Framework pro analýzu preferencí využívající metodu UTA

This bachelor’s thesis deals with the design and implementation of a framework for preference analysis based on the UTA method. The work was motivated by the existence of a historical web application that enabled the practical use of the method, but whose architecture was tightly coupled to a combination of several technologies and was difficult to maintain and extend. The aim of the thesis was therefore to analyse the original solution and, on this basis, design a more modern, modular and reusable software solution for the Python ecosystem. The thesis first places the topic in the context of multi-criteria decision analysis and preference disaggregation, explains the principles of the UTA method and then analyses the architecture of the original application. Based on this analysis, a new solution is designed and implemented, consisting of the uta_solver library core, an explicit data model for criteria and breakpoints, and a separate frontend layer built with Streamlit. The new solution supports the UTASTAR and UTANM variants,different criterion types, JSON and CSV import and export, and is complemented by example scenarios and tests. The main contribution of the thesis is the creation of an architecturally cleaner solution that separates computational logic from the presentation layer, improves testability and extensibility, and provides a foundation for further development of the UTA family of methods in Python.