Figure 2 - uploaded by Kian Wee Chen
Content may be subject to copyright.
Source publication
In order to support exploration in the early stages of the design process, researchers have proposed the use of population-based multi-objective optimisation algorithms. This paper focuses on analysing the resulting population of design variants in order to gain insights into the relationship between architectural features and design performance. T...
Contexts in source publication
Context 1
... demonstration explores how self-shading geometry, envelope materials and window area will affect the overall performance of a simplified build- ing located in the Singapore climate. The design schema is illustrated in Figure 2. The schema is based on two 4x2 grids stacked on top of each other ( Figure 2a). ...
Context 2
... demonstration explores how self-shading geometry, envelope materials and window area will affect the overall performance of a simplified build- ing located in the Singapore climate. The design schema is illustrated in Figure 2. The schema is based on two 4x2 grids stacked on top of each other ( Figure 2a). There are four options for the location of the ver- tical core of the building: columns 1 and 5, columns 2 and 6, columns 3 and 7, or columns 4 and 8 ( Figure 2b). ...
Context 3
... design schema is illustrated in Figure 2. The schema is based on two 4x2 grids stacked on top of each other ( Figure 2a). There are four options for the location of the ver- tical core of the building: columns 1 and 5, columns 2 and 6, columns 3 and 7, or columns 4 and 8 ( Figure 2b). One grid is chosen from each remaining column to create the building form (Figure 2c). ...
Context 4
... are four options for the location of the ver- tical core of the building: columns 1 and 5, columns 2 and 6, columns 3 and 7, or columns 4 and 8 ( Figure 2b). One grid is chosen from each remaining column to create the building form (Figure 2c). By staggering volumes on top of each other it is possible to create self-shading geometries. ...
Context 5
... are a total of 256 pos- sible building forms that can arise. All the external walls have windows, the heights of which range from 1.2 m to 3.6 m (Figure 2d). The walls and windows are assigned a material (Figure 2e). ...
Context 6
... the external walls have windows, the heights of which range from 1.2 m to 3.6 m (Figure 2d). The walls and windows are assigned a material (Figure 2e). The building can be rotated 360º (Figure 2f ). ...
Context 7
... walls and windows are assigned a material (Figure 2e). The building can be rotated 360º (Figure 2f ). The design schema can gen- erate 752,640 possible design variants. ...
Citations
... Fuchkina et al. [15] introduced a framework for exploring parametric modeling with node-link diagrams, parallel coordinates plots, and selforganizing maps. In the context of the fitness landscape, Chen at al. [7] and Stasiuk et al. [41] used k-means clustering Visual analytics framework for architectural design optimization. The main view offers two alternative representations for exploring the design space: a continuous density map and a discrete glyph-based map. ...
In architectural design optimization, fitness landscapes are used to visualize design space parameters in relation to one or more objective functions for which they are being optimized. In our design study with domain experts, we developed a visual analytics framework for exploring and analyzing fitness landscapes spanning data, projection, and visualization layers. Within the data layer, we employ two surrogate models and three sampling strategies to efficiently generate a wide array of landscapes. On the projection layer, we use star coordinates and UMAP as two alternative methods for obtaining a 2D embedding of the design space. Our interactive user interface can visualize fitness landscapes as a continuous density map or a discrete glyph-based map. We investigate the influence of surrogate models and sampling strategies on the resulting fitness landscapes in a parameter study. Additionally, we present findings from a user study (N = 12), revealing how experts’ preferences regarding projection methods and visual representations may be influenced by their level of expertise, characteristics of the techniques, and the specific task at hand. Furthermore, we demonstrate the usability and usefulness of our framework by a case study from the architecture domain, involving one domain expert.
... Experiments examining the resulting population of alternate designs and providing insight into the relationship between architectural features and design performance were conducted by (Chen, Janssen and Schlueter, 2015). The experiments show that it is possible to gain general knowledge by linking architectural features to design performance. ...
