SearchMAS: AI-Powered Knowledge Graph Exploration for MBSE

Overview

PoC #04 introduces SearchMAS, an advanced AI-based chat and visualization environment developed by the MSE of RWTH Aachen University. The tool is designed to manage the immense complexity of Model-Based Systems Engineering (MBSE) by mapping system data into an interactive Knowledge Graph.

How it Works: From Query to Graph

The core innovation of SearchMAS lies in its ability to translate natural language questions into precise database queries. Instead of searching through static documents, engineers can interact with their system data:

• Natural Language Interface: Users can ask questions like "Who is responsible for the customer function rear access?" Knowledge Graph Retrieval: The AI queries a graph database and identifies relevant nodes such as FLCustomerFunction, FLSubFunction, and SolutionContext.
• Traceability & Responsibility: The tool instantly reveals metadata, including responsible organizations, specific contact persons (e.g., for "Rear access functions"), and organizational units.

Features & Visualization

SearchMAS provides multiple perspectives on the system architecture:

1. Interactive Node View: Visualizes how customer functions are linked to sub-functions and logical components.
2. Tabular & JSON Export: Engineers can view raw edge-and-node data or export the results as JSON for further processing in other development tools.
3. End-to-End Tracing: The tool allows users to follow solution and use-case chains across different layers of the system model, ensuring complete traceability.

Impact

SearchMAS lowers the entry barrier for stakeholders by providing a "Google-like" experience for complex system models. It ensures that critical information—such as ownership, dependencies, and architectural impacts—is always just one chat message away, significantly accelerating decision-making in the automotive development process.