Conference PaperPDF Available

Optimization-Assisted Building Design Cases study of design optimization based on real-world projects

Authors:

Abstract and Figures

Computational design optimization has been widely considered a promising technique to help designers address complex design challenges regarding building performance. However, a barrier to applying it to real-world projects is the difficulty in incorporating functional requirements and constraints into the design optimization process. In response, this study presents an optimization-assisted design approach for early-stage architectural design. The approach combines the application of EvoMass, an integrated building mass design generation and optimization tool, and the soft constraint strategy. The combination allows designers to integrate various design requirements and constraints into the optimization, which makes it produce results with higher practical values. To demonstrate the efficacy of the approach, two case studies are presented, which show that the application of optimization facilitates designers to better formulate the design problem and rapidly investigate different design directions for exploration and information extraction. INTRODUCTION Building performance has become a prominent concern in early-stage architectural design in response to the goal of sustainable urban development, and many architects have begun to adopt the performance-based design approach in their design practice. An important responsibility for architects is to explore alternative design possibilities and compare the performance of these solutions in order to obtain more significant performance improvements. However, when this exploration is conducted manually, time restraints in real-world projects often limit the scope of architects' design exploration, which also makes early-stage design exploration both time-consuming and challenging. The capacity of computational design optimization to assist architects in early-stage design exploration has been demonstrated over the last decade by research advancements in the use of computational design optimization in architecture. Computational design optimization can automate the design generation and exploration process by combining parametric modeling, performance simulations, and optimization algorithms, freeing architects from the time-consuming trial-and-error process of finding high-performing solutions.
Content may be subject to copyright.
Optimization-Assisted Building Design
Cases study of design optimization based on real-world projects
Donglai Yang1, Likai Wang2, Guohua Ji3
1,2,3Nanjing University
1mg20360025@smail.nju.edu.cn 2wang.likai@outlook.com 3jgh@nju.edu.cn
Computational design optimization has been widely considered a promising technique to
help designers address complex design challenges regarding building performance.
However, a barrier to applying it to real-world projects is the difficulty in incorporating
functional requirements and constraints into the design optimization process. In response,
this study presents an optimization-assisted design approach for early-stage architectural
design. The approach combines the application of EvoMass, an integrated building mass
design generation and optimization tool, and the soft constraint strategy. The combination
allows designers to integrate various design requirements and constraints into the
optimization, which makes it produce results with higher practical values. To demonstrate
the efficacy of the approach, two case studies are presented, which show that the
application of optimization facilitates designers to better formulate the design problem
and rapidly investigate different design directions for exploration and information
extraction.
Keywords: Generative Design, Optimization, Design Exploration, Design Process,
EvoMass, Computational Design, Building Performance.
INTRODUCTION
Building performance has become a prominent
concern in early-stage architectural design in
response to the goal of sustainable urban
development, and many architects have begun to
adopt the performance-based design approach in
their design practice. An important responsibility for
architects is to explore alternative design
possibilities and compare the performance of these
solutions in order to obtain more significant
performance improvements. However, when this
exploration is conducted manually, time restraints in
real-world projects often limit the scope of architects’
design exploration, which also makes early-stage
design exploration both time-consuming and
challenging. The capacity of computational design
optimization to assist architects in early-stage design
exploration has been demonstrated over the last
decade by research advancements in the use of
computational design optimization in architecture.
Computational design optimization can automate
the design generation and exploration process by
combining parametric modeling, performance
simulations, and optimization algorithms, freeing
architects from the time-consuming trial-and-error
process of finding high-performing solutions.
Background
In research, the use of computational design
optimization in architectural design has been
noticed frequently. Many research studies have been
undertaken over the last decade to apply it to
architectural design challenges with varied
performance targets such as wind environment
Volume 1 – Co-creating the Future – eCAADe 40 | 609
(Chronis et al., 2011), sunlight (De Luca, 2017), and
energy consumption (Yi & Malkawi, 2012). However,
the application of computational design
optimization in early-stage architectural design is
still quite limited in practice. One of the reasons for
this is the dilemma between design effectiveness
and variability. This is because function-related
design requirements often introduce conflicts with
the optimization of building performance, and
existing applications of computational design
optimization in architectural design are often unable
to reconcile both building performance and
functional requirements at the same time.