Every time an object is built, it creates a relationship with the ground. Architects have a full responsibility to design the building by taking the ground into consideration. In the field of architecture, using data mining to identify any unusual patterns or emergent architectural trends is a nascent area that has yet to be fully explored. Clustering techniques are an essential tool in this process for organising large datasets. In this paper, we propose a novel proof-of-concept workflow that enables a machine learning computer system to cluster aspects of an architect's building design style with respect to how the buildings in question relate to the ground. The experimental workflow in this paper consists of two stages. In the first stage, we use a database system to collect, organise and store several significant architectural precedents. The second stage examines the most well-known unsupervised learning algorithm clustering techniques which are: K-Means, K-Modes and Gaussian Mixture Models. Our experiments demonstrated that the K-means clustering algorithm method achieves a level of accuracy that is higher than other clustering methods. This research points to the potential of AI in helping designers identify the typological and topological characteristics of architectural solutions and place them within the most relevant architectural canons
... The first method at high-level was simple unsupervised machine learning in the form of k-means clustering. K-means clustering is a centroid-based method to group data points into "k" groups or clusters according to their "distance" to each other (Everitt and Hothorn 2011, Han et al 2011, Chen et al 2015. The silhouette score, ranging from -1 to +1, is used to measure the data point's similarity to its cluster in contrast with other clusters (Rousseeuw 1987). ...
... Chen et al. define a twelve-dimensional parametric model of an abstract building geometry, resulting in a space of 752,640 design candidates [4]. They search this space with a multi-objective, Pareto-based GA that aims to minimize the building envelope's thermal transfer and cost and to maximize the available daylight, yielding 5,000 evaluated candidates. ...
... For example, Nagy et al. optimize an office layout in terms of six performance objectives [16]. Similarly to [4], they cluster and visualize a set of 10,000 design candidates, but have difficulty with identifying meaningful trends ( Figure 3). ...
This paper introduces performance-informed design space exploration (DSE) to question the relationship between explicit, quantitative optimization problems and "wicked", co-evolving architectural design problems and to support the reframing of architectural design optimization as a medium for reflection. The paper proposes selection, refinement, and understanding as key aspects of performance-informed DSE and surveys current approaches to performance-informed DSE: (1) Clustering and Pareto-based optimization support selection by reducing large numbers of parametric design candidates into smaller and more meaningful sets of choices. (2) Surrogate modelling supports refinement by approximating time-intensive simulations in real-time, which is important for interactivity. (3) Multi-variate visualizations and statistical analyses support understanding by providing insights into characteristics of design spaces and fitness landscapes. Finally, the paper discusses a novel tool for visual and interactive, performance-informed DSE, Performance Explorer. Performance Explorer combines the real-time feedback afforded by surrogate models with a multi-variate visualization of fitness landscapes. A user test of Performance Explorer uncovered several performance-informed DSE strategies followed by the participants. Consisting of different combinations of selection, refinement, and understanding, these strategies illustrate and-to some extent-validate the proposed framework for performance-informed DSE.
... Increasingly, researchers are exploring ways to implement data analysis and modeling techniques during the architectural design process. [32][33][34] One prominent application is data visualization for high-dimensional design spaces. 35,36 Others have proposed using statistical methods to consider relationships between variables and objectives, including Bayesian inference. ...
Many architectural designers recognize the potential of parametric models as a worthwhile approach to performance-driven design. A variety of performance simulations are now possible within computational design environments, and the framework of design space exploration allows users to generate and navigate various possibilities while considering both qualitative and quantitative feedback. At the same time, it can be difficult to formulate a parametric design space in a way that leads to compelling solutions and does not limit flexibility. This article proposes and tests the extension of machine learning and data analysis techniques to early problem setup in order to interrogate, modify, relate, transform, and automatically generate design variables for architectural investigations. Through analysis of two case studies involving structure and daylight, this article demonstrates initial workflows for determining variable importance, finding overall control sliders that relate directly to performance and automatically generating meaningful variables for specific typologies.
... Those algorithms might not generate an optimal solution but they help to understand the possible outcomes of the rules defined through the algorithm … Similarly, Chen et al. (2015) suggest that optimization methods should not only find high-performing solutions, but "give architects a better understanding of the relationship between architectural features and design performance." The framing of optimization as a generator of insights also is supported by (Bradner et al. 2014): ...
... … the aim is less on optimization per se and more on exploration: the results from optimization are about changing one's way of thinking more than choosing a single design and then realizing it. Schaffranek (2012), Chen et al. (2015), Bradner et al. (2014), and Stouffs and Rafiq (2015) emphasize that, for ADO, understanding optimization problems, i.e., fitness landscapes, is more important than finding "optimal" solutions. Johnson ...
... ure-found by a GA-according to 18 characteristics. Their analysis yields 80 clusters, i.e. 80 archetypal design candidates(Figure 4.1). However, due to this considerable number, a human designer might struggle to select a design for further development or to understand the characteristics of these 80 candidates in relationship to their performance.Chen et al. (2015) define a twelve-dimensional parametric model of an abstract building geometry, resulting in a space of 752,640 design candidates. They search this space with a multi-objective GA that aims to minimize the building envelope's thermal transfer and cost and to maximize the available daylight, yielding 5,000 evaluated candidates. They then ...
