ArticlePDF Available

Data-driven placemaking: Public space canopy design through multi-objective optimisation considering shading, structural and social performance


Abstract and Figures

In the context of ongoing densification of cities and aging urban populations, public spaces are a crucial infrastructure to support the physical and mental wellbeing of urban residents. The design of public space furniture elements is often standardised, and not considered in relation to environmental conditions and mechanisms of social interaction. This article presents a digital workflow to generate site-specific designs for shaded public seating, considering the relationships of local public places to their surroundings. A strategy for customised and site-specific design is developed through the use of multiple software tools, employing evolutionary algorithms and multi-objective optimisation. The method is applied to a small public space canopy prototype installed within a public housing estate in Hong Kong, incorporating additional criteria to achieve a low-cost and light-weight structure. Through multiple stages of refinement and optimisation, a material, structural and social performance-driven outcome was achieved that creates a shaded space for public seating, people watching and social interaction. As part of a larger research agenda exploring architectural form-finding and environmental psychology, the project represents potential new applications in the emerging field of socially driven computational design.
Content may be subject to copyright.
Data-driven placemaking: Public space
canopy design through multi-objective
optimisation considering shading, structural
and social performance
Jeroen van Ameijde
*, Chun Yu Ma
, Garvin Goepel
Clive Kirsten
, Jeff Wong
School of Architecture, The Chinese University of Hong Kong, Hong Kong Special Administrative
Region, Hong Kong, China
Lightweight Works, Hong Kong Special Administrative Region, Hong Kong, China
APS Research Ltd., Hong Kong Special Administrative Region, Hong Kong, China
Received 17 August 2021; received in revised form 6 October 2021; accepted 21 October 2021
Public space;
Tensile membrane
Structural design;
Abstract In the context of ongoing densification of cities and aging urban populations, public
spaces are a crucial infrastructure to support the physical and mental wellbeing of urban res-
idents. The design of public space furniture elements is often standardised, and not considered
in relation to environmental conditions and mechanisms of social interaction. This article pre-
sents a digital workflow to generate site-specific designs for shaded public seating, considering
the relationships of local public places to their surroundings. A strategy for customised and
site-specific design is developed through the use of multiple software tools, employing evolu-
tionary algorithms and multi-objective optimisation. The method is applied to a small public
space canopy prototype installed within a public housing estate in Hong Kong, incorporating
additional criteria to achieve a low-cost and light-weight structure. Through multiple stages
of refinement and optimisation, a material, structural and social performance-driven outcome
was achieved that creates a shaded space for public seating, people watching and social inter-
action. As part of a larger research agenda exploring architectural form-finding and environ-
mental psychology, the project represents potential new applications in the emerging field
of socially driven computational design.
ª2021 Higher Education Press Limited Company. Publishing services by Elsevier B.V. on behalf
of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license
* Corresponding author.
E-mail address: (J. van Ameijde).
Peer review under responsibility of Southeast University.
Please cite this article as: J. van Ameijde, C.Y. Ma, G. Goepel et al., Data-driven placemaking: Public space canopy design through multi-
objective optimisation considering shading, structural and social performance, Frontiers of Architectural Research,
2095-2635/ª2021 Higher Education Press Limited Company. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd.
This is an open access article under the CC BY-NC-ND license (
Available online at
journal homepage:
Frontiers of Architectural Research xxx (xxxx) xxx
1. Introduction
As cities are becoming increasingly dense to accommodate
population growth and limit their impact on surrounding
ecosystems, the design of urban public spaces is of crucial
importance to the quality of urban environments, experi-
ences and the social activities of everyday life. Private living
spaces become less affordable and spacious, and shared
spaces around residential buildings are used as part of
important socialising, recreational and community activ-
ities. Studies have shown that the quality of life of vulnerable
social groups such as low-income families or elderly can be
significantly improved through good quality public facilities
(Saunders et al., 2014;Gou et al., 2018). In high-density cities
such as Hong Kong, public open spaces with good environ-
mental comfort and opportunities for social interaction can
contribute to active aging, social integration, and to the
physical and psychological health of elderly residents (Yung,
2016;Lau and Murie, 2017;Zheng et al., 2015).
The layout of urban space affects social dynamics, as it
“facilitates co-presence and regulates interpersonal re-
lationships” (Madanipour, 2003, p. 206). People are more
inclined to inhabit or socialise in urban spaces that have
good visibility towards their surroundings (Gehl, 1987;
Whyte, 1980), and that offer privacy and protection
(Appleton, 1975,1984). Thus, the manipulation of spatial
relationships in urban spaces through landscape or archi-
tectural elements can help create conditions conducive to
socialising (Batty, 2001;Stamps, 2005). Research has shown
that morphological aspects such as the ‘edge effect’ (Gehl,
2010), and the concept of ‘triangulation’ - where urban
elements serve as a conversation starter, can stimulate
social interaction between strangers (Whyte, 1980). Chance
encounters and casual neighbouring can lead to social
integration, attachment to place, and the formation of
supportive communities (Low, 2000;Talen, 1999).
People’s willingness to inhabit public spaces is also
strongly affected by environmental conditions, especially in
hot summer conditions in cities such as Hong Kong (Ng and
Cheng, 2012). Research has determined that shading and
wind can improve human thermal comfort (Li, et al., 2018)
and that landscape elements that improve shading and wind
flow through urban public spaces can lead to increased social
activities (Huang, et al., 2016). However, the design of
public space furniture elements is often standardised, and
their location, orientation and formal qualities are often not
considered in relation to the environmental conditions and
social interaction of the sites in which they are placed (Lee
et al., 2013;Mesthrige and Cheung, 2020).
This article focuses on site-specific shaded public
seating, designed in consideration of the relationships of
specific places to their surroundings. It presents a novel
methodology for generating customised public space can-
opies, combining architectural form-finding and structural
optimisation with a data-driven approach to analysing
urban visibility and environmental conditions. The study
involved quantifying qualitative aspects of how people
experience public space, comfort and social relations, and
thus aimed to integrate some of the technical and psy-
chological design aspects of small urban spaces into a
comprehensive computational process. A customised
methodology and workflow were developed through the
integration of multiple software tools, exploring the use of
evolutionary algorithms and multi-objective optimisation to
produce a cost effective, structurally, socially, and envi-
ronmentally performative solution.
1.1. Form-finding and performative design in
Architectural form-finding is a broad concept that has most
often been associated with the self-optimisation of mate-
rial geometries and structures in relation to structural
performance (Adriaenssens et al., 2014;Turrin et al.,
2011). The term has been associated with lightweight ar-
chitecture and tensile membrane structures developed by
Frei Otto and others (Meissner and Mo
¨ller, 2015) and the
catenary structures of Gaudi (Huerta, 2006), Isler (Chilton
and Isler. 2000) or recently the Block Research Group
(Popescu et al., 2021). However, in recent year the notion a
single, optimised outcome at the equilibrium between
geometrical, structural and material properties (Burkhardt,
2016) has been expanded to include multiple objectives
and a more complex solution space.
Form-finding processes have since long been performed
through computational tools using physics-based algorithms
(Ahlquist and Menges, 2011), which enable the combination
of the previously separate processes of form generation and
evaluation (Nguyen et al., 2020) and transform form-finding
into an interactive dynamic process (Attar et al., 2010). The
use of the term ‘form-finding’ has been expanded to
include other criteria, such as the cost optimisation of
structures (Ekici et al., 2015), wind performance (Yao
et al., 2018), construction and shading (Agirbas, 2019).
These approaches expand the notion of form-finding to-
wards ‘performative design optimisation’, which has been
defined by Oxman (2008) as an approach which integrates a
generative design process within a simulated environment
in order to evaluate the design’s performative capabilities
in relation to that environment.
In this paper, we present the novel concept of form-
finding as a generative process that considers an architec-
tural membrane structure in relation to its environment’s
specific social interaction and sun shading requirements.
The work connects existing form-finding methodologies to
urban environment and visibility analysis, combining an
integrated computational modelling and simulation envi-
ronment with algorithms for multi-objective optimisation.
In recent years, research around computational perfor-
mance optimisation has increased significantly to include a
range of optimisation workflows and performance evalua-
tion criteria, using various swarm and evolutionary opti-
misation algorithms (Ekici et al., 2019). These algorithms
can address complex architectural design challenges, using
the process of Multi-Objective Optimisation (MOO) to
explore Multiple-criteria decision-making (MCDM) problems
which contain two or more optimisation criteria that are
conflicting: if a solution is improved towards one of the
goals, other criteria are reduced (Harkouss et al., 2018).