On the one hand, encoding the design
requirements as hard-coded rules in the parametric
model provides a simple way to avoid generating
invalid designs that do not satisfy functional
constraints (Janssen, 2005; Wang et al., 2016). This
method guarantees that the design obtained by the
optimization is valid. However, hard-coded
restrictions decrease the parametric model's design
variability, limiting optimization to a relatively
limited scope for high-performance solutions.
Furthermore, hard-coded rules impose unnecessary
limits on parametric models, reducing their
reusability for other design tasks.
On the other hand, some applications placed
more emphasis on design variability and adopted
flexible design generation approaches such as free-
form (Yi & Malkawi, 2012) and voxel geometries (Si &
Wang, 2015). While the design flexibility of these
approaches allows the optimization process to
examine a large range of design possibilities with
considerable changes, the lack of limitations often
results in solutions that do not match functional
requirements or design codes.
Paper overview
Considering the limitations of the previous studies,
this paper presents a study focusing on the
application of computational design optimization to
real-world architectural design projects. The study is
based on two design cases, in which EvoMass, a
building mass design generation and optimization
tool in Rhino-Grasshopper (Wang, Chen, et al.,
2020a) is used. EvoMass provides more flexibility to
be adapted for diverse site conditions and to
optimize the design for different performance
targets than other building design generating
algorithms created for specific design jobs. Over the
past few years, EvoMass has been applied to a variety
of design tasks and has shown its advantages in
assisting designers in early-stage exploration and
information extraction.
While previous applications have demonstrated
the potential of EvoMass in solving various
architectural design tasks, these applications often
fell short of the consideration of real-world design
constraints related to design specifications and
functional requirements. In this regard, this study
further exploits the potential of EvoMass in more
realistic design scenarios and proposes a design
approach incorporating EvoMass with other soft
constraint strategies. This approach enables
designers to rapidly explore different directions to
gather information about the design problem. In
addition, the use of soft constraints allows designers
to specify various functional requirements during
the optimization process. The combination of these
two components achieves an optimization-assisted
design approach that encourages designers to
interact with the computer for co-evolution and
human-in-the-loop design.
The proposed approach was developed during
our design of two real-world projects where design
codes and provisions imposed strong restrictions on
the application of the computational design process.
Hence, this paper uses these two projects as case
studies and demonstrates how the proposed
approach can be applied to the design process. As
such, one major contribution of this study is to
examine and demonstrate the utility of EvoMass in
more realistic design scenarios.
OPTIMIZATION-ASSISTED DESIGN
This study aims to explore the potential of
computational design optimization in architectural
design, with an emphasis on how optimization can
610 | eCAADe 40 – Volume 1 – Co-creating the Future
assist designers in design exploration involving
functional requirements. As stated above, the
approach is based on Rhino-Grasshopper, which
makes it easier for design practitioners to use it.
Technically, the approach is achieved by using
EvoMass and other performance simulation tools
such as Ladybugs and ClimateStudio. Regarding
requirements related to design codes and
regulations, the approach uses soft constraints to
transform these requirements into penalty or award
functions that navigate and guide the optimization
process.
Design approach
Compliance with design rules such as setback and
spacing is one of the most crucial, yet time-
consuming, duties for architects. More recently,
increased demands for the environment,
sustainability, and people's health and well-being
have added to the design challenges, making
simulation an imperative in design routines. As a
result, trial-and-error design processes using
simulation verification are becoming increasingly
common in most design firms and offices (Figure 1
top). Such design approaches limit the designer's
ability to explore numerous design possibilities
during the fast-paced conceptual design stage
because there isn't enough time to verify that each
design option fits all performance goals as well as
functional needs and restrictions. Therefore, it is
typical to see that designers will soon delve into a
specific design scheme and then focus on subtle
adjustments to the design to meet all requirements,
resulting in the designer's inability to coordinate all
aspects of the design as a whole.
The above-mentioned limitations highlight the
need for agile and rapid design exploration that
allows the designer to investigate different
directions at the outset of the design process. In this
regard, design optimization can be used to
substitute designers’ work on design tweaking and
code compliance. With this approach, designers can
be relieved from repetitive routine tasks and devote
more effort to design exploration and ideation. In
other words, the computer becomes a co-designer,
actively searching for high-performance but also
satisfying the design needs of the human designer.
