Complex Flowchart Text Generation Based on Large Language Models

As a common business process modeling tool, flowcharts are widely used in software engineering, business process management, architecture framework modeling, and other fields. However, with the increase of business complexity, the scale and complexity of flowcharts also increase, which brings challenges to the understanding and text description of flowcharts. To solve this problem, this paper proposes a text generation method for complex process diagrams based on large language model (LLM), which realizes high-quality text description of complex process diagrams through RPST (Refined Process Structure Tree) decomposition and Least-to-Most prompting strategy.