面向建成环境的可解释智能城市设计转型:综述与框架  

Explainable Intelligent Urban Design Transformations for theBuilt Environment: A Review and Framework

作　　者：杨俊宴[1,2] 孙昊成 黄雨悦 YANG Junyan;SUN Haocheng;HUANG Yuyue(School of Architecture,Southeast University;Southeast University,Smart City Research Institute of Southeast;Southeast University)

机构地区：[1]东南大学建筑学院 [2]东南大学智慧城市研究院副院长 [3]东南大学

出　　处：《世界建筑》2025年第3期67-70,共4页World Architecture

基　　金：国家自然科学基金,项目批准号:52394224;江苏省重点科技研发计划,项目批准号:BE2023799。

摘　　要：中国城镇化发展已进入中后期阶段,智能城市设计成为第四代数字化城市设计演进的重要方向之一。然而,人工智能技术的泛化应用导致设计同质化与“黑箱”生成问题,制约了城市设计的在地性与原创性,亟需构建可解释的智能设计方法。本文梳理了应对建成环境复杂性的4类智能设计方法,并引入可解释模型提升决策透明度,通过混合智能人机交互模式,整合人工智能与城市设计多主体的专家经验,最终构建了融合建成环境自组织规律与规划管控要求的可解释智能设计框架。As the development of urbanisation in China progresses into the middle and late stages,intelligent urban design has emerged as one of the significant directions in the evolution of the fourth-generation digital urban design.However,the widespread application of artificial intelligence technology has given rise to design homogenisation and the creation of black-box problems,which in turn restrict the locality and originality of urban design.There is an urgent need to develop explainable intelligent design methods.This paper reviews four types of intelligent design methods that address the complexity of the built environment and introduces explainable models to enhance decision-making transparency.By integrating artificial intelligence with the expertise of multiple stakeholders in urban design through a hybrid intelligent human-machine interaction mode,an explainable intelligent design framework that combines the self-organisation principles of the built environment with planning and control requirements is ultimately constructed.

关 键 词：建成环境 智能城市设计 可解释人工智能 人机交互 

分 类 号：TU984[建筑科学—城市规划与设计]