Design Knowledge Graph and User Profiling-Driven Product Innovation Design Problem Solving

Conceptual design is a crucial stage in product innovation design problem solving, which requires design knowledge mainly presented in natural language, images, and semantic expressions of customer demands. These design knowledge exhibit significant features such as cross-domain, fragmented distribution, redundancy, and multimodality, making it difficult for product designers to understand, associate, and accurately reuse design knowledge. As a result, the design problem, design goals, and design iterations are not clearly defined in the conceptual design phase, leading to inaccurate recommendations for classical design solutions and the inability to generate innovative design solutions. To effectively enhance the personal level and capabilities of designers and address the contradiction between “knowledge overload” and “knowledge shortage” in the design problem-solving process, a top-down hierarchical structure model was constructed to model cross-domain product conceptual design knowledge and complete the construction of a cross-domain multimodal graph. By building user profiles and establishing multidimensional feature models, a collaborative filtering recommendation algorithm based on clustering knowledge graphs was developed to achieve personalized design path and principle solution solving. Finally, a case study of home product innovation design problem solving was conducted to verify the effectiveness of applying knowledge graphs in the field of product innovation design problem solving.