Design Optimization of Cultural Creative Products Based on Cultural-Aware Generative Graph Transformer

Cultural creative products embody regional heritage, artistic tradition, and cultural storytelling, thus being critical to both economic development and cultural conservation. Traditional design processes, however, depend on human expertise or universal computer-aided design software, leading to low levels of creativity, poor cultural authenticity, and poor market adaptation. Conventional AI-based methods, although providing limited automation, often do not incorporate structured cultural knowledge and do not optimize over several real-world limitations including manufacturability, price, and end-user preferences. This paper presents the Cultural-Aware Generative Graph Transformer (CAGT), a new system that integrates cultural knowledge graph embedding, multimodal feature integration, and generative design with multi-objective optimization. The CAGT model infuses symbolic and semantic cultural knowledge into a graph transformer, combines it with vision-language embeddings, and produces design ideas that meet at the same time aesthetic attractiveness, functional viability, and cultural appropriateness. An additional preference-tuning module ensures alignment of outputs with desired target audience expectations. Comparative tests with conventional and state-of-the-art AI-aided systems reveal that CAGT obtains better cultural relevance, greater design diversity, and higher manufacturability compliance. Both quantitative measures and qualitative assessments affirm its capacity to provide innovative, culturally responsive, and market-ready designs, offering an expandable solution to current cultural creative product development.