科研管理

科研管理 ›› 2025, Vol. 46 ›› Issue (12): 65-77.DOI: 10.19571/j.cnki.1000-2995.2025.12.007

• 论文 • 上一篇    下一篇

芯片制造领域“知识-技术-产品”多层创新网络连通性研究

何喜军，安嘉骏，董颖，石安杰，张志豪   

  1. 北京工业大学经济与管理学院，北京100124
  • 收稿日期:2024-07-05 修回日期:2025-05-07 接受日期:2025-05-07 出版日期:2025-12-20 发布日期:2025-12-15
  • 通讯作者: 张志豪
  • 基金资助:
    国家自然科学基金项目：“异构信息网络下技术供需匹配模型与对接路径研究”（71974009，2020.01—2023.12）；北京市自然科学基金项目：“大模型赋能的北京技术交易市场供需数据质量评估方法与应用研究”（9252002,2025.01—2027.12)；国家自然科学基金青年项目：“大模型赋能的高价值专利个性化交易推荐方法与应用研究”（72404020，2025.01—2027.12）；国家资助博士后研究人员计划（C档）：“大模型驱动下基于专利供需知识图谱的交易推荐研究”（GZC20230184，2025.01—2027.12）。

Research on the connectivity of the multi-layer innovation network of  "knowledge-technology-product" in the field of chip manufacturing

He Xijun, An Jiajun, Dong Ying, Shi Anjie, Zhang Zhihao   

  1. School of Economics and Management, Beijing University of Technology, Beijing 100124, China

  • Received:2024-07-05 Revised:2025-05-07 Accepted:2025-05-07 Online:2025-12-20 Published:2025-12-15

摘要：     芯片制造是实现科技自立自强与高质量发展的战略支柱产业，其创新网络的连通性水平直接影响我国在全球创新链、产业链、供应链中的竞争优势与地位。现有研究鲜有从多层网络渗流视角探索该领域知识、技术与产品等创新要素之间的交互连通性。本文以2013—2022年芯片制造领域的论文、专利及产品供应合作数据为基础，构建中国及全球“知识-技术-产品”多层创新网络，基于渗流理论的生成函数方法，刻画多层网络动态渗流过程，测度网络交互连通性并揭示其演变规律。研究发现：（1）中国芯片制造领域单层创新网络连通性差异显著，知识网络保持稳定连通，技术网络自2018年后处于断裂状态，产品网络连通性呈倒U型波动。（2）“知识-技术-产品”全过程创新链的交互连通性水平较低，知识向技术、产品的扩散路径通而不畅，技术向产品的扩散路径断裂。（3）多层创新网络极大子团的主体分布不均衡，异构创新资源聚集度较低，整网韧性不足。本研究为多层创新网络渗流连通性研究提供了方法参考，并为中国芯片制造领域创新网络治理提供了策略支撑。

关键词: 芯片制造, 多层创新网络, 渗流理论, 交互连通性

Abstract:    The chip manufacturing industry is a strategic pillar for China′s technological self-reliance and high-quality development. The connectivity of its innovation networks directly affects China′s position and competitive advantage in the global innovation, industrial, and supply chains. However, existing studies have seldom explored the interactive connectivity of knowledge, technology, and product elements from a percolation-based multi-layer network perspective. This paper constructed a multi-layer innovation network for China and the world in the "knowledge-technology-product" dimensions, using the data from research papers, patents, and product supply collaborations in the chip manufacturing sector from 2013 to 2022. By applying the generating function method based on the percolation theory, the study characterized the dynamic percolation process, measures the inter-layer and intra-layer connectivity, and revealed its structural evolution and connectivity thresholds. The findings are as follows: (1) There are significant differences in the connectivity of single-layer innovation networks in China′s chip manufacturing sector. The knowledge network remains stably connected, while the technology network has exhibited disconnection since 2018, and the product network shows an inverted U-shaped trend in connectivity. (2) The interactive connectivity of the "knowledge-technology-product" innovation chain is relatively low. The diffusion paths from knowledge to technology and from knowledge to products are technically connected but inefficient, and the path from technology to products is disrupted. (3) The distribution of the largest subgroups in the multi-layer innovation networks is uneven, with low aggregation of heterogeneous innovation resources, resulting in insufficient resilience across the overall network. This article provides a methodological reference for investigating percolation connectivity in multilayer innovation networks and offers strategic guidance for innovation network governance in China′s chip manufacturing sector.

Key words: chip manufacturing, multilayer innovation network, percolation theory, interaction connectivity