Nástroj pro procesní modelování s využitím UPMM

Abstract

This master's thesis explores the automation of process modeling using Large Language Models (LLMs) and metamodels. The objective is to design and implement a tool that transforms informal natural language descriptions of software processes into structured process models. At the core of the solution is the integration of artificial intelligence, enabling automated model generation and subsequent mapping onto a reference Scrum Formal Model.

The thesis emphasizes process validation and the identification of gaps between real-world practice and methodological standards. In addition to static structural modeling, the application provides a dynamic system analysis through Causal Loop Diagrams. These diagrams offer insights into interdependencies and process pathologies within agile environments. The result is a modular platform that addresses the challenges of AI non-determinism in process engineering and establishes a foundation for future automated compliance diagnostics in software teams.