住区布局多目标自动寻优的模拟方法

1)深圳大学建筑与城市规划学院,广东深圳518060; 2)深圳市建筑环境优化设计研究重点实验室, 广东深圳 518060

城乡规划与设计; 性能模拟; 多目标优化; 参数化设计; 优化算法; 住区布局

A multi-objective auto-optimizing simulation method of residential layout design
YUAN Lei and LI Bingyao

YUAN Lei and LI Bingyao1)College of Architecture and Urban Plan, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China2)Shenzhen Key Laboratory for Optimizing Design of Built Environment, Shenzhen 518060, Guangdong Province, P.R.China

urban planning and design; performance simulation; multi-objective optimizing; parametric design; optimization algorithm; residential design

DOI: 10.3724/SP.J.1249.2018.01078

备注

为解决传统住区布局设计流程中性能模拟独立于主要设计过程之外的不足,提出一种由物理环境性能模拟驱动,以多目标优化为核心并集合了参数化建模技术的住区布局形态自动生成的集成设计方法——基于模拟的多目标自动优化设计. 采用该方法设计4组布局生成实验,分别以容积率、夏季得热、日照满足率、整体采光以及景观条件等要素的组合作为优化目标,以建筑单体的平面比例、高度、相对位置及组合形式参数等作为控制单体形态和总平面布局形态的自变量,进行住区布局设计寻优.结果表明,该方法可以在不同需求条件下提供满足不同目标侧重的多样化最优设计解集,并可得到最优解各项目标函数数据及其相互间量化的竞争关系.该方法在大幅提高设计优化效率的同时,还能为方案决策提供数据支持.

Traditional residential layout design is inadequate duo to its performance simulation independent of the main process of design.In order to solve the problem, we propose an integrated auto-design method, called MOOD-S(Multi-objective optimizing design based on simulation), which is driven by physical performance simulations, running with multi-objective optimizing algorithm, and coupled with parametric modelling. We introduce 4 different design tests using MOOD-S. In these tests, parameters including plot ratio, summer solar radiation gain, sunshine satisfaction rate, daylighting factor and sight view factor are selected to assemble the objective sets. The geometrical parameters of single building, relative positions and combination form characters are used as independent variables, playing the roles of controlling design optimization. The results show that this method can respond to different types of needs by providing feasible solution sets with high diversity. In addition, it can show the quantified competitive relations among the involved objectives which are helpful in scenario screening and selection,while greatly improving the efficiency of design optimization.

·