Instead of a single solution, there are a number of “Pareto
optimal” solutions (Miettinen, 1999). The analysis of the
solution space between these objectives helps to identify
J. van Ameijde, C.Y. Ma, G. Goepel et al.
and quantify the trade-offs between and demonstrate the
consequences of prioritising each of the different criteria
(Navarro-Mateu et al., 2018).
Examples of the wide range of architectural applications
are exploring the trade-offs between architecture function-
ality, structural and environmental performance of a stadium
design (Yang et al., 2018), energy and daylight optimisation of
fac¸ade shading devices (Kirimtat, 2019), and the testing of
residential neighbourhood layouts in relation to wind and
microclimate comfort (Wu et al., 2021). Scholars have defined
the combined of MOO with the paradigm of performance
simulation evaluation as defined by Oxman (2008),as
Simulation-Based Multi-Objective Optimisation (SBMOO)
(Nguyen et al., 2014). The study presented here builds on
these examples and methods by employing a methodology
that evaluates the spatial relationships that are produced by
the architectural form, defined in relation to its spatial and
social environment. The term‘performance’ is expanded with
the notion of ‘social performance’ which is quantified through
measuring visibility across urban spaces.
1.2. Placemaking methods in urban design
Since the 1970s, the term ‘placemaking’ has been increas-
ingly used amongst the social sciences, including within the
field of urban planning (Friedmann, 2010). While there is a
range of different definitions of the term, in the context of
this article we refer to notion of placemaking as the process
of designing urban spaces to promote public life and
pedestrian activity (Walljasper, 2004). The notion refers to
‘place’ as opposed to ‘space’, indicating human attachment
to places that are attractive and suitable for their desired
modes of use (Schneekloth and Shibley, 1995). Cresswell
(2014, p.39) defines ‘place’ as “constituted through reiter-
ative social practice”, emphasizing that the value of a place
lies in its ability to stimulate events and social behaviours.
Quantitative research into the relationships between
human activities and urban morphology has been conducted
since the 1980s (Gehl, 1987;Hillier and Hanson, 1984;Whyte,
1980). A range of methods has been developed around the
theories of Space Syntax, incorporating analysis of circulation
network configuration and visibility mapping across complex
spaces (Batty, 2001;Hillier, 1996). Neurophysiological
research and behavioural theory suggest that boundary and
visibility conditions regulate social behaviour in urban envi-
ronments (Stamps, 2005). An increasing number of data-
driven studies in recent years has focused both on the evalu-
ation of predictive computational models (Askarizad and
Safari, 2020;Karimi, 2018;Loit, 2021) and on empirical
studies around the sociability of spaces (Koohsari et al., 2015;
Vroman, 2017), confirming the association between visibility
and sociability (De Stefaniand Mondada, 2018;Zakariya et al.,
2014) used asa performance measurein our study.While these
insight,relational principles and methodologies arestarting to
be implemented in generative architectural design (Goldstein
et al., 2020), their application towards the design of small-
scale urban spaces is rare. This study explores this knowl-
edge gap as part of a wider research agenda exploring envi-
ronmental psychology and data-driven design strategies for
sociable urban spaces.
Related to the sociability of public spaces is the research
into urban micro-climate and human comfort. Many studies
have been conducted around the relationships between
urban morphology and urban ventilation, establishing that
urban porosity and greening can increase comfort and
pollution dispersion (Abdollahzadeh and Biloria, 2021;Ng,
2009;Ng et al., 2011;Xue et al., 2017). While these
studies provide valuable insights around configurational
aspects of urban spaces at the larger scale, there is limited
research on the formal aspects of urban places at the
micro-scale. A number of studies around public seating
highlights the importance of wind blocking, people watch-
ing and human comfort factors including temperature, solar
radiation, and relative humidity (Zhou et al., 2015). Legge
(2020) poses that seating design features should accom-
modate psychological comfort, physical comfort, and
pleasure, and Hadavi et al. (2015) found that seating areas
which encourage socialising are most preferred. Carroll
et al. (2020) studied seating needs of elderly in residen-
tial areas and concluded that more sheltered outdoor
spaces are needed to avoid seasonal limitations to neigh-
bourhood socialising. Our study incorporates insights from
previous studies into its methodology, using the environ-
mental comfort and socialising properties of public seating
as objectives for the performance of a generative public
seating canopy design.
2. Schematic design methodology
This section describes the case study project used to
develop a computational workflow, focusing on an urban
site analysis, and the setting up and testing of a multi-
objective optimisation process that incorporated site rela-
tionship criteria as well as internal performance objectives.
2.1. Case study site analysis and selection
The location for our prototype is a rooftop public space in
between a retail centre and several high-rise residential
buildings of the Yau Oi Estate in Tuen Mun, Hong Kong, a
public housing estate opened in 1982 that accommodates a
large proportion of elderly residents. The area was analysed
using a series of algorithms including DeCodingSpaces
(Abdulmawla et al., 2017), Ladybug (Roudsari, 2013), and
customised scripts developed by the authors. The specific
site was chosen based on surveillance (‘being seen’), visi-
bility (‘seeing’), enclosedness, summer sun and connectivity
parameters (Fig. 1). Five possible locations were chosen for
quantitative analysis and comparison. The locations are
indicated in Fig. 1 and their spatial and environmental
analysis parameters are given in Table 1. All values are nor-
malised to a 0e1 range except Solar Radiation.
The location for the prototype (Site 4) was chosen due to
its good people watching opportunities and as it could
benefit the most from blocking views from surrounding
buildings to reduce the sensation of ‘being seen’. Table 1
shows that Site 4 has high levels of surveillance, medium-
high values for visibility and low values for enclosedness,
which indicate that the location has good opportunities to
survey all of its surroundings. The medium-low levels of
Frontiers of Architectural Research xxx (xxxx) xxx
connectivity (network integration) show that the location
experiences less pedestrian traffic, which makes it attrac-
tive to longer periods of resting and socialising. The lower
solar radiation value indicates a shorter duration of sunlight
exposure, which implies that a smaller size canopy can
make a significant impact on the usability of the location as
the roof covering can be adjusted to a smaller range of
different sunlight vectors.
2.2. Multi-objective optimisation for urban
As part of the ambition to develop a low-cost and
demountable public space canopy system, a tensile mem-
brane with steel support system was chosen for the project.
While traditional tensile membrane systems often use poles
and cables around the outside of the canopy, a centralised
frame principle was chosen to avoid obstacles in the public
space and simplify installation on site. While this basic ty-
pology was predetermined in relation to budgetary, regu-
latory and site management constraints, the specific
morphology of the canopy on site was subject to a wide
range of possibilities.
For our computational workflow, the traditional process
of membrane form finding, which has a single optimisation
criterion, was extended with additional fitness criteria
related to the urban setting such as shading and visibility.
Additional project goals such as minimising costs and
structural weight meant that a multi-objective optimisation
process was necessary. As part of the generative process
and multi-criteria decision analysis workflow, Evolutionary
Algorithms (EA) were used. The modelling setup was con-
structed in Rhinoceros 3D with Grasshopper, and using the
Galapagos (Rutten, 2013) and Wallacei (Makki et al., 2018)
evolutionary solver plugins. The plugins were selected from
a range of available optimisation tools available for the
Rhino3D/Grasshopper platform, including Octopus
(Vierlinger, 2012), SilverEye (Cichocka et al., 2017), and
Opossum (Wortmann, 2017) and Optimus (Cubukcuoglu
et al., 2019). Galapagos was used in the process design
stage due to its user-friendly interface and ability to
quickly visualise the flexibility in the generative model
setup, while Wallacei was used in the design research and
development stage due to its detailed analytical function-
alities, and its comprehensive tools for the selection and
visualisation of the solution space.
The evolutionary algorithms use several interconnected
steps, including defining a parametric definition of a ge-
notype (the base typology described above), the generation
of a population of individuals (phenotypes) containing
random variations and mutations, and the evaluation of
these using fitness objectives. The algorithms select the top
performing individuals to generate a new set of improved
individuals, across a predefined number of successive gen-
erations. The Galapagos software uses Simulated Evolution
and Simulated Annealing algorithms, and the Wallacei
software uses the NSGA-II algorithm (Deb et al., 2002). Both
algorithms produce progressive multi-objective evolu-
tionary processes to achieve a diverse Pareto optimal set
within a limited time span and with reasonable computer
equipment (Navarro-Mateu et al., 2018).
