- 规划与建设
- 文章编号：1009-6000(2026)02-0040-07
- 中图分类号：TU984 文献标识码：B
- Doi：10.3969/j.issn.1009-6000.2026.02.006
- 项目基金：国家自然科学基金面上项目（52078246）；国家自然科学基金面上项目（52478060）；国家自然科学基金面上项目（52478061）。
- 作者简介：戴思懿，南京大学建筑与城市规划学院硕士研究生；
秦萧，通信作者，南京大学建筑与城市规划学院副教授，博士生导师；
张姗琪，南京大学建筑与城市规划学院助理教授，硕士生导师；
杨晟一，南京大学建筑与城市规划学院硕士研究生；
许媛媛，南京大学建筑与城市规划学院硕士研究生。
- 基于多智能体模型的居住空间利用模拟与优化策略——以南京市花园路片区为例
- Simulation and Optimization Strategies of Residential Space Utilization Based on Agent-Based Modeling: A Case Study of Huayuan Road Area in Nanjing
- 戴思懿 秦萧 张姗琪 杨晟一 许媛媛
- DAI Siyi QIN Xiao ZHANG Shanqi YANG Shengyi XU Yuanyuan
- 摘要：
存量规划时代，精细化和人本化成为居住空间更新的重要方向，准确把握居民行为是更新的关键。传统方法难以满足这一需求，多智能体建模从使用者视角“自下而上”模拟人的空间行为，为居住空间优化提供了新思路。文章以居民与环境、居民间的互动关系分析为抓手，选取南京市花园路片区为研究对象，利用多智能体方法构建居住空间利用模型，通过调研数据对模型进行校验，并提出功能混合、用地开放、更新有序等居住空间优化策略。这不仅为传统居住空间利用研究提供新的视角与方法，也为面向未来的社区建设实践提供有效的路径参考。 - 关键词：
居住空间利用；多智能体模型；居民行为模拟；空间优化 - Abstract: In the era of stock-based planning, refinement and human-centered approaches have become key directions for the renewal of residential spaces, and accurately understanding residents’ behaviors is critical to effective renewal. Traditional methods struggle to meet this need. Agent-based modeling, which simulates human spatial behaviors from a bottom-up, user-centered perspective, offers a new approach to optimizing residential spaces. Focusing on interactions between residents and their environment as well as among residents themselves, this study takes the Huayuan Road area of Nanjing as a case study. An agent-based model of residential space utilization is constructed and validated using survey data. Based on the results, the study proposes optimization strategies for residential spaces, including functional mixing, open land use, and orderly renewal. These findings not only provide new perspectives and methods for traditional research on residential space utilization, but also offer practical pathways to guide future community development.
- Key words: residential space utilization; agent-based modeling; resident behavior simulation; spatial optimization