Optimization of a stock portfolio using multi-criteria decision-making methods

The thesis deals with portfolio optimization using a two-stage method. The thesis aims to construct a stock portfolio using multi-criteria decision-making methods, focusing particularly on the pre-selection of efficient investment instruments. The theoretical part of the thesis clarifies the basic principles of the financial market and introduces the methods of multi-criteria decision-making. Furthermore, various approaches used for portfolio optimization using multi-criteria decision-making methods are presented. This part provides insight into the theoretical foundations and methodology that are the basis for practical application. The practical part of the thesis applies the two-stage method to a dataset containing 1375 titles of the U.S. stock market titles. In the first stage of portfolio optimization, multi-attribute decision-making methods are used to select efficient investment instruments for three theoretical investors with different attitudes towards risk, specifically using the methods TOPSIS, TOPSIS-sort-B, and ELECTRE I. The second stage presents the final portfolios for the aforementioned theoretical investors based on the application of the lexicographic goal programming model. The constructed portfolios are compared with the benchmark, which is the market index S&P500. The results of the thesis favour the TOPSIS-sort-B method, which generally provides the best results among all constructed portfolios when comparing their performance in the year 2023. Thus, the thesis presents a possible guide to the use of the two-stage method in the field of investment decision-making and shows the importance of the appropriate selection of efficient instruments in portfolio optimization.