A Knowledge Graph-Driven Approach for Architecture Design Space Generation and Reduction

The process of architecture design for complex systems relies on domain knowledge, which involves identifying architecture decision issues from text-based requirements and generating the architecture design space. As the complexity of the system increases, higher requirement is proposed for the storage and reuse of design knowledge. To address this challenge, this paper proposes a knowledge graph-driven approach for architecture design space generating and reduction. To represent and store knowledge in architecture decisions, a decision-oriented design knowledge graph (DDKG) is constructed by analyzing the relationship among requirements, functions, and knowledge in architecture decisions. The nodes in DDKG to generate the architecture design space are matched with decision issue keywords extracted from the user’s requirement text. Moreover, the design space is generated from nodes matched in DDKG and reduced by forming a Constraint Satisfaction Problem (CSP) model. The method proposed in this paper has been validated in the generation and reduction of architectural design space for the first-stage separation systems of launch vehicles.