At the same time, the human designer reflects on the
solution produced by the computer and iteratively
indicates a new or more detailed direction for the
computer to further explore (Figure 1 bottom). As
such, a feedback loop between designers and
computers is established.
EvoMass
EvoMass is the core component in the proposed
optimization-assisted design approach. On the one
hand, EvoMass's reusability allows it to be applied to
a variety of design jobs, relieving the designer of the
burden of parametric modeling and speeding up the
Figure 1
Comparison
between manual
design and
optimization-
assisted design
workflow
Volume 1 – Co-creating the Future – eCAADe 40 | 611
design process. When using EvoMass, designers can
customize the generated building design with
various user-defined parameters. The design
variability of EvoMass, on the other hand, allows the
optimization to actively respond to changes in the
design target and optimization objectives.
In EvoMass, two generative components encode
the subtractive and additive form generation
principles respectively (Wang, Chen, et al., 2020b).
Because of the geometrical freedom inherited by
these two concepts, the produced designs have a lot
of topological variety. In addition to the design
generation, EvoMass also includes an optimization
algorithm called the steady-state island evolutionary
algorithm (SSIEA) (Wang, Janssen, et al., 2020). SSIEA
integrates an island-based approach and a steady-
state replacement strategy into an evolutionary
algorithm. The island-based approach subdivides
the design population into multiple subpopulations,
each guided to focus on different regions in the
design space, which is aimed at increasing the
diversity in the design population. The steady-state
replacement strategy aims to exclude inferior
designs from the design population more rapidly,
compared with the generational replacement
strategy. Thus, this strategy facilitates the fast
convergence of the optimization process and
shortens the time for the optimization. The
combination of the two strategies in SSIEA ensures
optimum efficiency and effectiveness when using
EvoMass.
Soft constraints and design workflow
The use of EvoMass lays the foundation for the
applicability of the proposed optimization-assisted
design exploration, and previous studies have
shown that combining EvoMass with other
performance simulation tools can enable effective
building design optimization for performance
improvement. However, employing EvoMass alone
is insufficient to produce desirable solutions that
satisfy both the demand for improved building
performance and functional restrictions. Thus,
further intervention is a necessity for the effective
application of EvoMass to practice. In the proposed
design approach, soft constraints are adopted to
establish communication between designers and
computers. Soft constraints have been proved to be
effective to navigate the design optimization
process and compel the optimization to search for
the design satisficing various input objectives and
constraints (Chen et al., 2022).
By combining EvoMass and the soft constraint
strategy, a human-in-the-loop design workflow can
be established (Figure 2). In this design workflow,
designers first customize EvoMass to regulate the
building mass design generation and specify
performance objectives and functional constraints
Figure 2
Demonstration of
the design
workflow
612 | eCAADe 40 – Volume 1 – Co-creating the Future
for the design optimization. Regarding design
optimization, a single-objective optimization mode
is used in this approach, where performance
objectives and functional constraints are integrated
into a fitness evaluation function using weight-
product approaches. This study does not apply
Pareto optimization to handle multiple objectives
since the Pareto set can contain too many options,
which could impede designers’ information
extraction and pattern recognition. The design
feedback, from the optimization, can stimulate the
designer to further adjust the optimization
objectives or design constraints in order to explore
other alternative directions. In comparison, such
iterative design workflows are often absent in most
relevant studies due to the inflexibility of the design
tool. Thus, this study can also provide examples of
how to apply optimization to realistic design
problems.
CASE STUDY
Two case studies, one for a kindergarten and the
other for a primary school, are presented to
demonstrate the viability of the proposed design
method. Both design tasks are based on real-world
design projects that require consideration of a
variety of functional requirements and constraints.
Moreover, the presented case studies also reflect our
actual design process in dealing with these issues.
Regarding the kindergarten design, this is the
first task where we applied EvoMass to a real-world
design project. Thus, this case study demonstrates
how we become aware of the problem in a routine
application of computational design optimization
and how we applied the soft constraint strategies to
coordinate the requirements of performance
improvements and functional constraints.
