FoAR 基于自组织邻接图生成平面布局图。论文标题:Generating floor plan diagrams using a self-organising adjacency graph
期刊:Frontiers of Architectural Research
作者:Mohamed Zaghloul, Ludger Hovestadt
发表时间:June 2026
DOI:10.1016/j.foar.2025.11.011
微信链接:点击此处阅读微信文章

FoAR是由高等教育出版社和东南大学建筑学院联合主办的全英文学术期刊
建筑学 / 城乡规划 / 风景园林
本刊已被 AHCI / EI / CSCD / Scopus / DOAJ / CSTPCD 收录
中国科技期刊卓越行动计划 英文领军期刊

ManuscriptTitle
论文题目
Generating floor plan diagrams using a self-organising adjacency graph
基于自组织邻接图生成平面布局图
Authors
作者
Mohamed Zaghloul (a)(b)* , Ludger Hovestadt (c)
(a) Healthy and Sustainable Built Environment Research Center (HSBERC), College of Architecture, Art and Design (CAAD), Ajman University, Ajman, United Arab Emirates
(b) Artificial Intelligence Research Center (AIRC), Ajman University, Ajman, United Arab Emirates
(c) Chair of Digital Architectonics, ETH Zu¨rich, Zu¨rich 8093, Switzerland
Abstract
论文摘要
In this study, we introduce a novel self-organising map algorithm-based approach to floor plan diagram generation. A graph distance function was used instead of the standard Euclidean distance, together with a graph grid that fits inside any floor plan shape. The cosine distance was used to determine the best-matching neuron in the training process. Machine learning was then used to optimise the diagram on the fly; specifically, an unsupervised learning technique automatically generated initial floor plan diagrams without the need for an existing training dataset. The proposed method uses an adjacency matrix that describes the relationship between spaces. As a vital part of the process, the user can initialise fixed positions for some zones and elements, such as the entrance, thereby influencing the adjacency zoning localisation outcome. A Mathematica interface allows users to set input entrance positions, draw floor plan outlines and specify the direct relations between spaces. To validate the feasibility of the proposed approach, we performed an experiment using a residential apartment. The main contribution of this study was in improving the automation of the initial planning and design stages, which are the first crucial stages in designing floor plans.
本研究提出一种全新的、基于自组织映射算法的建筑平面图解生成方法。该方法摒弃传统欧氏距离,改用图距离函数,并搭配可适配任意平面轮廓形态的图网格;训练阶段采用余弦距离筛选最优匹配神经元。研究引入机器学习实现平面图解实时优化:具体而言,该无监督学习方法无需预先提供训练数据集,即可自动生成初始平面布局方案。本文所提方法依托邻接矩阵表征各空间单元之间的拓扑关联。设计流程中支持人工设定部分功能分区与构件的固定点位(如建筑出入口),以此约束、引导空间邻接分区的排布结果。配套开发的 Mathematica 交互界面可供使用者录入出入口位置、绘制建筑外轮廓、定义空间之间的直接邻接关系。为验证该生成方法的可行性,本文选取住宅公寓案例开展试验。本研究核心创新在于提升建筑方案前期规划设计阶段的自动化程度,而前期策划是平面设计流程中至关重要的首要环节。
Keywords
关键词
Floor plan diagram / 平面图解
Machine learning / 机器学习
Self-organising map / 自组织映射
Unsupervised learning / 无监督学习
Graph distance function / 图距离函数
Cosine distance function / 余弦距离函数
Sections Title
章节标题
1. Introduction / 引言
2. Related research / 相关研究综述
3. Methodology / 研究方法
3.1. Floor plan graph initiation / 平面拓扑图初始化
3.2. Feature engineering / 特征工程
3.3. Weight initialisation / 权重初始化
3.4. SOAG training algorithm Input / 自组织邻接图训练算法输入
3.5. Output and SOAG interpretation / 输出结果与自组织邻接图解析
4. Experiment / 实验验证
4.1. Application interface—side input panel / 应用界面:侧边输入面板
4.1.1. Application interface—graph zones and relationships / 应用界面:拓扑分区与空间关系
4.1.2. Application interface—graph representation / 应用界面:拓扑图表征
4.1.3. Export and run controls / 导出与运行控制
5. Results and discussion / 结果与讨论
5.1. Qualitative comparison / 定性对比分析
5.2. Quantitative comparison / 定量对比分析
6. Conclusions / 结论
Illustrations
主要插图

▲ 图 1:左图:基于C. Lueder(2012)的文献整理,借鉴勒?柯布西耶思想绘制的概念泡泡图。中图:基于诺布斯(1937)的记载,参照Nobbs, P.E. 1932年方案绘制的空间连接概念图解。右图:依据怀特(1975)本人论述,参考E.T. White 1975年提出的概念泡泡图。©本文作者

