long-term investment strategies that take into account the personal context of an individual investor. The subject of the study is models, methods and information technologies of building long-term investment strategies based on the personal context of an individual investor. The aim of the study is to develop methods for integrating the personal context of an individual investor into the optimization process by developing new and improving existing methods for long-term portfolio optimization based on controlled Markov processes. The following scientific results were obtained in this study: For the first time: - the specific relationships between the utility function, the type of optimal investment strategy, and the resulting portfolio efficiency are investigated; - it is proven that the inclusion of the risk aversion parameter in the utility function allows modeling a wide range of investor preferences, and a higher level of risk aversion leads to more conservative strategies, especially for low levels of initial capital and among older investors, reflecting a greater emphasis on capital preservation; - it is proved that that total return with utility provides a powerful metric for quantifying the expected performance of a portfolio based on an investor's risk preferences; - a method of constructing surfaces of general expected utility was developed, which show the trade-off between risk and return for different levels of risk aversion; - information technology to support financial decisions was created, which demonstrated significantly higher efficiency in statistical testing on historical data compared to standard portfolio optimization methods It has been proven: - the impact of external financing on the formation of an optimal investment strategy. The presence of regular contributions or income streams allows taking a greater financial risk, especially at the early stages of the investment horizon, and emphasizes the importance of taking into account the financial situation of the sources of financing and the life context of a specific investor for optimizing their investment strategy. - path-dependent utility functions and loss aversion reflect investors' preferences and behavioral biases. - the influence of macroeconomic indicators on the aggregate behavior of assets. The dependence of asset return correlations on indicators such as general inflation, the key rate of the US Federal Reserve, and expected market volatility is tested and proven using historical data. - the impact of including individual utility functions, market information, and the life context of the individual investor in the optimization framework on additional computational complexity and input requirements, which calls for careful consideration and empirical calibration of optimization algorithms. Keywords: computer models, dynamic programming, controlled Markov processes, long-term investments, computing resources, investor life cycle, leverage, optimization of long-term investment strategies, decision-making, optimal allocation, stock markets, risk control, strategic portfolio management.