Knowledge graph-assisted design-by-analogy: promoting product innovation through structured analogical knowledge retrieval

Abstract Design-by-analogy (DbA) is a powerful method for product innovation design, leveraging multidomain design knowledge to generate new ideas. Previous studies have relied heavily on designers’ experiences to retrieve analogical knowledge from other domains, lacking a structured method to organize and understand multidomain analogical knowledge. This presents a significant challenge in recommending high-quality analogical sources, which needs to be addressed. To tackle these issues, a knowledge graph-assisted DbA approach via structured analogical knowledge retrieval is proposed. First, an improved function-effect-structure ontology model is constructed to extract functions and effects as potential analogical sources, and six semantic matching rules are established to output entity triplets, and the DbA knowledge graph (DbAKG) is developed. Second, based on the knowledge of semantic relationships in DbAKG, the domain distance and similarity between the design target and the analogical sources are introduced to establish an analogical value model, ensuring the novelty and feasibility of analogical sources. After that, with function as the design target, analogical sources transfer strategy is formed to support innovative solution solving, and TRIZ theory is used to solve design conflicts. Finally, a pipeline inspection robot case study is further employed to verify the proposed approach. Additionally, a knowledge graph-assisted analogical design system has been developed to assist in managing multidomain knowledge and the analogical process, facilitate the adoption of innovative design strategies, and assist companies in providing more competitive products to seize the market.