▲ 图 2:自组织邻接图的工作流程。©本文作者

▲ 图 3:基于自组织邻接图(SOAG)的平面图解工具操作界面,以随机彩色圆形区域区分。©本文作者

▲ 图 4:部分分区对应的神经元权重。白色单元格代表该分区匹配概率为 100%;黑色单元格代表该分区匹配概率为 0%。©本文作者

▲ 图 5:分位数对比图:对角线代表真实基准(GT)面积,散点为自组织邻接图(SOAG)生成方案对应的面积值。©本文作者

▲ 图 6:(a) 分区权重值与自组织邻接图生成结果对比。(b) 豪斯多夫距离直方图。©本文作者
Authors Information
作者简介

Mohamed Zaghloul*
Assistant Professor
Healthy and Sustainable Built Environment Research Center (HSBERC), College of Architecture, Art and Design (CAAD), Ajman University, United Arab Emirates
He is interested in developing machine-learning based tools that start from abstract features of any model and proceed into indexing, synthesizing designs, and developing probabilistic-based emulators of building performances in order to create intuitive and fast design performance. He joined the chair in November 2012. He holds MSc degree in architecture from and worked as a teaching assistant at the Faculty of Fine Arts Department of Architecture at Alexandria University from 2003 to 2017 and lecturer from 2017 to 2019. He is Co-founder and Research Director at Encode Studio.

Ludger Hovestadt
Professor
Chair of Digital Architectonics, ETH Zürich, Zürich 8093, Switzerland
Ludger Hovestadt is Professor of Architecture and CAAD (Computer-Aided Architectural Design) at the Institute for Technology in Architecture, ETH Zurich. Between 1997 and 2000 Ludger Hovestadt was a visiting professor in the Department of CAAD at University of Kaiserslautern, Germany. In 2000, he was appointed Full Professor at the Department of Architecture at ETH. Since 2018 he is a Visiting Professor at Southeast University in Nanjing, China. He studied architecture at the RWTH Aachen, Germany and the HfG in Vienna, Austria. Upon completion of his diploma in 1987, he started his academic career with Prof. Fritz Haller at TU Karlsruhe for whom he worked as a scientific researcher for over ten years. Under his supervision Ludger Hovestadt completed his doctorate at TU Karlsruhe and Carnegie Mellon in 1994, expanding Haller’s school of thought of minimalist-functional mannerism into the digital domain, thereby laying the foundations for digital architectonics.
▼ 点击下方词条 往期精彩不容错过
#期刊快讯#系列
主编王建国院士团队荣获国家科技进步奖一等奖
JCR 2025 影响因子4.4建筑学第一
CiteScore 2025 指数上升为7.8
FoAR 被中国科技论文与引文数据库CSTPCD收录
FoAR 荣获2025科爱三项期刊奖
2025年度中科院分区表大类小类FoAR保持一区
FoAR 入选"卓越行动计划"英文领军期刊
FoAR 进入 EI 工程索引数据库
FoAR 在国际影响力 TOP 期刊榜单中排名稳步上升
#主办会议# 系列
会议回顾 FoAR青年学者之夜
会议回顾 智汇浦江: FoAR青年学者学术沙龙
会议回顾 FoAR专题会议:历史空间的跨媒介再现
会议回顾 FoAR专题会议:面向气候环境可持续的街区设计
会议回顾 皖美遇见: FoAR青年学者学术沙龙
会议回顾 FoAR专题会议: 高性能建筑维护结构
活动招募 FoARxJoA建筑国际写作营
#年度报告# 系列
FoAR 2020年度报告
FoAR 2021年度报告
FoAR 2022年度报告
FoAR 2023年度报告
FoAR 2024年度报告
FoAR 2025年度报告
#新刊上线#系列
2025年第四期
2025年第五期
2025年第六期
2026年第一期
2026年第二期
2026年第三期
#FoAR投稿指南#系列
投稿流程
Guide for Author中文版
审稿流程及发表流程
如何提交修改稿
录用文章出版流程
#期刊知识科普#系列
SCI之父尤金·加菲尔德的传奇人生
国际核心期刊数据库大解析
如何使用最强工具Web of Science
如何发现一本好期刊
国内核心期刊有哪些
版面费与期刊影响力
期刊联络
高等教育出版社:010-58556484
东南大学:025-83795543
刊物邮箱:foar@pub.seu.edu.cn
FoAR英文期刊交流QQ群:21608832
在线投稿
www.editorialmanager.com/foar
刊物主页
http://www.sciencedirect.com/science/journal/20952635
http://journal.hep.com.cn/foar
来源:Frontiers of Architectural Research