Knowledge-based Retrieval Methods for Enhancing Aerospace Model Software Documentation

Addressing the critical challenges of poor structure, weak relevance, and limited reusability in aerospace model software development documents, we propose a knowledge-based retrieval method. This method constructs a knowledge base by analyzing key document features, including storage methods, file formats, content structure, inter-document associations, and naming conventions. The system implements structured document management and associates documents based on software configuration information. By partitioning the retrieval domain, constructing semantic vectors, and applying multi-dimensional weight configurations, the method enables efficient retrieval in domains characterized by dense information and frequent term repetition. An aerospace model software document retrieval system has been developed and preliminarily implemented in an aerospace enterprise, demonstrating its effectiveness in improving retrieval efficiency and enhancing the reuse of development knowledge.