Regarding the primary school design, we
learned from the kindergarten design and
incorporated multiple functional constraints at the
beginning of the design process. In this case study,
we applied the workflow to quickly explore two
different school planning layouts to test the
feasibility of the design.
Kindergarten design
The project is located in Nanjing, China, and the
kindergarten is required to have 15000 m2 with a
maximum floor number of four. As shown in Figure
3, there are several high-rise office buildings on the
east of the building site, which produce noticeable
context shading on the kindergarten site. Since
sunlight accessibility is an important factor in
kindergarten design, the context shading can cause
a critical design challenge for this project.
For the design generation, the user-defined
parameters will be set based on common
architectural design parameters. The column grid
spacing is set to be 6 and 4 (meters) in the X and Y
directions; the maximum floor number is four with a
4.2 m floor height.
Considering the design objective, we first target
to maximize the sunlight accessibility of the building
and the kindergarten yard/playground for outdoor
activities. According to the Kindergarten Design
Code, nursery units must have at least two hours of
direct sunlight from 9:00 to 15:00 on the winter
solstice, and more than half of the yard must enjoy
more than three hours of direct sunlight. In order to
calculate the proportion of the building façade
surface and the area of the yard that cannot match
the above-mentioned requirements, we translate
these two requirements into two performance
indicators. As a result, the optimization goal is to
reduce the value of these two indicators.
For the optimization, we use the weight product
approach to integrate the above evaluation into a
single-objective design evaluation function. Figure 4
Figure 3
Site condition of
the kindergarten
project
Volume 1 – Co-creating the Future – eCAADe 40 | 613
illustrates the four optimized designs selected from
the optimization results. These four design options
show a shared tendency that the building spreads
alongside the southwest edge of the site in order to
maximize its sunlight. In addition, due to the context
shading, the northeast corner of the building site is
left empty as this region has the least sunlight
accessibility.
While these designs effectively address the
sunlight accessibility and solar exposure issues, the
problem is that these designs typically have a large
proportion of the building volume on the fourth
floor. The reason for this is that the higher the floor,
the less sunlight is blocked by the surrounding
structures. These design solutions also reveal a trend
to extend the west façade to boost afternoon solar
exposure in order to maximize sunlight accessibility.
Excessive solar exposure to the west, on the other
hand, may cause the building to overheat, increasing
the cooling load in the summer.
By reflecting on the flaws in the optimized
designs, we further added three additional
constraints to the optimization. First, we calculate
the floor area ratio from the first to the third floor
with respect to the GFA of the whole building.
Because only the first three floors are accessible to
children, this criterion decides whether the first three
floors have enough space to accommodate all of the
nursery units.
Second, we change the measurement of the
sunlight accessibility of the building. We only
include the façade surface below the fourth floor in
the second stage to avoid the fourth floor's influence
on the design results.
Third, maximizing the south-facing façade area
is included in the optimization to avoid the design
with a smaller south-facing façade area. This
constraint is important because it is easier for the
design with a smaller south-facing façade to achieve
a higher proportion of façade surfaces with sufficient
sunlight accessibility. Such designs, nevertheless,
have little practical value as only a few nursery units
can have south-facing windows.
With the modification of the design evaluation,
the performance metrics for the second design
optimization run are summarized in Table 1. On this
basis, the second optimization run is conducted.
Figure 5 demonstrates the result of the optimization.
It is worth noting that the inclusion of additional
functional metrics has a significant impact on the
optimization results.
Compared with the results of the first
optimization run, the optimized designs of the
second optimization run have higher practical value
for the subsequent design development. While, in
general, the building is spread alongside the
southwest edge, changes in the measurement of
sunlight accessibility force the design to elongate
further along the east-west direction to enlarge the
south-facing façade. In addition, the overall building
volume is also lowered compared with the
optimized design found from the previous
optimization result. This provides more space for the
designer to arrange the nursery units. However, due
to the enlargement of the building massing, the
sunlight accessibility of the kindergarten yard is
reduced. Hence, these two optimization runs also
provide clues to the trade-off between the sunlight
Figure 4
Selected high-
performing designs
from the first
optimization run
614 | eCAADe 40 – Volume 1 – Co-creating the Future
Optimization
objectives
Target
value Abbr
gross floor area 15000m2 p_area
building coverage 30% p_covera
ge
south façade area maximize area_sout
h
The proportion of the
building façade with less
than three hours of
sunlight accessibility
minimize sunlight_f
acade
The proportion of the
yard area above three
hours of direct sunlight
>50% sunlight_
yard
accessibility of the building and the yard.