2.3. Multi-objective optimisation setup
To set up a generative computational model to work with the
multi-objective optimisation process, a basic typology of a
fabric canopy was constructed within the digital model of the
urban site environment. A surface was created in between
four coordinate points. The location of these points was set
as variable inputs to be tested, constrained to a maximum
boundary area on site. An internal ‘funnel’ geometry was
Fig. 1 Public space analysis mapping the levels of (aþb) surveillance, (c) enclosedness, (d) environmental comfort, (e) visibility
and (f) connectivity across the different site areas.
Table 1 Site options and measurements, with the
selected option in bold.
Site 1 Site 2 Site 3 Site 4 Site 5
Surveillance 0.15 0.25 0.40 0.45 0.30
Enclosedness 0.69 0.76 0.67 0.64 0.62
Solar Radiation (kWh/m2) 4.35 3.14 2.85 2.80 3.67
Visibility 0.79 0.75 0.73 0.76 0.81
Connectivity 0.77 0.16 0.90 0.57 0.93
J. van Ameijde, C.Y. Ma, G. Goepel et al.
created automatically by scaling and extruding the perim-
eter polyline, and removing the faces oriented towards the
open space (Fig. 2a). Using the real-time physics simulation
tool Kangaroo (Piker,2013), a double-curved minimal surface
geometry was generated (Fig. 2b). This geometry and ma-
terial system was chosen as tensile fabric structures can span
large areas with minimal use of material. The system also
allows easy installation on site, and removal and storage if
The experimental setup required the translation of spatial
social relationships andenvironmental performance goalsinto
a set of mathematical variables, relating to the parametric
definition and evaluation of our architectural geometry. The
open nature of the Rhino/Grasshopper parametric environ-
ment allows to construct such complex definitions, as it en-
ables several third-party plugins to be combined, connecting
the data flows of generating the phenotypes with tools and
rulesets to evaluate their performance.
As part of the parametric definition of the genotype, we
set up four coordinate points relative to a chosen site
boundary area, shown as 1x & 1y, 2x & 2y, 3x & 3y and 4x &
4y in Fig. 2. The values were normalised between 0 and 1.
With two decimal places set in the definition, this resulted in
a total of 808 slider values and 1.1 x 1016 possible variations.
As fitness objectives, four sets of parametric relation-
ships were set up, using a sun analysis tool, surface area
calculation and two sets of line-of-sight relationships which
were checked for connectivity or blocking. The general aim
was to optimise the location and shape of the canopy in
relation to three existing benches. Simplified volumes were
modelled to represent seated human body positions on the
benches, and their surfaces were used to analyse whether
any direct sunlight was received. Surfaces were modelled
to represent the human eye level zone of seated people on
the benches, and line-of-sight relationships were defined
with reference to important elements in the surroundings
(Fig. 3). The following Fitness Objectives (FO) were defined
to optimise the form finding process towards several per-
formance goals for the canopy:
1. FO1: Maximise shading for people seated on the three
selected benches located across from each other, thus
stimulating social interaction by inviting people to sit
face to face. Weather data for Hong Kong was imported
in the Ladybug plugin, and the analysis date was set to
September 15, 11.00 a.m.e15:00 p.m. April and
September have the lowest sun vectors and represent
the start and end of the period where shading is required
for comfortable outdoor seating.
2. FO2: Minimise the total surface area of the canopy
geometry, to reduce overall project cost in relation to
the use of materials, and wind loads which would need
to be accommodated for with the steel structure. The
definition evaluated the surface area of the final mesh
shown in Fig. 2b. A minimum surface area of 15 m
set to avoid too many unusable results.
3. FO3: Minimise ‘being seen’ by blocking views from the
shopping gallery and lower rows of residential windows
directly behind the site, to improve the feeling of pri-
vacy. People will feel less ‘overlooked’ and the solid
canopy behind them provides a feeling of strategic
safety, creating a more intimate space. The definition
counted the number of viewing lines intersecting with
the canopy geometry, shown as blue lines in Fig. 3.
4. FO4: Maintain ‘seeing’ lines of sight to the key public
spaces around the benches, to allow for the seated
persons to keep an eye on their surroundings, to increase
a feeling of safety and control. The definition counted
the number of blocked viewing lines towards the main
entrances to the public space, shown as red lines in
Fig. 3.
A current limitation of the Wallacei plugin is that while it
enables to define a larger number of fitness objectives, its
algorithm NSGA-II was developed for two objective opti-
misation problems. As in our study, FO3 and FO4 are set up
as constraints, the simulation is expected to produce cor-
rect near-optimal results. Future development of our
methodology will involve exploring other evolutionary
Fig. 2 Generative canopy typology setup, using four points within a predefined boundary area. The double curvature is generated
by the physics engine and a smoothing operation is performed on the final geometry.
Frontiers of Architectural Research xxx (xxxx) xxx
algorithms developed for multi-objective optimisation with
three or more objectives.
2.4. Evolutionary process results
The evolutionary algorithm in the Wallacei plugin was
configured to produce 50 individuals per generation (con-
taining random variations), to be evaluated and optimised
across 100 iterative generations. This produced a popula-
tion size of 5000 individuals within a computing time of
around 8.5 min. This fast simulation time allowed for many
rounds of iterative testing and improvement of the para-
metric setup, refining the evaluation mechanisms of the
performance indicators, and simulation settings to balance
between exploration, and selection and optimisation.
As the multi-objective evolutionary solver tried to opti-
mise all four fitness objectives simultaneously and in par-
allel, the resulting population set shows a wide range of
geometric variations. This variety is also produced by
properties of the algorithm that promotes wide exploration
of the solution space, to avoid premature convergence to-
wards a local optimum (Navarro-Mateu et al., 2018). Fig. 4
shows a selection of individuals from generation 0, 19, 39,
59, 79 and 99.
The software calculates and visualises the Standard
Deviation (SD) values for each generation, plotting blue
curves for advanced generations on top of red curves which
represent the earlier sets. A solid cluster of blue lines
indicate convergence towards an optimised solution space,
which occurs in our experiment with fitness objective 1, 2
and 3 (Fig. 5). Fitness objective 4 (maintaining ‘seeing’
lines towards the surroundings) does not transition to an
optimum as this is set up as a ‘binary condition’, promoting
solutions that do not block any viewing lines towards the
surroundings, which most of the more advanced solutions
can achieve well. The software also plots trendlines for the
Standard Deviations and Mean Values (Fig. 5, right side).
Gradually increasing SD values and decreasing Mean Values
indicate increasing fitness per generation, which in our
experiment was achieved gradually for fitness objective 1,
while 2 and 3 achieved this more rapidly as the parametric
setup made these criteria easier to achieve.
The multi-objective optimisation software stores the
performance data of all individuals and ranks them
accordingly. The highest-ranking individuals for each of the
fitness objectives can be retrieved, giving insights in which
specific geometry solutions fulfil those objectives most
effectively. Fig. 6 shows the phenotype, performance data
and radar charts for each fitness objective: FO1 Shading,
FO2: Minimise Surface Area, FO3: Blocking ‘being seen’,
FO4: Maintain ‘seeing’ lines. Visualisation of the sun hours
calculation on the seating area surface produced by
Ladybug are also shown. The highest-ranking phenotype for
FO1 shows a large canopy surface area covering most of the
benches so there is only a small amount sun hours received.
The second phenotype, optimised to FO2, does not show
the smallest surface area, but an area close to the mini-
mum required surface of 15 m
, as set up in the algorithm.
The third geometry shows optimal blocking of ‘being seen’
lines (FO3), resulting in a flat and narrow surface parallel to
the fac¸ade which’ window sightlines are aimed to be
blocked. The fourth phenotype is a random sample of the
solution range, as most individual satisfy the criteria for
maintaining surveillance from the benches (FO4).
While the specialised individuals give insights into how
separate fitness objectives could be achieved, we are mainly
interested in final geometry solutions that strike a balance
between trade-offs. As some of the objectives are contra-
dictory, there is no ‘perfect’ solution that satisfies all
criteria. Wallacei offers two methods for identifying solu-
tions that perform well across all objectives. The Relative
Difference option aims to find individuals that perform at an
equilibrium between objectives, while the Fitness Average
method could produce a solution that prioritises one
objective by weakening others (Navarro-Mateu et al., 2018).
In our process (Fig. 7), the Relative Difference solution was
Fig. 3 The canopy basic surface typology at three existing benches, human eye level zones of seated people on the benches
(shown in orange), and lines of sight relationships in relation to elements in the surroundings.
