Benchmarking Requirement-to-Architecture Generation with Hybrid Evaluation
This addresses a crucial step in software development for automating architecture generation, though it is incremental as it focuses on benchmarking and evaluation.
The authors tackled the lack of datasets for generating software architecture from requirements by introducing R2ABench, a benchmark with real-world projects and expert diagrams, and found that LLMs achieve strong syntactic validity but struggle with relational reasoning, leading to fragmented architectures.
Recently, Large Language Models (LLMs) have demonstrated significant potential in automating software engineering tasks. Generating software architecture designs from requirement documents is a crucial step in software development. However, there is currently a lack of functional datasets tailored for this task. To bridge this gap, we introduce R2ABench (Requirement-To-Architecture Benchmark), a novel benchmark comprising diverse real-world software projects paired with comprehensive Product Requirements Documents (PRDs) and expert-curated PlantUML reference diagrams. Furthermore, we propose a multi-dimensional, hybrid evaluation framework that assesses generated diagrams across three complementary layers: Structural Graph Metrics, Multi-dimensional Scoring, and Architecture Anti-pattern Detection. Using this framework, we conducted a comprehensive empirical study evaluating state-of-the-art models and agentic workflows. Our study shows that LLMs show strong syntactic validity and robust entity extraction but fundamentally struggle with relational reasoning, leading to structurally fragmented architectures. Code-specialized models partially alleviate this limitation, while agent frameworks introduce significant instability rather than consistent improvements. R2ABench provides a robust and standardized foundation for advancing LLM-driven software architecture generation.