In summary, the first case study demonstrates a
design process that echoes Schon’s reflective
practitioners (Schön, 1992). In this presented design
process, a moving-seeing-moving process emerges.
Compared to the manual design modification
process, the modification in this case study depends
on the definition of the optimization problem.
Therefore, it is foreseeable that the designer can
identify more constraints in the design and adjust
the objectives and constraints in design
optimization by continuing the reflective design
process.
Primary school design
The second case study describes a primary school
project also located in Nanjing, China. As shown in
Figure 6, the building site of the project is
surrounded by several high-rise residential towers
on the west and south, and there is a middle school
on the east. According to the design brief, the target
GFA of the school is 25000m2.
The Primary School Design Code, like the
kindergarten design, stipulates that the school has
adequate sunlight accessibility and natural
daylighting. Furthermore, as compared to
kindergarten design, primary school design must
take into account noise reduction and
communication between classrooms and other
amenities. Therefore, based on the optimization
objective defined in the kindergarten project, we
include three additional factors.
Figure 5
Selected high-
performing designs
from the second
optimization run
Table 1
Optimization
objectives of the
first case study
Figure 6
Site condition of
the primary school
project
Volume 1 – Co-creating the Future – eCAADe 40 | 615
Optimization
objectives
Target
value Abbr
gross floor area 25000 m2 p_area
Building coverage 25% p_density
south façade area maximize area_south
Similar to the first
case study minimize sunlight_faca
de
Spatial Daylight
Autonomy maximize sDA
Noise from the
sports field minimize p_noise
number of massing 1 p_num_massi
ng
floor area ratio >75% floor_ratio
8
First, spatial daylight autonomy (sDA) determines
the daylight accessibility of the building, which
indicates the proportion of the interior spaces that
can have sufficient daylight. Second, sports fields are
frequently the main source of noise in primary
schools. Therefore, a buffer zone around the sports
field is defined, and the overlap area between the
buffer zone and the building is calculated to
measure the indoor space that is more likely to be
affected by the noise from the sports field. Third, in
order to improve connection, the number of
building blocks was evaluated, which shows the
degree of separation between the buildings. Table 2
outlines the first case study's aims and restrictions, as
well as the factors considered in the second case
study's optimization. These variables are combined
using a weight-product function, just as they were in
the first case study.
Table 2
Optimization
objectives of the
second case study
Figure 7
The layout result of
the sports field in
the southeast
corner and the
scores of various
indicators (the four
best designs)
Figure 8
The layout result of
the sports field in
the southwest
corner and the
scores of various
indicators (the four
best designs)
616 | eCAADe 40 – Volume 1 – Co-creating the Future
Aside from the optimization objectives and
restrictions, the location of the sports field is also
taken into account in this project. The location of the
sports field in primary school design is critical in
deciding the overall building layout, which can have
an impact on the design's performance. As a result,
two prospective locations were specified in order to
explore the impact of the sports field on this design
project, and the design parameters and limitations
for the building mass generation were altered
correspondingly.
Lastly, the column spacing in the X and Y
directions is set to 12 and 8 (meter) in the design
generation, which is based on the standard
classroom dimensions; the maximum floor number
is five with a 4.2 m floor height.
Figures 7 and 8 show the optimization results
based on two different sports court locations within
the building site. In general, placing the sports field
on the southeast side (Figure 7) is more conducive to
achieving better results in terms of overall fitness,
and the corresponding design typically has higher
sDA and sunlight accessibility. Taking both
optimization results into account, it can be found
that dividing the buildings into three rows is a viable
building plan choice (options 1-2, 1-4, 2-3, 2-4). On
the contrary, the designs with relatively higher
fitness (options 1-1, 2-1) achieve better overall
performance, typically due to connectivity. As all the
building blocks are mutually connected, this feature
benefits the accessibility of the design. However, it
undermines the daylight and sunlight accessibility
because of the mutual shading between the
building blocks.