J. van Ameijde, C.Y. Ma, G. Goepel et al.
less suitable as the sunlight access to the seating was not
fully blocked, while the Fitness Average solution performed
best in relation to blocking sun, ‘being seen’ viewing lines
and maintaining surveillance of the surroundings. This solu-
tion, Generation 90, Individual 25 was selected for further
testing and development. Fig. 8 shows a separate simulation
with Ladybug on the surface area surrounding the canopy, to
visualise the solar radiation impact on the site on several
dates and times of the day. The analysis shows that signifi-
cant parts of the benches would remain shaded throughout
the day, and across different months.
3. Membrane design development
For the further development of the project, a multi-
disciplinary collaboration phase was started between aca-
demics and industry professionals specialising in lightweight
and tension membrane architecture. Using previous experi-
ence with the design development, fabrication and instal-
lation, the project underwent several stages of refinement
by considering the material, structural and manufacturing
complexity of the design.
The first stage of design development was to recreate
the surface geometry produced in Rhino/Grasshopper and
Kangaroo with the specialised software EASY by Technet, a
Germany-based company that supports the development of
Membrane- and Cable Net Structures. The canopy design
was remodelled by inputting the four control points
generated by the multi-objective optimisation process, and
the two lower attachment points resulting from the ‘fun-
nel’ typology encoded in the parametric definition. Fig. 9
shows the digital model of the surface, simulated as a
membrane in tension combined with steel edge cables. The
software performs its own form finding process, based on
dynamic relaxation of the surface in between the
Fig. 4 A selection of individuals from different generations of the multi-objective optimisation process.
Frontiers of Architectural Research xxx (xxxx) xxx
attachment points and in relation to pretension forces
applied at those points. Higher pretension forces and
straighter edge curves would generate large reaction forces
which the supporting structure should be able to accom-
modate. Hence, a balanced solution was found between
the final surface curvature and manageable tension forces.
The geometries produced by EASY were exported back to
Rhino3D and Grasshopper during several stages of the
development process, to check whether the required sun
shading function was still achieved as the geometry was
evolving towards a constructable solution.
The specialised software also allowed for several types of
calculations relating to further materialisation. The edge
cable forces calculation informed the specification of a steel
cable diameter of 10 mm. The calculation of stresses and
deformation of the membrane surface informed the selection
of a range of suitable fabric types. As the scale of the project
was relatively small compared to the range of projectsseen in
Fig. 5 The Standard Deviation and Mean Values plot lines and trendlines.
Fig. 6 Top ranked individuals in relation to each of separate fitness objectives.
J. van Ameijde, C.Y. Ma, G. Goepel et al.
this field, it was determined that one of the thinner types of
material, Pes PVC type 1 specification could be used.
One of the key issues to address during design develop-
ment was the simulation of wind load scenarios, which
would exert forces onto the fabric in addition to the pre-
tension forces. The first aim with tension membrane
structures is that wind-related stresses do not greatly
exceed the pretension stresses, so that wind gusts do not
result in major deformations of the fabric. The resulting
stress is kept well within the material properties. The
second aspect of considering wind load scenarios is to
determine the reaction forces that the supporting structure
would need to accommodate. Besides the pretension re-
action forces, the EASY software enables the calculation of
additional reaction forces based on a given wind speed.
As the project described here classifies as a mobile
installation, it does not need to withstand the highest wind
loads associated with typhoons under the Hong Kong
Fig. 7 Highest ranked individuals and the Parallel Coordinate Plot selected by analysing the Relative Difference and Average of
fitness ranks.
Fig. 8 Testing the selected canopy solution through solar radiation analysis (15 April, 11:00e12:00).
Frontiers of Architectural Research xxx (xxxx) xxx
regulations. These high winds typically occur in the
‘typhoon season’ which covers August and September. The
site managers of the installation agreed that the fabric
structure would be taken down during this season, or when
strong winds of over 15 m/s (30 knots, 54 km/h) would be
forecasted. Analysing weather data from the Hong Kong
observatory, we found that average wind speeds between
8.0 and 13.8 m/s occur 0.3% annually, rising to 1.6% when
only looking at the months of June/July/August. Average
wind speeds higher than 13.8 m/s occur close to 0% during
non-typhoon conditions. However, as the Observatory
documents sustained wind speeds calculated as 10-min
Mean values, the maximum wind gust speeds need to be
considered, which are instantaneous and usually higher
than the sustained wind speeds. This variation informed the
safety factors applied to the design of the steel structure,
as well as the material specifications and detailing.
4. Steel structure design
As previously described, some of the requirements and
early design decisions for the project included the capa-
bility to have a freestanding, movable and demountable
structure that does not requires alterations to the site. As
support structure for the fabric canopy, a centralised frame
principle was chosen, which would also create unob-
structed open space around to invite socialising.
4.1. Initial structure design testing
Following from the previous stage of design development,
the six anchor points and membrane geometry were defined
and considered ‘frozen’ for the next stage of structure
design. The reaction forces were known, as well as the
specific vector directions of these forces. Based on studying
precedents and previous experiences, it was decided that
the central structure would be a truss-like spatial frame,
constructed out of Circular Steel Hollow Sections (CHS). In
an initial design phase, several different design configura-
tions were developed and tested in conversation with the
team member with structural engineering expertise.
The structural layouts were modelled manually in the
Rhino3D environment and analysed with the cloud-based
structural analysis software ‘SkyCiv Structural 3D’ (Fig. 10).
In relation to the detailing and fabrication of the connec-
tions, it was determined that all structural members should
be of the same diameter. After calculating the uplift and
down-press reaction forces in EASY, these vectors were
used as input point loads in the structural analysis of the
steel frame. For structural safety analysis, the combined
stress (bending þshear þtorsion) should not exceed
220 MPa, considering a steel structure with S275 grade steel
and full-penetration welded joints. A ULS load factor of 1.4
(for wind load) was applied for the structural analysis, in
line with the Code of Practice for the Structural Use of Steel
(2011). For this temporary structure, the deflection limit
was set at 30 mm at the end of the cantilevered branches,
and different steel sections were analysed to find the most
lightweight solution.
It was found that with a CHS section of 114.3 mm
6.3 mm, most of the structure would withstand the stress
resulting from the wind load, except one of the cantilev-
ering members that was not aligned with the reaction
vector. The structural design was then iterated in order
avoid areas of disproportionate stress, and to reduce the
sizes of steel sections and overall weight of the structure if
4.2. Structure design optimisation
In the structure optimisation stage, a computational work-
flow was again used, similar in nature to the multi-objective
optimisation process described earlier in this paper. The
general aim was to allow the final composition of the spatial
frame structure to be informed by an optimisation algo-
rithm, in response to the reaction forces of the pre-designed
canopy and the wind forces in Hong Kong. Therefore, a
parametric model was set up in Grasshopper3D, based on
several three-dimensional coordinate points located be-
tween the edges of the canopy and the boundary of the
footprint. These nodes were connected by lines, to form a
coherent structural framework in a truss configuration con-
taining internal diagonals for lateral stability.
Fig. 9 Form-finding and material behaviour analysis in Technet software for membrane structures.
J. van Ameijde, C.Y. Ma, G. Goepel et al.
Within the Rhino/Grasshopper parametric platform, the
parametric structural engineering tool Karamba3D
(Preisinger, 2013), was used to evaluate the parameterized
geometric model, and connect this process to an optimisa-
tion algorithm. Similar to the initial design stage, the
structure design development required optimisation ac-
cording to multiple and conflicting fitness objectives. We
employed the Octopus plugin developed by Vierlinger
(2012), which optimises against multiple goals at the same
time according to the Pareto-principle, similar to Wallacei.
The parametric setup used several gene pool compo-
nents to position the 3D points in space. These gene pools
were connected into Octopus to allow the algorithm to run
through the spectrum of available solutions (Fig. 11). The
objectives of the optimisation were to (1) reduce the
overall mass, (2) have a minimal displacement, and (3) have
minimal intersections between beams and fabric canopy.
The pre-calculated reaction forces as described above were
integrated as a load case in Karamba3D, as well as addi-
tional load case scenarios simulating horizontal wind di-
rections across the site to optimise the steel structure
against all likely load conditions.
