An Intelligent and Automatic System Design Approach for Complex Equipment Conceptual Design

ABSTRACT Addressing the challenges of transforming natural language requirements and the lengthy design cycles inherent in complex systems such as aerial bomb guidance control systems, this study proposes an automated system design method based on large language models (LLMs) and knowledge graphs (KGs). A novel instruction framework, REQUIRE, is proposed to guide the transformation of unstructured natural language requirements into a structured function ontology using an LLM. Subsequently, automatic mapping from the function ontology to a component ontology is achieved through KG‐based semantic reasoning. Finally, an optimized module combination generation algorithm is utilized to swiftly generate component‐level design solutions, and the performance of these solutions is evaluated using a comprehensive performance scoring model. Experimental results demonstrate that, compared with traditional methods, this approach can generate system design solutions that meet the specified requirements more expeditiously, thus providing a novel approach for the intelligent design of complex systems.