In this case study, the optimization facilitates the
design exploration and allows the designer to
investigate two different directions, and additional
design exploration can be stimulated after the
existing optimization runs. The use of optimization
can help designers rapidly identify promising
solutions or hidden constraints in the design,
without which designers may need to spend more
time and effort to find these solutions and
constraints. Thus, the computer in this project acts as
a co-designer that helps designers to delve into each
design direction and provide feedback to designers’
decision-making.
DISCUSSION AND CONCLUSION
The two case studies show that the combination of
EvoMass and soft constraint strategies allows for an
iterative and human-in-the-loop design process
assisted by computational design optimization. In
the design process, the computer helps the designer
conduct the tedious “number-crunching” for finding
satisfying designs and code compliance.
The case studies also demonstrate the
applicability of the proposed design approach to
practical building projects. From the design
perspective, EvoMass’ capability of providing timely
feedback and abstract design information is key to
achieving its utility. In other words, the proposed
approach is not intended to provide designers with
definite solutions, but, instead, the result of the
optimization is aimed at serve as a “medium for
reflection” (Wortmann & Schroepfer, 2019).
To conclude, this study presents an
optimization-assisted design approach for early-
stage architectural design. As shown in the cases, the
use of computational design optimization facilitates
the designer’s design exploration and information
extraction. Moreover, in comparison to existing
studies, this study is also focused on how to include
functional requirements and constraints from the
real-world project into the optimization process. The
result demonstrates that the inclusion of these
requirements and constraints makes the
optimization produce results with greater practical
values, which assists designers in rapidly identifying
promising design directions and hidden constraints.
ACKNOWLEDGEMENTS
This study is funded by China Postdoctoral Science
Foundation (2021M701664) and National Natural
Science Foundation of China (52178017).
Volume 1 – Co-creating the Future – eCAADe 40 | 617
REFERENCE
Chen, Y., Lu, Y., Gu, T., Bian, Z., B, L. W., & Tong, Z.
(2022). From Separation to Incorporation - A
Full-Circle Application of Computational
Approaches to Performance-Based
Architectural Design. In Proceedings of the 2021
DigitalFUTURES (pp. 189–198). Springer
Singapore.
Chronis, A., Turner, A., & Tsigkari, M. (2011).
Generative Fluid Dynamics: Integration of Fast
Fluid Dynamics and Genetic Algorithms for
Wind Loading Optimization of Free Form
Surfaces. SimAUD2011, Malkawi 2004, 79–85.
De Luca, F. (2017). From Envelope to Layout.
Buildings Massing and Layout Generation for
Solar Access in Urban Environments. Sharing
Computational Knowledge! ShoCK! Proceedings
of the 35th International Conference on Education
and Research in Computer Aided Architectural
Design in Europe (ECAADe) Vol. 2, 2(Willis 1995),
431–440.
Janssen, P. (2005). A design method and
computational architecture for generating and
evolving building designs. The Hong Kong
Polytechnic University.
Schön, D. A. (1992). Designing as reflective
conversation with the materials of a deisgn
situation. Knowledge-Based Systems, 5(1), 3–14.
Si, F., & Wang, T. (2015). Building massing
optimisation in early design stage. 20th
International Conference of the Association for
Computer-Aided Architectural Design Research in
Asia, 583–592.
Wang, L., Chen, K. W., Janssen, P., & Ji, G. (2020a).
ENABLING OPTIMISATION-BASED EXPLORATION
FOR BUILDING MASSING DESIGN-A Coding-free
Evolutionary Building Massing Design Toolkit in
Rhino-Grasshopper Design Automation Lab View
project Advanced Manufacturing Method for
Concrete Buildings View project ENABLI.
Wang, L., Chen, K. W., Janssen, P., & Ji, G. (2020b).
Algorithmic generation of architectural massing
models for building design optimisation:
Parametric modelling using subtractive and
additive form generation principles. RE:
Anthropocene, Design in the Age of Humans -
Proceedings of the 25th International Conference
on Computer-Aided Architectural Design Research
in Asia, CAADRIA 2020, 1, 385–394.
Wang, L., Janssen, P., & Ji, G. (2020). SSIEA: a hybrid
evolutionary algorithm for supporting
conceptual architectural design. Artificial
Intelligence for Engineering Design, Analysis and
Manufacturing, 34(4), 458–476.