In addition to the Evolutionary Structural Optimisation
(ESO) process enabled by Karamba3D, the Bidirectional
Evolutionary Structural Optimisation (BESO) functionality
was used to trace the internal flow of forces through the
structure and to evaluate if the least strained elements can
be removed. After testing, it was decided to retain all truss
elements in order to be able to rely on spatial triangulation
principles for lateral stability rather than on the stiffness
and welding quality of the joints. Similar to previous pro-
jects developed with Karamba3D, it was found that fitness
objectives defined towards displacement and self-weight
are effective in the optimisation of structural designs
through genetic algorithms, and while this produces asym-
metrical or irregular layouts, these solutions can be
increasingly effective (Preisinger et al., 2011).
The outcomes of the evolutionary process showed
structure layouts in which the orientation of the upper
beams aligned with the vector orientation of the pre-
calculated reaction forces (Fig. 12). From a range of
optimised solutions across the Pareto front within
Octopus, a solution was picked based on aesthetic con-
siderations and constructability, in particular in relation to
the geometric complexity of the joints. Consistent
checking and feedback from the structural engineering
perspective, as well as final evaluation of the design so-
lution in SkyCiv Structural 3D confirmed that the refined
structure is more effective in response to the canopy’s
reaction forces and wind forces than the earlier, manually
designed solutions. The computational optimisation has
resulted in the elimination of all bending stress within the
members. As the real-world setup might still encounter
eccentricities, which could result in the buckling of the
cantilever members, a final CHS section of 88.9 mm
4 mm was prescribed.
The structural design refinement and finalisation de-
marcates the last stage of the design development pro-
cess, in which the overall canopy design, material and
structural system have been defined in relation to
external and internal requirements. Further steps have
been undertaken to develop the material detailing,
connections and manufacturing sequencing but as the
central focus of this paper is on data-driven design, these
aspects will be presented in a separate publication. The
completion of the architectural, urban and structural
design development through the use of computational
tools has resulted in a unique and affordable prototype
project that is scheduled for manufacturing and instal-
lation on site in the near future (Fig. 13a and b). At this
stage, we are able to formulate valuable insights and
conclusions regarding the difficulties and opportunities
around the computational design approach demonstrated
in this project, based on our experiences with multi-
Fig. 10 One of the earlier structural frame options, tested with the fabric tensioning reaction forces.
Frontiers of Architectural Research xxx (xxxx) xxx
objective optimisation across multiple stages of design
5. Discussion
This paper has covered several stages of a project concep-
tion and development process that mirror the typical stages
in an architectural or urban design project, but that are
novel in their integration of computational methods. We
have shown how an urban site analysis and a canopy design
process can be augmented through digital tools for visibility,
environmental and structural analysis, and attempted to
demonstrate how additional data can lead to more informed
insights about the various performative qualities of archi-
tectural structures within their contexts. The existing liter-
ature on urban morphology analysis and generative, material
performance-based design is separated into discrete do-
mains. There is a wide array of research projects focusing on
experimental architecture, yet these projects often focus on
novel materials, complex geometry and computational
mathematics without quantifying the qualities of the human
experiences resulting from these innovative spatial
Fig. 11 Some of the layout versions generated during the evolutionary structural optimisation process.
Fig. 12 The final selected configuration, showcasing the alignment of its cantilevering members with the reaction forces.
J. van Ameijde, C.Y. Ma, G. Goepel et al.
structures. In these demonstrator projects of autonomous
digital explorations, explicit relationships to their urban
contexts are often lacking. The project we have presented
employs a different use of computational tools, by focusing
on the spatial and environmental relationships between
discrete places and their wider context, and using quantified
descriptions of these relationships as design drivers. By
focusing on visibility, climate comfort and other placemaking
qualities that can promote social interaction, we deploy
computation in a human-centric design process, aimed at
producing solutions for basic human needs.
The use of evolutionary algorithms and multi-objective
optimisation processes present a powerful new toolkit for
architectural and urban design problems, incorporating
more of the complex performance criteria that play a part
in the real-world complexity of human experiences and
interactions in public spaces. In our project, the objec-
tives and range of possible outcomes was still quite limited
in relation to the possibilities of these new tools, due to
the self-imposed limitations of the chosen typology and
material system. The main goals of achieving sun shading
with a minimum amount of material, while improving local
privacy levels and maintaining opportunities to survey the
surroundings, might have been solved in a ‘manual’ digital
modelling and testing process. The value of the multi-
objective optimisation workflow, however, is not in pro-
ducing an outcome that works, but in generating all
possible solutions that could achieve satisfactory out-
comes in different ways. The process is faster, more pre-
cise and more transparent than traditional design methods
as the solution space is comprehensively explored, ana-
lysed and visualised. This helps to understand the way in
which design solutions represent a balanced solution, or
compromise, between a range of ambitions around a
project, which are often contradictory. During the canopy
design process, detailed insights in the amount of sunlight
Fig. 13 Visualisations of the final design within the site, showcasing the shading qualities and social atmosphere created around
the existing benches.
Frontiers of Architectural Research xxx (xxxx) xxx
or shadow created at the chosen site allowed to make
informed decisions about the shape and size of the
structure, understanding the trade-off between produc-
tion cost and placemaking benefits. The use of data-driven
methods allowed us to evaluate how (and how many)
people would be able to experience improved climate
comfort and social interaction opportunities in relation to
the required budget for construction. It is easy to see how
the open-ended and data-driven nature of the digital
workflow explored here can be expanded to encompass
more complex architectural or urban design problem, and
their evaluation criteria.
6. Conclusions
In this paper, we have offered a reinterpretation of the
notion of architectural ‘form-finding’, from a material
performance based concept with a single optimisation cri-
terion to a multi-faceted concept incorporating structural,
social and environmental performance. The public seating
canopy project discussed here represents an experimental
process in enhanced digital design methods using recent
computational tools, as part of a larger research agenda
around data-driven processes for socially performative
public structure and spaces. By making the principles of
social interaction and climate comfort in outdoor public
spaces explicit, spatial and mathematically quantified, it is
possible to describe the performance of architectural
structures more distinctly, and optimise them accordingly.
Multi-objective optimisation methods enable a more
transparent and insightful exploration and discussion of the
different possible solutions to a design problem, high-
lighting which design objectives could be prioritised, while
understanding their impact on other objectives. In our
public seating canopy project, a balanced solution between
fabrication costs and placemaking benefits was found,
paving the way for approval and implementation on site.
The developed workflow can be employed for a range of
future projects at different sites, using 3D data of the
surroundings to generate different site-specific solutions at
each location. The current method allows to produce other
canopy designs with the same material palette and struc-
tural principles, while the geometric typology could be
expanded to address larger sites, different social activities
and more refined environmental performance
From a conceptual point of view towards the methodol-
ogy explored here, our project is part of the wider shift to-
wards data-driven design in many disciplines, using data
analytics, machine learning and evolving modes of human-
computer interactions to enhance the quality of design
processes and their outcomes. It signals a move from tradi-
tional linear and reductionist processes led by a single
human designer to a collaborative, multi-disciplinary and
integrative process. The computational tools employed here
refer to ‘the biological paradigm’, using digital methods to
mimic natural processes of mutation, selection and evolu-
tion to generate design solutions that are suitable for con-
texts that are increasingly understood as complex and
dynamic environments. Our future work will continue to
explore the implications of these methodologies, focusing on
digital methods for analysing urban places, aiming to capture
their social and environmental complexities and address
these through to generative design, producing meaningful
architectural and urban design solutions that contribute to
the improved well-being of individuals and communities.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
This project was supported by a grant from Design Trust,
Hong Kong. The authors would like to thank Poonam Sar-
desai, Karen Lee, and People’s Place Hong Kong for their
Abdollahzadeh, N., Biloria, N., 2021. Outdoor thermal comfort:
analyzing the impact of urban configurations on the thermal
performance of street canyons in the humid subtropical climate
of Sydney. Frontiers of Architectural Research 10 (2), 394e409.
Abdulmawla, A., Bielik, M., Bus, P., Chang, M., Dennemark, M.,
Fuchkina, E., Miao, Y., Knecht, K., Koenig, R., Schneider, S.,
2017. DeCodingSpaces Toolbox for Grasshopper: Computational
Analysis and Generation of Street Network, Plots and Buildings.
Adriaenssens, S., Rhode-Barbarigos, L., Kilian, A., Baverel, O.,
Charpentier, V., Horner, M., Buzatu, D., 2014. Dialectic form
finding of passive and adaptive shading enclosures. Energies 7
(8), 5201e5220.
Agirbas, A., 2019. Fac¸ade form-finding with swarm intelligence.
Autom. ConStruct. 99, 140e151.
