【摘要】 高校科研空间布局影响科研人员在日常的步行路径上偶然接触其他学术信息的潜力，进而影响科研成果产出与创新，但这一关系还未得到充分量化验证。以麻省理工学院（MIT）为例，通过实地调研数据构建一种普遍的科研空间步行ABM仿真模型，获取空间变化前后三年科研人员日常路径上接触的学术信息量作为空间作用指标，获取Web of Science数据库中相同时间段的论文发表数据作为成果指标，通过回归分析探讨两种指标的关联影响。结果表明，MIT科研空间布局影响下的学术信息量与其论文产出具有显著的正向相关性，空间布局有效促进了科研创新。本文的研究思路与结论为研究型高校空间环境与创新的相关研究与实践提供了有益的参考。更多还原

【Abstract】 In the era of the knowledge economy, universities are undergoing functional renovations to transform them from venues of knowledge transmission into bases of research innovation. Consequently, how architectural spatial design improves interdisciplinary exchange and academic innovation has emerged as a core research imperative within architectural studies of universities. The existing literature indicates that research-space layout in universities influences researchers’ potential for serendipitous exposure to relevant academic information during routine pedestrian circulation, thereby affecting research output and innovation. Yet this mechanism remains insufficiently quantified, with prior quantitative studies on the correlation between space layout and innovations mainly focusing on specific cases. These studies primarily analyze how static spatial attributes affect interpersonal interactions through space syntax, thereby influencing work performance and innovation in a company or a department. There is a dearth of studies of the overall planar layout of universities and how researchspace layout changes are related to academic innovation. A case study based on the research area consisting of 49 academic buildings at the Massachusetts Institute of Technology(MIT) was carried out to explore the correlation among research-space layout, associated academic-information contact, and research outputs. Since the concept of academic innovation is difficult to quantify accurately, the research-paper output of MIT scholars was used to represent academic innovation. Different from interpersonal communication in offices, collaboration among researchers who are unfamiliar with each other requires specific catalysts. Serendipitous exposure to academic information during daily work—especially research from other disciplines—helps researchers understand the progress of peers, develop new research ideas, and motivate collaborative exchanges and studies that ultimately generate novel scholarly outputs.In this study, 3,792 articles of 1,365 authors included in the Web of Science database in 3 successive years before and after the expansion of the primary research area in MIT were collected as output indicators. Meanwhile, an agentbased modeling(ABM) experimental foundation framework was established through field-survey data, and evaluation indicators of building-level academicinformation density were set. Serendipitous exposure to information along routine paths of researchers on campus across two periods was simulated and applied as the spatial-intervention indicator. Changes in building-level information exposure and publication volume before and after the space renewal were analyzed. This method allowed us to eliminate or weaken impacts of policy systems, culture, and other influencing factors on the research-space layout. The linear regression analysis of academic information exposure(independent variable) against publication volume(dependent variable) reveals significantly positive correlations across the two periods investigated. Moreover, changes in academic-information exposure have a significantly positive effect on changes in publication volume(p=0.007**, regression coefficient=0.729). Academic-information exposure under the influence of the research-space layout at MIT shows a significantly positive correlation with publication volume, with spatial changes substantively helping researchers to access more academic information and improve their own academic innovations. Although this spatial effect accounts for just 30% of the variance in research productivity(R2= 0.300), it confirms that spatial renovations will facilitate academic innovation, even if only unconsciously. A reasonable research-space layout in universities can enhance the potential for researchers to access cross-disciplinary academic information during daily work, thus facilitating their scientific research outputs and innovations. The findings not only provide beneficial references for theory and practice related to spatial environments and innovation in research-based universities, but also suggest ways to promote campus renovation goals emphasizing scientific research outputs.更多还原