Wang, L., Tan, Z., & Ji, G. (2016). Toward the Wind-
related Building Performative Design: A wind-
related building performance optimization
design system integrating Fluent and
Rhinoceros based on iSIGHT. Living Systems and
Micro-Utopias: Towards Continuous Designing,
Proceedings of the 21st International Conference
on Computer-Aided Architectural Design Research
in Asia (CAADRIA 2016), 209–218.
Wortmann, T., & Schroepfer, T. (2019). From
optimization to performance-informed design.
In S. Rockcastle, T. Rakha, C. C. Davila, D.
Papanikolaou, & T. Zakula (Eds.), 2019
Proceedings of the Symposium on Simulation for
Architecture & Urban Design (Vol. 51, Issue 8, pp.
261–268).
Yi, Y. K., & Malkawi, A. M. (2012). Site-specific
optimal energy form generation based on
hierarchical geometry relation. Automation in
Construction, 26, 77–91.
618 | eCAADe 40 – Volume 1 – Co-creating the Future
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
Using performance-based optimization to explore unknown design solutions space has become widely acknowledged and considered an efficient approach to designing high-performing buildings. However, the lack of design diversity in the design space defined by the parametric model often confines the search of the optimization process to a family of similar design variants. In order to overcome this weakness, this paper presents two parametric massing generation algorithms based on the additive and subtractive form generation principles (food4rhino.com/node/2974). By abstracting the rule of these two principles, the algorithms can generate diverse building massing design alternatives. This allows the algorithms to be used in performance-based optimization for exploring a wide range of design alternatives guided by various performance objectives. Two case studies of passive solar energy optimization are presented to demonstrate the efficacy of the algorithm in helping architects achieve an explorative performance-based optimization process.
Conference Paper
Full-text available
This paper presents an evolutionary design toolkit for performance-based building massing design optimisation (food4rhino.com/node/2974). The toolkit is aimed to assist architects in exploring a wide range of building massing design alternatives guided by various performance objectives, thereby encouraging architects to incorporate evolutionary design optimisation for enriching design ideation at the outset of the design process. The toolkit is implemented in the Rhino-Grasshopper environment and includes components of a diversity-guided evolutionary algorithm and two pre-defined parametric models capable of generating a wide range of massing designs. The evolutionary algorithm can yield diverse design variants from the optimisation process and present more informative results with higher design differentiation. The pre-defined parametric models require minimal customisation from the architects. By using the toolkit, architects can readily explore high-performing building design with performance-based design optimisation with ease, and the coding-free optimisation workflow also streamlines the design process.
Article
Full-text available
Significant research has been undertaken focusing on the application of evolutionary algorithms for design exploration at conceptual design stages. However, standard evolutionary algorithms are typically not well-suited to supporting such optimization-based design exploration due to the lack of design diversity in the optimization result and the poor search efficiency in discovering high-performing design solutions. In order to address the two weaknesses, this paper proposes a hybrid evolutionary algorithm, called steady-stage island evolutionary algorithm (SSIEA). The implementation of SSIEA integrates an island model approach and a steady-state replacement strategy with an evolutionary algorithm. The combination aims to produce optimization results with rich design diversity while achieving significant fitness progress in a reasonable amount of time. Moreover, the use of the island model approach allows for an implicit clustering of the design population during the optimization process, which helps architects explore different alternative design directions. The performance of SSIEA is compared against other optimization algorithms using two case studies. The result shows that, in contrast to the other algorithms, SSIEA is capable of achieving a good compromise between design diversity and search efficiency. The case studies also demonstrate how SSIEA can support conceptual design exploration. For architects, the optimization results with diverse and high-performing solutions stimulate richer reflection and ideation, rendering SSIEA a helpful tool for conceptual design exploration.
Conference Paper
Full-text available
The use of daylight for the inhabitants health and comfort purposes and for the energy efficiency of buildings influences significantly the shape and outlook of urban environments. The solar envelope and solar collection surface are methods to define the massing of buildings for direct solar access requirements. They have been recently improved to be used in the design of buildings in relation to the Estonian daylight standard. Nevertheless the solar collection method can be applied only to single buildings with simple shape. The present research investigates the direct solar access performance of building clusters with multiple layouts in different urban areas in the city of Tallinn. Result show that different patterns perform in significant different ways whereas the same cluster types have the best and the least performances in all the cases.