Ahlquist, S., Menges, A., 2011. Behavior-based Computational
Design Methodologies: Integrative Processes for Force Defined
Material Structures. ACADIA 11: Integration through Computa-
tion [Proceedings of the 31st Annual Conference of the Associ-
ation for Computer Aided Design in Architecture (ACADIA)].
Banff (Alberta), pp. 82e89.
Appleton, J., 1975. The Experience of Landscape. John Wiley &
Sons, London and New York.
Appleton, J., 1984. Prospects and refuges re-visited. Landsc. J. 3
(2), 91e103.
Askarizad, R., Safari, H., 2020. Investigating the role of semi-open
spaces on the sociability of public libraries using space syntax
(Case Studies: sunrise Mountain and Desert Broom Libraries,
Arizona, USA). Ain Shams Engineering Journal 11 (1), 253e264.
Attar, R., Aish, R., Stam, J., Brinsmead, D., Tessier, A., Glueck, M.,
Khan, A., 2010. Embedded rationality: a unified simulation
framework for interactive form finding. Int. J. Architect.
Comput. 8 (4), 399e418.
Batty, M., 2001. Exploring isovist fields: space and shape in archi-
tectural and urban morphology. Environ. Plann. Plann. Des. 28,
Burkhardt, B., 2016. Natural structures-the research of Frei Otto in
natural sciences. Int. J. Space Struct. 31 (1), 9e15.
Carroll, S., Jespersen, A., Troelsen, J., 2020. Going along with
older people: exploring age-friendly neighbourhood design
through their lens. J. Hous. Built Environ. 35 (2), 555e572.
Chilton, J., Isler, H., 2000. Heinz Isler. Thomas Telford.
Cichocka, J.M., Migalska, A., Browne, W.N., Rodriguez, E., 2017.
SILVEREYE ethe implementation of particle swarm
J. van Ameijde, C.Y. Ma, G. Goepel et al.
optimization algorithm in a design optimization tool. In:
, G., O
¨zkar, M., Gu
¨l, L., Gu
¨rer, E. (Eds.), Computer-Aided
Architectural Design. Future Trajectories. CAADFutures 2017.
Communications in Computer and Information Science, vol.
724. Springer, Singapore.
Cresswell, T., 2014. Place: A Short Introduction. Wiley-Blackwell.
Cubukcuoglu, C., Ekici, B., Tasgetiren, M.F., Sariyildiz, S., 2019.
OPTIMUS: self-adaptive differential evolution with ensemble of
mutation strategies for grasshopper algorithmic modeling. Al-
gorithms 12 (7), 141.
De Stefani, E., Mondada, L., 2018. Encounters in public space: how
acquainted versus unacquainted persons establish social
and spatial arrangements. Res. Lang. Soc. Interact. 51 (3),
Deb, K., Pratap, A., Agarwal, S., Meyarivan, T.A.M.T., 2002. A fast
and elitist multiobjective genetic algorithm: NSGA-II. IEEE
Trans. Evol. Comput. 6 (2), 182e197.
Ekici, B., Cubukcuoglu, C., Turrin, M., Sariyildiz, I., 2019. Perfor-
mative computational architecture using swarm and evolu-
tionary optimisation: a review. Build. Environ. 147, 356e371.
Ekici, B., Kutucu, S., Sarıyıldız,
I.S., Tas
getiren, M.F., 2015. May).
Addressing the high-rise form finding problem by evolutionary
computation. In: 2015 IEEE Congress on Evolutionary Compu-
tation (CEC). IEEE, pp. 2253e2260.
Friedmann, J., 2010. Place and place-making in cities: a global
perspective. Plann. Theor. Pract. 11 (2), 149e165.
Gehl, J., 1987. Life between Buildings: Using Public Space. Van
Nostrand Reinhold, New York.
Gehl, J., 2010. Cities for People. Island Press, Washington.
Goldstein, R., Breslav, S., Walmsley, K., Khan, A., 2020. Space-
Analysis: a tool for pathfinding, visibility, and acoustics analyses
in generative design workflows. In: Proceedings of the 11th
Annual Symposium on Simulation for Architecture and Urban
Design (SimAUD ’20), pp. 1e8, 5.
Gou, Z., Xie, X., Lu, Y., Khoshbakht, M., 2018. Quality of life (QoL)
survey in Hong Kong: understanding the importance of housing
environment and needs of residents from different housing
sectors. Int. J. Environ. Res. Publ. Health 15 (2), 219.
Hadavi, S., Kaplan, R., Hunter, M., 2015. Environmental affor-
dances: a practical approach for design of nearby outdoor
settings in urban residential areas. Landsc. Urban Plann. 134,
Harkouss, F., Fardoun, F., Biwole, P.H., 2018. Multi-objective
optimization methodology for net zero energy buildings. Jour-
nal of Building Engineering 16, 57e71.
Hillier, B., 1996. Space Is the Machine: a Configurational Theory of
Architecture. Space Syntax, London.
Hillier, B., Hanson, J., 1984. The Social Logic of Space. Cambridge
University Press, Cambridge; New York.
Huang, J., Zhou, C., Zhuo, Y., Xu, L.,Jiang, Y., 2016. Outdoorthermal
environments and activities in open space: an experiment study in
humid subtropical climates. Build. Environ. 103, 238e249.
Huerta, S., 2006. Structural design in the work of Gaudi. Architect.
Sci. Rev. 49 (4), 324e339.
Karimi, K., 2018. Space syntax: consolidation and transformation of
an urban research field. J. Urban Des. 23 (1), 1e4.
Kirimtat, A., Krejcar, O., Ekici, B., Tasgetiren, M.F., 2019. Multi-
objective energy and daylight optimization of amorphous
shading devices in buildings. Sol. Energy 185, 100e111.
Koohsari, M.J., Mavoa, S., Villanueva, K., Sugiyama, T.,
Badland, H., Kaczynski, A.T.,, et al., 2015. Public open space,
physical activity, urban design and public health: concepts,
methods and research agenda. Health Place 33, 75e82.
Lau, K.Y., Murie, A., 2017. Residualisation and resilience: public
housing in Hong Kong. Hous. Stud. 32 (3), 271e295.
Lee, Y.B., Chan, L.D., Tang, M.X., 2013. Park Seating Furniture
Design in Hong Kong: a Case Study of Inclusive Design and its
Relation to User Interaction.
Legge, K., 2020. Public seatingesmall important places. In: The
Routledge Handbook of Placemaking. Routledge, pp. 439e448.
Li, J., Niu, J., Mak, C.M., Huang, T., Xie, Y., 2018. Assessment of
outdoor thermal comfort in Hong Kong based on the individual
desirability and acceptability of sun and wind conditions. Build.
Environ. 145, 50e61.
Loit, D., 2021. Children’s Access to Playgrounds: A Space Syntax
Assessment of the Urban Integration of Built Playgrounds and
Homeplayground Accessibility (Stockholm).
Low, S., 2000. On the Plaza: the Politics of Public Space and Cul-
ture. University of Texas Press, Austin, TX.
Madanipour, A., 2003. Public and Private Spaces of the City.
Routledge, London.
Makki, M., Showkatbakhsh, M., Song, Y., 2018. Wallacei: an
Evolutionary and Analytic Engine for Grasshopper 3D. Retrieved
from. (Accessed 12 July 2020).
Meissner, I, Mo
¨ller, E, 2015. Frei Otto: forschen, bauen, inspirieren
/ a life of research, construction and inspiration. DETAIL,
Mesthrige, J.W., Cheung, S.L., 2020. Critical evaluation of ‘ageing
in place’in redeveloped public rental housing estates in Hong
Kong. Ageing Soc. 40 (9), 2006e2039.
Miettinen, K., 1999. Nonlinear Multiobjective Optimization. Kluwer
Academic Publishers, Boston, MA.
Navarro-Mateu, D., Makki, M., Cocho-Bermejo, A., 2018. Urban-
tissue optimization through evolutionary computation. Mathe-
matics 6 (10), 189.
Ng, E. (Ed.), 2009. Designing High-Density Cities: for Social and
Environmental Sustainability. Routledge.
Ng, E., Cheng, V., 2012. Urban human thermal comfort in hot and
humid Hong Kong. Energy Build. 55, 51e65.
Ng, E., Yuan, C., Chen, L., Ren, C., Fung, J.C., 2011. Improving the
wind environment in high-density cities by understanding urban
morphology and surface roughness: a study in Hong Kong.
Landsc. Urban Plann. 101 (1), 59e74.
