Structures 43 (2022) 1334–1341

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

GA based design automation and optimization of earthquake resisting CFS

structures in a BIM environment
Nouredine Bourahla a, *, Salim Tafraout a, Yasser Bourahla b, Adelane Sereir-El-Hirtsi a,
Abdeslam Skoudarli a
a
LGSDS Laboratory, Civil Engineering Department, National Polytechnic School, Algiers, Algeria
b
Univ. Grenoble Alpes, Inria, CNRS, Grenoble INP, LIG, F-38000 Grenoble, France

A R T I C L E I N F O A B S T R A C T

Keywords: The trend in systemization and automation of the design process in the architecture, engineering and con­
Cold-formed steel struction (AEC) industry is relatively slow because of the diversity of partners, the specificity of each project and
Design optimization the iterative nature of the design. Cold-formed steel structures which are widely used as a construction system,
Design automation
lends itself well to industrialisation. Taking advantage of the building information modelling (BIM) features, this
Genetic algorithm
paper, develop an optimization procedure to help automatizing the interface between the architecture and the
Building information modelling
Shear wall panel structural design of this type of framing. The concept relies on using a genetic algorithm to generate optimized
shear wall panels (SWP) layouts at each level of an early architectural IFC model of a multi-storey CFS building in
BIM environment. The generated layouts fulfil the load bearing continuity through the building height and are in
compliance with main earthquake resisting structural criteria as per codes of practice. The performance and the
robustness of the algorithm were demonstrated using different CFS prototype buildings. The procedure achieved
very promising results and more importantly, the algorithm can accommodate more structural code-based
criteria or constructional provisions as constraints or as controlling parameters of the multi-objective function.
In practice, the proposed technique can be easily developed and implemented in a BIM platform to assist both the
architect and the engineer to design optimized CFS buildings to resist earthquake loading.

1. Introduction CFS constructions contributed towards the application of the seismic

performance-based design method [14] and provided valuable infor­
The construction industry has achieved several milestones in the mation on seismic behaviour of braced sheathing panels as well as on
industrialization, automation and optimization course [1–3]. Yet, the how non-structural elements of CFS buildings can affect the seismic
architecture, engineering and construction (AEC) sector is still facing response of this type of building [15]. Nonetheless, the overall behav­
many challenges in this trend, because of the diversity of the partners iour can reasonably be portrayed using simple design assumptions as
and disciplines involved in the AEC process and the iterative nature of demonstrated by case studies [16] and code based seismic design of CFS
the design procedure [4]. buildings [17–20]. On the other hand, a variety of methods have
The use of building information modelling (BIM) has greatly effectively been applied to CFS structures optimization, which resulted
enhanced the information sharing and the inter-disciplinary check in economical and efficient construction solutions [21,22]. Artificial
during the conceptual phase, the drafting, manufacturing and further intelligence (AI) techniques, especially metaheuristics algorithms, have
stages [5–8]. emerged as means of choice for resolving the optimization and auto­
The process of cold-formed steel (CFS) construction is reputed to be mation of the design process [23,24]. In an attempt to enhance the
well suited for industrialization [9]. It can significantly benefit from the flexural load-bearing capacities of individual CFS members, a swarm
automation and optimization of its design and fabrication procedure optimization (PSO) method was used to optimize their sections [25]. At
[10,11]. Over the past decades, tremendous research effort on individ­ a multi-components system level, a design optimization, including the
ual members [12,13] and full scale physical and numerical modelling of spacing, pitch of the frame and section sizes of CFS elements,

* Corresponding author.
E-mail address: [email protected] (N. Bourahla).

https://doi.org/10.1016/j.istruc.2022.07.041
Received 5 April 2022; Received in revised form 30 June 2022; Accepted 15 July 2022
Available online 30 July 2022
2352-0124/© 2022 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
N. Bourahla et al. Structures 43 (2022) 1334–1341