Increasing applications of parametric design and performance simulations by architectural designers present opportunities to design more resource and energy efficient buildings via optimization. But Architectural Design Optimization (ADO) is less widespread than one might expect, due to, among other challenges, (1) lacking knowledge on simulation-based optimization, (2) a bias towards inefficient optimization methods—such as genetic algorithms (GAs)—in the building optimization literature, (3) lacking state-of-the-art, easy-to-use optimization tools, and, perhaps most importantly, (4) the problematic integration of optimization with architectural design. This problematic integration stems from a contrast between “wicked” or “co-evolving” architectural design problems, which exhibit vague and changing problem definitions, and optimization problems, which require problem definitions to be explicit and unchanging.
This thesis presents an interdisciplinary study of ADO that draws on design theory, building optimization, mathematical optimization, and multivariate visualization. To address the first three challenges, the thesis (1) surveys existing optimization methods and benchmark results from the mathematical and building optimization literatures, (2) benchmarks a representative set of optimization methods on seven problems that involve structural, energy, and daylighting simulations, and (3) provides Opossum, a state-of-the-art, easy-to-use optimization tool. Opossum employs RBFOpt, a model-based optimization method that simultaneously “machine-learns” the shapes of fitness landscapes while searching for well-performing design candidates. RBFOpt emerges as the most efficient optimization method from the benchmark, and the GA as the least efficient. To mitigate the contrast between architectural and optimization problems, the thesis (4) proposes performance-informed design space exploration (DSE), a novel concept that emphasizes selection, refinement, and understanding over finding highest-performing design candidates, (5) presents Performance Maps, a novel visualization method for fitness landscapes, (6) implements Performance Maps in the Performance Explorer, an interactive, visual tool for performance-informed DSE, and (7) evaluates the Performance Explorer through a user test with thirty participants. The Performance Explorer emerges as more supportive and enjoyable to use than manual search and/or optimization from this test. In short, the thesis offers tools for ADO and performance-informed DSE that are more efficient and that better acknowledge the “wickedness” of architectural design problems.
... 7 One approach to elevate optimization to understanding comes from unsupervised machine learning. Chen et al. 8 search a space containing 752,640 design variants with a multi-objective genetic algorithm that aims to minimize the building envelope's thermal transfer and cost and to maximize the available daylight, yielding 5000 evaluated (i.e. simulated) variants. ...
This article presents a method to visualize high-dimensional parametric design spaces with applications in computational design processes and interactive optimization. The method extends Star Coordinates using a triangulation-based interpolation with Barycentric Coordinates. It supports the understanding of design problems in architectural design optimization by allowing designers to move between a high-dimensional design space and a low-dimensional Performance Map. This Performance Map displays the characteristics of the fitness landscape, develops designers’ intuitions about the relationships between design parameters and performance, allows designers to examine promising design variants, and delineates promising areas for further design exploration.
... Recently, ADO has received new understandings both as a generative design tool that provides starting points for further design exploration (Bradner et al. 2014) and as a representational tool that aids the understanding of design problems (Wortmann et al. 2015). Chen et al. (2015) attempt to group large numbers of design variants -found with a genetic algorithm-with a clustering method to better understand the relationship between design features and performance. Their effort is symptomatic of the need for human-understandable representations of fitness landscapes. ...
This paper presents a novel method to visualize high dimensional parametric design spaces with applications in computational design space exploration. Specifically, the visualization method presented here supports the understanding of design problems in architectural design optimization by allowing designers to move between a high dimensional design space and a low dimensional "performance map". This performance map displays the characteristics of the fitness landscape, develops designers' intuitions about the relationships between design parameters and performance, allows designers to examine promising design variants and delineates promising areas for further design exploration.
Designers in the built-environment disciplines increasingly solve problems using generative design methods, which promise novel and performant solutions to design problems but produce large design spaces that are challenging to explore. Design Space Exploration (DSE) interfaces have been used to understand, refine and narrow design spaces. Still, a critical analysis of current DSE interfaces shows a gap between their features and how designers explore and make decisions. We conducted a design study with domain experts to develop a DSE interface (DesignSense) that tightly integrates and adds to several features found separately in current DSE systems. We present a formative focus group evaluation, which suggested more areas for improvement and highlighted the need to distinguish designers from scientists as two user groups of DSE systems with varying needs, amongst other findings.