Nguyen, A.T., Reiter, S., Rigo, P., 2014. A review on simulation-
based optimization methods applied to building performance
analysis. Appl. Energy 113, 1043e1058.
Nguyen, T.N., Hien, T.D., Nguyen-Thoi, T., Lee, J., 2020. A unified
adaptive approach for membrane structures: form finding and
large deflection isogeometric analysis. Comput. Methods Appl.
Mech. Eng. 369, 113239.
Oxman, R., 2008. Performance-based design: current practices and
research issues. Int. J. Architect. Comput. 6 (1), 1e17.
Piker, D., 2013. Kangaroo: form finding with computational physics.
Architect. Des 83 (2), 136e137.
Popescu, M., Rippmann, M., Liew, A., Reiter, L., Flatt, R.J., Van
Mele, T., Block, P., 2021. June). Structural design, digital
fabrication and construction of the cable-net and knitted
formwork of the KnitCandela concrete shell. In: Structures, vol.
31. Elsevier, pp. 1287e1299.
Preisinger, C., 2013. Linking structure and parametric geometry.
Architect. Des 83, 110e113.
Preisinger, C., Vierlinger, R., Hofmann, A., Bollinger, K., 2011.
Evolutionary Structural Optimization Revisited. IASS Morphology.
Roudsari, M.S., Pak, M., Smith, A., 2013, August. Ladybug: a
parametric environmental plugin for grasshopper to help de-
signers create an environmentally-conscious design. In: Pro-
ceedings of the 13th International IBPSA Conference Held in
Lyon. France Aug, pp. 3128e3135.
Rutten, D., 2013. Galapagos: on the logic and limitations of generic
solvers. Architect. Des 83 (2), 132e135.
Saunders, P., Wong, H., Wong, W.P., 2014. Deprivation and poverty
in Hong Kong. Soc. Pol. Adm. 48 (5), 556e575.
Schneekloth, L., Shibley, R., 1995. Placemaking: the Art and
Practice of Building Communities. Wiley, New York.
Stamps, A., 2005. Isovists, enclosure, and permeability theory.
Environ. Plann. Plann. Des. 32, 735e762.
Frontiers of Architectural Research xxx (xxxx) xxx
Talen, E., 1999. Sense of community and neighbourhood form: an
assessment of the social doctrine of new urbanism. Urban Stud.
36, 1361e1379.
Turrin, M., Von Buelow, P., Stouffs, R., 2011. Design explorations of
performance driven geometry in architectural design using
parametric modeling and genetic algorithms. Adv. Eng. Inf. 25
(4), 656e675.
Vierlinger, R., 2012. Octopus. BollingerþGrohmann Engineers.
Vroman, L., Lagrange, T., 2017. Human movement in Public spaces:
the use and development of motion-oriented design strategies.
Des. J. 20 (Suppl. 1), S3252eS3261.
Walljasper, J., 2004. Pedestrian Power. Utne. May 1, 2004.
Whyte, W.H., 1980. The Social Life of Small Urban Spaces. Con-
servation Foundation, Washington, D.C.
Wortmann, T., 2017. Opossum-introducing and evaluating a model-
based optimization tool for grasshopper. In: Janssen, P., Loh, P.,
Raonic, A., Schnabel, M.A. (Eds.), Protocols, Flows, and Glitches -
Proceedings of the 22nd CAADRIA Conference, Xi’an Jiaotong-
Liverpool University, Suzhou, China, 5-8 April 2017, pp. 283e292.
Wu, Y., Zhan, Q., Quan, S.J., Fan, Y., Yang, Y., 2021. A surrogate-
assisted optimization framework for microclimate-sensitive
urban design practice. Build. Environ. 195, 107661.
Xue, F., Gou, Z., Lau, S.S.Y., 2017. Green open space in high-dense
Asian cities: site configurations, microclimates and users’ per-
ceptions. Sustainable cities and society 34, 114e125.
Yang, D., Ren, S., Turrin, M., Sariyildiz, S., Sun, Y., 2018. Multi-
disciplinary and multi-objective optimization problem re-
formulation in computational design exploration: a case of
conceptual sports building design. Autom. ConStruct. 92,
Yao, J.W., Lin, Y.Q., Zheng, J.Y., Yuan, P.F., 2018. A “dynamic
form-finding” approach to environmental-performance building
design. International Journal of High-rise Buildings 7 (2),
Yung, E.H.K., Conejos, S., Chan, E.H.W., 2016. Social needs of the
elderly and active aging in public open spaces in urban renewal.
Cities 52, 114e122.
Zakariya, K., Harun, N.Z., Mansor, M., 2014. Spatial characteristics
of urban square and sociability: a review of the City Square,
Melbourne. Procedia-Social and Behavioral Sciences 153,
Zheng, W., Shen, G., Wang, Hao, Lombardi, p., 2015. Critical issues
in spatial distribution of public housing estates and their im-
plications on urban renewal in Hong Kong. Smart and Sustain-
able Built Environment 4 (2), 172e187.
Zhou, D., Zou, Y., Jiang, J., 2015. A research into the design
strategies for public seating in a windy environment. Interna-
tional Conference on Arts, Design and Contemporary Education
(ICADCE 2015) 380e383.
J. van Ameijde, C.Y. Ma, G. Goepel et al.
... General Secretary Xi Jinping said at the Intermediate Rural Work Conference held at the end of 2013: "China should be strong, and agriculture and animal husbandry should be strong; China should be beautiful, and the countryside should be beautiful; for China to be rich, farmers must be rich [1]." ...
Full-text available
At this stage, the most important content of rural construction includes the construction of rural public space, which is related to the environment of all rural areas, the happiness of villagers, and the inheritance and development of rural areas. Especially under the guidance of the rural revitalization strategy, the necessity of public space environment construction has become increasingly prominent, becoming an irreplaceable core content in rural construction. Therefore, we must closely focus on the development needs of rural areas at this stage, dig deep into the development of the public space environment, find corresponding design strategies, and maximize the completion of the real construction of beautiful rural areas. Rural public spaces have experienced the pain of new urbanization and rural intelligence. The most important public spaces in traditional agricultural societies, such as deep-water wells, trees, and sundecks, will slowly decline and fade. In the future, public space in rural areas must function in urban areas. Rural public spaces with complex roles and diverse types of public activities are the main orientations for the transformation of rural spaces in the future.
Modern urban architecture is innovative, involves integrated solar technologies to increase sustainability, by providing efficient energy solutions, and involves stakeholders from multiple domains, with specific concerns. Such development, however, leads to complex interrelationships among disciplines and relies on models to simulate different aspects. To address these, the parametric design approach is widely adopted, enabling the generation of several urban designs in the early stage, and evaluating these across several objectives. This provides possibilities for creating collaborative, multi-disciplinary workflows for energy projects at different scales. This paper provides a critical overview of the possibilities offered by the parametric design, to increase the sustainability of advanced buildings, by putting a focus on solar energy. This paper analyses the parametric design across five categories: design scope, application, technology, objective, and implementation method. No review article currently highlights all these aspects to observe overall research in the field. The results indicate that the parametric design is prominent on a single building-scale, while windows in building-scale and building forms in district-scale models are widely manipulated parameters. Most models are limited to BIPV, shader, and solar façade technologies, daylight and solar power applications, and energy and comfort-related objectives, the majority of which are implemented using dedicated parametric modeling tools. Further ongoing research trends emphasize how parametric design transforms the field and creates opportunities for designers working in building energy research. The analytical results help to find current barriers, and future initiatives to implement the parametric design at a wider spread for solar integration into buildings.
Full-text available
The quality of outdoor space is becoming increasingly important with the growing rate of urbanization. Visual, acoustic, and thermal balance degradation are all negative impacts associated with outdoor comfort in dense urban fabrics. Urban morphology thus needs assessment and optimization to ensure favorable outdoor thermal comfort (OTC). This study aims to evaluate the thermal performance of streets in residential zones of Liverpool, NSW, Australia, and tries to improve their comfort index (Physiological Equivalent Temperature) to reveal optimum urban configurations. This evaluation is done by investigating the following urban design factors affecting OTC using computational simulation techniques: street orientation, aspect ratio, building typology, and surface coverage. Our findings reveal that street canyon orientation is the most influential factor (46.42%), followed by aspect ratio (30.59%). Among the influential meteorological parameters (air temperature, wind speed, humidity and solar radiation), wind velocity had the most significant impact on the thermal comfort of the outdoor spaces in this coastal region, which typically experiences intense airflow. The results of our analysis can be utilized by multiple stakeholders, allowing them to understand and extract the most vital design factors which contextually influence the thermal comfort of outdoor spaces. Outdoor thermal comfort has a direct effect on the health and wellbeing of occupants of outdoor spaces.
