大语言模型助力社区生活圈规划与治理研究
The Empowerment of Large Language Models in Community Life Circle Planning and Governance
张文佳
同济大学建筑与城市规划学院 长聘教授，博士生导师
李博洋
北京大学城市规划与设计学院 硕士研究生
黄诺贤
北京大学城市规划与设计学院 硕士研究生
王雨润（通信作者）
北京大学城市规划与设计学院 博士研究生，yura_wang@stu.pku.edu.cn
武钰林
北京大学城市规划与设计学院 博士研究生
牛璐瑶
北京大学城市规划与设计学院 硕士研究生
摘要: 伴随着信息通信技术的发展，社区生活圈规划与治理的智能化水平不断提升，但仍面临动态服务智能定制和用户实时交互能力不足的挑战，Large Language Models (LLMs)技术的自然语言理解和知识整合能力为破解该难题带来机遇。通过时空知识图谱集成、地理邻近性增强检索与垂域大模型动态决策，搭建融合LLMs与检索增强生成技术的社区公共设施信息服务平台，实现本地公共服务设施信息智能助手、邻里在线社交平台、社区生活圈资源实时评估与优化等应用场景。深圳高校型社区与超高密度混合社区的试点结果显示，该体系能有效提升社区公共服务供给效率，增强居民互动与社区治理参与。展望LLMs在生活圈规划和治理中的双重应用，包括自上而下的政策宣导、规划反馈以及自下而上的需求预测与公众参与、项目建议和社区服务优化，反映大模型等新技术对生活圈规划与治理的潜在价值。
Abstract: With the advancement of information and communication technology (ICT), the intelligence of community life circle planning and governance has increased. Yet, challenges persist in dynamically customizing services and enabling real-time resident interaction.The natural language understanding and knowledge integration capabilities of Large Language Models (LLMs) present an opportunity to address these challenges. By integrating spatiotemporal knowledge graphs, geographically proximity-enhancedRetrieval-Augmented Generation (RAG) methodology, and domain-specific LLMs, this study develops a community public facility information service platform. This platform enables smart assistants for local public service facilities, online neighborhood social platforms, and real-time evaluation and optimization of community life circle resources. Pilot results from university-type communities and ultra-high-density mixed communities in Shenzhen show that this system effectively improves the efficiency of community public service provision and strengthens resident interaction and governance participation. The study also envisions the dual applications of LLMs in life circle planning and governance, including top-down policy communication and planning feedback, as well as bottom-up demand prediction, public participation, project recommendations, and community service optimization, highlighting the potential value of new technologies like large models in this field.
关键词：大语言模型（LLMs）；检索增强生成（RAG）；社区生活圈；社区治理
Keyword: Large Language Models (LLMs); Retrieval-Augmented Generation (RAG); community life circle; community governance
中图分类号：TU984
文献标识码: A
上海市规划和国土资源管理局. 15分钟社区生
活圈规划导则（试行）[EB/OL]. （2016-05-01）
[2025-03-15]. https://up.caup.net/file/life-circle.pdf.
Shanghai Municipal Bureau of Planning and
Natural Resources. 15-minute community life circle planning guidelines (trial)[EB/OL]. (2016-
05-01) [2025-03-15]. https://up.caup.net/file/life-
circle.pdf.
张文佳，李春江，罗雪瑶，等. 机器学习与社区生
活圈规划：应用框架与议题[J]. 上海城市规划，
2021（4）：59-65.
ZHANG Wenjia, LI Chunjiang, LUO Xueyao, et al. Machine learning and community life circle
planning: application framework and topics[J].
Shanghai Urban Planning Review, 2021(4): 59-65.
ZHENG Y, LIN Y, ZHAO L, et al. Spatial planning of urban communities via deep reinforcement
learning[J]. Nature Computational Science, 2023,
3(9): 748-762.
席广亮，甄峰，项欣怡，等. 智能技术作用下的城
市生活服务供需匹配研究进展与展望[J]. 地理
科学进展，2023，42（11）：2231-2241.
XI Guangliang, ZHEN Feng, XIANG Xinyi, et al. Research progress and prospects on supply-
demand matching of urban life services under the
influence of intelligent technology[J]. Progress in Geography, 2023, 42(11): 2231-2241.
孔宇，甄峰，张姗琪，等. 智能技术支撑的社区规
划：概念模型与技术框架[J]. 城市规划，2023，
47（1）：15-24.
KONG Yu, ZHEN Feng, ZHANG Shanqi, et al. Community planning supported by intelligent
technology: conceptual model and technical framework[J]. City Planning Review, 2023, 47(1):
15-24.
张姗琪，甄峰，孔宇，等. 基于虚实空间交互的
社区生活圈服务设施评估与优化配置：研究进
展与展望[J]. 自然资源学报，2023，38（10）：
2435-2446.
ZHANG Shanqi, ZHEN Feng, KONG Yu, et al. Evaluation and optimal allocation of community life circle service facilities based on virtual and
real space interaction: research progress and prospects[J]. Journal of Natural Resources, 2023, 38(10): 2435-2446.
蔡兴飞，王浩，李莉，等. 社区生活圈评估应用实
践、挑战及展望[J]. 规 划 师，2023，39（5）：
47-52.
CAI Xingfei, WANG Hao, LI Li, et al. Assessment and application practice, challenges and prospects
of community life circle[J]. Planners, 2023, 39(5):
47-52.