Conference Paper
Full-text available
The integration of optimization algorithms and building performance simulation tools make it possible to carry out performa-tive design or performance-driven design, which aims to guide the design synthesis process of the simulation results to continuously improve the design. However, the associated research work of wind-related building performance is still deficient, resulting from lack of applicable interface and the time consumption. Meanwhile, in the industrial design realm, the aerodynamics or fluid-dynamics behaviour of the production under development has been vastly analysed and optimized based on the multi-discipline optimization (MDO) techniques. Owing to offering numerous built-in interface and integrated optimization algorithm, MDO application software has begun to be used in building optimization design with the complex relationship between various objectives. With the advantage of MDO tools and aimed to provide an efficient optimization approach from the perspective of architect , this paper proposes a wind-related building performance optimization design system integrating Rhinoceros and Fluent based on iSIGHT-a MDO application software. In addition, the lighting performance is considered in this research as well for implementing the multi-objective optimization. Two case studies of tall building optimization design based on varied generative approaches are introduced to investigate the effect and efficiency of this system.
Conference Paper
Full-text available
This paper proposes a performance-driven design workflow based on Total Sunlight Hours evaluation. The objective is to investigate an optimal solution of a building massing design meeting solar radiation criteria as early as in the conceptual design stage. In our paper, such a process is demonstrated through a case study on an Experimental Social Housing project. We illustrate how design constraints are encoded with the evaluation criterion, Total Sunlight Hours (TSH), through an integrated computational workflow. Alongside with such a computation-intensive process, we also experimented with the same design project using a conventional design approach. The advantages and disadvantages of using a performance-driven computational workflow over a conventional design process are discussed and presented. In particular, we examine how a performance-driven design workflow can be integrated within the iterative design process and how human designers interact with computation to investigate optimal design solutions.
Article
Full-text available
"October 2004" Thesis (Ph.D.)--The Hong Kong Polytechnic University, 2005. Includes bibliographical references.
Article
The paper considers what it means to capture design knowledge by embodying it in procedures that are expressible in a computer program, distinguishing several possible purposes for such an exercise. Following the lead of David Marr's computational approach to vision, emphasis is placed on ‘phenomenological equivalence’ — that is, first defining the functions of designing, and then specifying how people design. The paper goes on to describe design phenomena that a computational strategy of this kind would have to reproduce. All of them are integral to a view of designing as reflective conversation with the materials of a design situation, and depend on the idea of distinctive design worlds constructed by the designer. These phenomena include: the designer's seeing-moving-seeing, the construction of figures from marks on a page, the appreciation of design qualities, the evolution of design intentions in the course of the design process, the recognition of unintended consequences of move experiments, the storage and deployment of prototypes, which must be placed in transaction with the design situation, and communication across divergent design worlds. Considered as performance criteria for a phenomenologically equivalent computational designer, these phenomena are formidable and threatening. Considered as performance criteria for the construction of a computer-based design assistant, however, they may be highly evocative.
Article
The use of computers in architecture has extended beyond computer-aided design (CAD) systems to include dynamic modeling and animation. The use of this technology moves beyond the rendering and refining of ideas to the generation of form. These programs have become collaborators in the creative process.Building simulation offers the potential to provide an alternative method for generating architectural forms. Currently, building simulation tools are used for analysis only, rather than for synthesis. This paper presents an innovative design method to generate building forms utilizing dynamic outdoor conditions to optimize the forms for reducing thermal load. Integration of Computational Fluid Dynamics (CFD) and Energy Simulation (ES) has been used as a first step to more effectively analyze and model buildings and their surroundings, and new optimization technique is proposed to facilitate developing models for synthesis and to generate designs based on performance.
From Separation to Incorporation -A Full-Circle Application of Computational Approaches to Performance-Based Architectural Design
  • Y Chen
  • Y Lu
  • T Gu
  • Z Bian
  • Z Tong
Chen, Y., Lu, Y., Gu, T., Bian, Z., B, L. W., & Tong, Z. (2022). From Separation to Incorporation -A Full-Circle Application of Computational Approaches to Performance-Based Architectural Design. In Proceedings of the 2021