摘要
随着全球城市化进程的加速,建筑能耗逐步增加,行人风舒适度的保障也变得愈加重要。如何同时优化场地布局和建筑布局以降低能耗并改善风环境,已成为亟待解决的关键问题。为此,本文提出了一种结合生成式AI与多目标优化的框架,旨在实现场地布局和建筑布局的快速生成与联合优化。该框架通过学习真实布局,采用扩散模型融入规划条件,生成多样式且可控的建筑和户型。在能效预测方面,引入XGBoost模型加速建筑能耗预测,显著提高计算效率。同时,采用扩散模型进行风环境舒适度的生成,使用NSGA-II算法在能效和风舒适度之间实现多目标优化。整体而言,本文提出的框架为可持续形态设计提供了创新的技术路径,为未来AI生成设计与优化结合的研究提供了重要参考。
With the rapid acceleration of global urbanization,building energy consumption has been steadily rising,and guaranteeing pedestrian wind comfort has become increasingly critical.How to simultaneously optimize site layout and building configuration to reduce energy demand while improving the wind environment is therefore a pressing challenge.To address this issue,we propose a framework that integrates generative AI with multi-objective optimization,enabling rapid generation and joint optimization of both site and building layouts.By learning from real-world examples,a diffusion model conditioned on planning constraints produces diverse and controllable building and floor-plan variations.For energy-performance assessment,an XGBoost model is introduced to accelerate energy-consumption prediction,greatly enhancing computational efficiency.Meanwhile,a diffusion model is also employed to generate wind-comfort evaluations.An NSGA-II algorithm then performs multi-objective optimization between energy efficiency and wind comfort.Overall,the proposed framework offers an innovative technological pathway for sustainable form-finding and provides a valuable reference for future research that couples AI-based generation with performance-driven optimization.
作者
高金顶
梁超
陈康
刘小平
GAO Jinding;LIANG Chao;CHEN Kang;LIU Xiaoping(School of Geographic Science and Planning,Sun Yat sen University;Guangdong Institute of Natural Resources and Environment;Zhongshan University)
出处
《世界建筑导报》
2025年第4期20-25,共6页
World Architecture Review
关键词
场地布局
建筑布局
能源消耗
风环境舒适度
优化
site layout
building layout
energy consumption
pedestrian wind comfort
optimization