魏伟，洪梦谣，谢波. 基于供需匹配的武汉市15
分钟生活圈划定与空间优化[J]. 规划师，2019，
35（4）：11-17.
WEI Wei, HONG Mengyao, XIE Bo. Demarcation and spatial optimization of Wuhan's 15-minute life circle based on supply-demand matching[J].
Planners, 2019, 35(4): 11-17.
赵鹏军，罗佳，胡昊宇. 基于大数据的生活圈范
围与服务设施空间匹配研究——以北京为例[J].
地理科学进展，2021，40（4）：541-553.
ZHAO Pengjun, LUO Jia, HU Haoyu. Research on spatial matching between life circle range and service facilities based on big data: a case study
of Beijing[J]. Progress in Geography, 2021, 40(4):
541-553.
李颖，颜婷，曾艺元，等. 行为量化分析视角下的
公共服务设施有效使用评估研究[J]. 规划师，
2019，35（2）：66-72.
LI Ying, YAN Ting, ZENG Yiyuan, et al. Evalua-
tion of the effective use of public service facilities
from the perspective of behavioral quantitative
analysis[J]. Planners, 2019, 35(2): 66-72.
车冠琼，仇保兴，王倚天，等. 基于居民步行活动
模拟的社区生活圈规划：研究框架与议题[J]. 西
部人居环境学刊，2024，39（2）：139-143.
CHE Guanqiong, QIU Baoxing, WANG Yitian, et al. Community life circle planning based on simulation of residents' walking activities: research framework and topics[J]. Journal of Human
Settlements in West China, 2024, 39(2): 139-143.
罗雪瑶，张文佳，柴彦威. 15分钟生活圈的建成
环境阈值效应研究[J]. 地理研究，2022，41（8）：
2155-2170.
LUO Xueyao, ZHANG Wenjia, CHAI Yanwei. Study on the threshold effect of the built environ-
ment in 15-minute life circle[J]. Geographical Research, 2022, 41(8): 2155-2170.
伍磊，牛强，阿吉艾比布拉·艾尼瓦尔，等. 虚实融
合的线上线下社区生活圈：迭代升级与规划探
索[J]. 城市规划学刊，2024（2）：25-33.
WU Lei, NIU Qiang, AJIAIBIBULA Aniwaer,
et al. Online and offline community life circle integrating virtual and real worlds: iterative upgrading and planning exploration[J]. Urban Planning Forum, 2024(2): 25-33.
肖作鹏，林潇泓，韩来伟，等. 北京即时配送行为
网络结构识别及对生活圈规划的启示[J]. 北京
规划建设，2024（4）：93-98.
XIAO Zuopeng, LIN Xiaohong, HAN Laiwei, et al. Identification of network structure of instant delivery behaviors in Beijing and its implications for life circle planning[J]. Beijing Planning and
Construction, 2024(4): 93-98.
刘冰，王舸洋，朱俊宇，等. 基于共享单车大数据
的骑行生活圈识别及其活动网络模式分析[J].
城市规划学刊，2023（4）：32-40.
LIU Bing, WANG Geyang, ZHU Junyu, et al. Identification of cycling life circle and analysis of its activity network patterns based on shared
bicycle big data[J]. Urban Planning Forum, 2023(4): 32-40.
金红，陈礼珂，游兰，等. 基于地理时空关联和社
会影响的兴趣点推荐[J]. 计算机科学，2025，
52（5）：128-138.
JIN Hong, CHEN Like, YOU Lan, et al. Points of interest recommendation based on geographic
spatiotemporal association and social influence[J].
Computer Science, 2025, 52(5): 128-138.
郝赫，姚恩建，陈荣升，等. 基于深度学习的多模
式公交线网乘客路径诱导策略研究[J]. 交通工
程，2025，25（2）：1-10.
HAO He, YAO Enjian, CHEN Rongsheng, et al. Research on passenger path guidance strategies for multimodal bus networks based on deep learning[J]. Journal of Transportation Engineering,
2025, 25(2): 1-10.
上海市杨浦区人民政府. 量子城市 共植未来 |
2025年上海量子城市时空创新建设“上数营”
联创活动举行[EB/OL].（2025-03-14）[2025-03-
15]. https://www.shyp.gov.cn/shypq/xwzx-ypyw/ 20250314/476089.html.
Yangpu District Government. Quantum city, co-
planting the future: the 2025 Shanghai quantum city
spatiotemporal innovation and construction "Shang
Shu Ying" joint innovation activity held[EB/OL]. (2025-03-14) [2025-03-15]. https://www.shyp.gov.
cn/shypq/xwzx-ypyw/20250314/476089.html.
NI H, WANG Y, LIU H. Planning, living and judging: a multi-agent LLM-based framework for cyclical urban planning[J]. arXiv, 2024: 2412.20505.
ZHOU Z, LIN Y, JIN D, et al. Large language model for participatory urban planning[J]. arXiv, 2024: 2402.17161.
ZHU H, ZHANG W, HUANG N, et al. PlanGPT:
enhancing urban planning with tailored language model and efficient retrieval[J]. arXiv, 2024: 2402.19273.
PIAO J, YAN Y, ZHANG J, et al. AgentSociety: large-scale simulation of LLM-driven generative agents advances understanding of human behaviors and society[J]. arXiv, 2025: 2502.08691.