Full-text available
Nowadays, in spite of the positive efforts made in recent years, active public libraries are still faced with challenges in attracting and retaining their users through the provision of up-to-date services to meet their scientific, cultural and recreational requirements. Accordingly, the need to create spaces with the approach of sociability in public libraries seems to be necessary. In this way, the Space Syntax method, as a predictor of the social aspects of architectural spaces, has been used in order to evaluate the role of semi-open spaces on the sociability of public libraries in the existing and hypothetical models using Depthmap. Finally, the findings of the two simulated models were examined and compared with each other using SPSS and the independent sample t-test. The findings from the Sunrise Mountain Library analysis indicated that in the existing model, the integration level was 20.49. However, after eliminating the semi-open spaces, the level of integration decreased to 15.48. Also, the findings obtained from the Desert Broom Library indicated that in the existing model, the integration level was 10.10. However, in the hypothetical model, its degree of integration changed to 8.15. Results show that semi-open spaces have a positive effect on the sociability of public libraries.
Full-text available
Neighbourhoods are extremely important to older people, as this is where a great deal of their everyday life is spent and where social interaction happens. This is particularly the case in deprived neighbourhoods, where people with limited economic resources or physical limitations find it challenging to venture outside the neighbourhood. A growing body of research suggests studying age-friendly neighbourhoods from a bottom-up approach which takes the diversity of the age group into account. This paper aims to investigate how the go-along method can serve to co-construct knowledge about age-friendly neighbourhood design in a deprived neighbourhood of Copenhagen with a diverse group of older people. Sixteen go-along interviews were carried out with older people aged 59–90. The participants took on an expert role in their own everyday life and guided the researcher through the physical and social environments of their neighbourhood. The go-alongs were documented with a GoPro camera. The data was analysed using situational analysis and was grouped into thematic categories. Our findings conclude that social interaction is the overall motivator for going outdoors and that dimensions of pavements, the seating hierarchy, the purpose of lawns, sheltered spaces and ‘unauthorised’ places are all neighbourhood design elements that matter in this regard. The findings suggest to consider age-friendly details as the starting point for social interaction, to target the appropriate kind of age-friendly programs and to enhance empowerment through physical spaces. The go-along interview as a research method holds the potential for empowering older people and appreciating their diversity.
Full-text available
Most of the architectural design problems are basically real-parameter optimization problems. So, any type of evolutionary and swarm algorithms can be used in this field. However, there is a little attention on using optimization methods within the computer aided design (CAD) programs. In this paper, we present Optimus, which is a new optimization tool for grasshopper algorithmic modeling in Rhinoceros CAD software. Optimus implements self-adaptive differential evolution algorithm with ensemble of mutation strategies (jEDE). We made an experiment using standard test problems in the literature and some of the test problems proposed in IEEE CEC 2005. We reported minimum, maximum, average, standard deviations and number of function evaluations of five replications for each function. Experimental results on the benchmark suite showed that Optimus (jEDE) outperforms other optimization tools, namely Galapagos (genetic algorithm), SilverEye (particle swarm optimization), and Opossum (RbfOpt) by finding better results for 19 out of 20 problems. For only one function, Galapagos presented slightly better result than Optimus. Ultimately, we presented an architectural design problem and compared the tools for testing Optimus in the design domain. We reported minimum, maximum, average and number of function evaluations of one replication for each tool. Galapagos and Silvereye presented infeasible results, whereas Optimus and Opossum found feasible solutions. However, Optimus discovered a much better fitness result than Opossum. As a conclusion, we discuss advantages and limitations of Optimus in comparison to other tools. The target audience of this paper is frequent users of parametric design modelling e.g., architects, engineers, designers. The main contribution of this paper is summarized as follows. Optimus showed that near-optimal solutions of architectural design problems can be improved by testing different types of algorithms with respect to no-free lunch theorem. Moreover, Optimus facilitates implementing different type of algorithms due to its modular system.
Technical Report
Full-text available
Most of the current approaches to implement structural topology optimization use highly sophisticated computational environments and necessitate scripting skills. The aim of our research is a user-friendly implementation of Finite Element (FE)-and Bidirectional Evolutionary Structural Optimisation (BESO)-components for intuitive usage within the parametric 3D modelling software package Rhino and Grasshopper. The approach investigated in this paper consists of combining genetic algorithms (GA) with Bidirectional Evolutionary Structural Optimization (BESO) applied to beam structures: The GA adjusts the parameters for BESO which acts on regular and irregular ground structures. In case of irregular structures, a GA also breeds the base geometry. The product of self weight and internal elastic energy reflects structural effectiveness and serves as the fitness criteria that drives the genetic algorithm.
Simulations can often benefit microclimate-sensitive urban design by offering insightful abstractions of stochastic urban system behaviors, yet many of them are difficult to use and generally consume significant computational resources. We adapt an advanced surrogate-assisted evolutionary optimization algorithm instead of other empirical multi-objective evolutionary algorithms commonly used to search for optimal design alternatives to confront the challenge. Moreover, a parametric design module is hybridized with this surrogate-assisted evolutionary optimization algorithm to create a working scheme for mathematically modeling microclimate-sensitive urban design problems. This seven-step scheme is tested using a hypothetical case, a spatial planning problem in a residential block, to search for design proposals that maximize project development profits and facilitate the needed creation of a comfortable wind environment. Moreover, by utilizing three optimization solvers, we obtained a near-optimal site plan with a wind velocity ratio of 0.36, a wind velocity Gini index of 0.31, and a gross profit of 4.05 × 10⁸ RMB. Also, the case study results show that the proposed optimization framework, which consists of a global surrogate (additive Gaussian process model) and a local surrogate (gradient boosted regression trees model), converges faster and provides better optimal solutions to a high-dimensional design problem compared to the algorithm which only uses a single surrogate. Built with a flexible structure, we believe the proposed framework can address the emerging demands for a wide range of microclimate-sensitive design tasks, especially those with costly simulations and small experimental datasets.
Different from large deflection analyses of plates and shells, for analysis of an isotropic membrane structure a form finding procedure is first performed to find the new geometry of membrane. With this new geometry, large deflection analysis is conducted and the membrane works in a pure membrane state. In the conventionally numerical approaches for analysis of membranes, form finding and large deflection analysis are performed separately and quite complexly by different techniques. To overcome this difficulty, this paper proposes a unified adaptive approach (UAA) for analysis of membranes in the framework of isogeometric analysis (IGA). In UAA, both form finding and large deflection analysis are simultaneously and simply performed based on a modified Riks (M-R) nonlinear solver. As another advantage of UAA, it can be applied to various types of structures without any limitation or modification. For membranes, the proposed approach performs both form finding and large deflection analysis naturally. It is convenient that UAA automatically performs large deflection analysis without form finding when plates or shells are considered. Based on the adaptation of UAA, nonlinear responses of plates/shells and membranes are first obtained at the same time in this paper. Especially, a new type of equilibrium path of membranes is first proposed via UAA. The path includes two parts: one is obtained by form finding and another is from large deflection analysis. High accuracy and efficiency of the proposed approach and the present formulation are verified via some benchmark problems. Effects of slenderness ratio, curvature and boundary condition on form finding and nonlinear responses of membranes are rigorously investigated. Some new equilibrium paths of square and cylindrical membranes are first provided.
This paper describes the structural design, digital fabrication and construction of KnitCandela, a free-form, concrete waffle shell with KnitCrete, a falsework-less formwork approach using a custom prefabricated knitted textile as multi-functional, structural shuttering layer and a form-found cable net as the main load-bearing formwork. The digitally designed and fabricated textile provided integrated features for inserting and guiding elements such as cables and inflatables that helped shape the sophisticated mould. With a total weight of only 55 kg, the 50 m² formwork was easy and compact to transport. On site, the formwork was tensioned into a timber and steel rig, the pockets were inflated, and then coated with a thin layer of custom-developed, fast-setting cement paste. This paste served as a first stiffening layer for the textile, minimising the formwork’s deformations during further concrete application. Fibre-reinforced concrete was manually applied onto the formwork to realise a 3 cm-thick shell with 4 cm-deep rib stiffeners. The novel approach, for the first time applied at architectural scale in this project, enables the building of bespoke, doubly-curved geometries in concrete, with a fast construction time and minimal waste, while also reducing the cost and labour of manufacturing complex parts.