demonstrated the efficiency of a genetic algorithm to converge to a best “stud” members, which fit into unlipped channel “track” sections at the
portal frame with the minimum cost, whilst fulfilling the design criteria top and bottom into the foundations. The lateral forces, induced by the
[26,27]. While CFS members and structural sub-systems optimization seismic or wind loads, are directly transmitted to the foundation through
has been widely investigated [28], very limited or no work has been the shear wall panels (SWP) or bracings assuming that the floors are
done on automating the conceptual design and optimizing the overall rigid diaphragms in-plane and the capacity design criteria are satisfied
structures with SWPs [29]. To partially fill this literature gap, this paper [16]. In this study, the SWP is the primary lateral load resisting system;
applies a genetic algorithm to derive from an architectural BIM model of it is composed of CFS C-shaped framing members (studs and tracks)
a multi-storey CFS building, optimized shear wall panels (SWP) layouts attached to steel/wood sheathing using screw fasteners. The SWP is
at each floor level in compliance with the load path continuity and other designed to enforce inelastic behaviour in the connection zone between
structural and seismic requirements. The Algerian seismic code the CFS frame and the sheathing up to the failure mechanism. For the
RPA99v2013 [30] and Eurocode8 [31] are referenced in this instance, purpose of the preliminary design, a nominal strength of the SWP will be
still, other criteria of any code of practice can also be integrated. This used to determine the required length/floor. Cross-section design checks
research work is part of an ongoing project, which initiated with the will be performed in a subsequent phase of the structural design.
work of Tafraout et al. [32], with the objective of developing an intel­
ligent protocol to automatically generate optimized structural designs
3.1. Floor joists direction
for architectural configurations extracted from BIM models in IFC (In­
dustry Foundation Classes) format. The proposed approach enables to
In this type of structures, vertical loads are distributed on slabs and
explore wider range of potential structural solutions at an early stage of
transmitted to wall elements (studs) through joists. The spans and the
the architectural design.
wall openings are the two main criteria that control the orientation of
the joists in a given slab. Some other constructional conditions can in­
2. Concept of intelligent structural design
fluence also the choice of the joists’ direction, however for simplification
reason, only the first two criteria are considered in this study for rect­
Defining an initial optimized layout (conceptual design) at an early
angular slab cells.
stage of the architectural design is of paramount importance, but in most
When the length in one direction is dominant such as:
cases is not straightforward [33]. Hence, an intelligent structural design
concept based on a genetic algorithm [32] is used in this study to Rxy < 0.50 orRxy > 2 for Lmax < 5.0 m (1)
generate optimized structural layouts of CFS shear wall panels layouts.
Starting from a BIM model of an architectural preliminary design, the Rxy < 0.67or Rxy > 1.5 forLmax > 5.0m (2)
genetic algorithm defines first the entire geometric locations where Then the direction of joists is chosen to be parallel to the shortest
shear wall panels (SWP) can be positioned at each level with respect to span Lmin = min (Lx, Ly).
the architectural partitions and the load path continuity. Then, the key If the two conditions are not met, which corresponds to geometric
structural criteria are expressed as constraints and scores to formulate a configurations with comparable dimensions Lx and Ly, the optimal joists’
multi-objective function to be optimized using a genetic algorithm. In direction is assigned parallel to the direction of walls that have less
the present study, the weights of the objective function are varied in openings min (Lxt, Lyt).
order to investigate the effect of each parameter on the performance of Where:
the resulting optimized layouts.
Lx
Rxy = (3)
3. Structural constraints OF CFS buildings Ly

A cold-formed steel (CFS) structure as illustrated in Fig. 1 refers to a Lmax = max(Lx , Ly ) (4)
light-frame construction where the vertical and horizontal structural
∑
nox
elements are primarily formed by a system of repetitive framing mem­ Lxt = 2Lx − Loxi (5)
bers, in which the slab and walls are rigidly connected to form a i=1
monolithic system providing a structural redundancy that enhances the
global stability of the building.
noy
∑
Lyt = 2Ly − Loyi (6)
The vertical loads are supported by floor joists or ceiling rafters and i=1
transmitted to wall members which are typically vertical lipped channel
nox and noy are the number of openings in x and y direction

Floor joists

Shear wall panel

Studs

Fig. 1. Typical CFS structures.

1335
N. Bourahla et al. Structures 43 (2022) 1334–1341

respectively (Fig. 2). 3.5. SWP in plan distribution

A joist length limit can be used as a gravity load resistance criterion
to be compared to the maximum joist length that has been determined During earthquakes, torsional vibrations can induce further damage
according to the above conditions. to buildings. In order to reduce the building’s torsional response, the
SWPs should be distributed in such a way to minimize the torsional
3.2. Length of individual SWP eccentricities e0x and e0y in the x and y axes respectively.
Considering rigid floors, SWPs having the same material properties,
The shear wall panels are composed of C-shaped framing members sheathing and storey height, would have their corresponding stiffness
(chord studs, studs and tracks) attached to sheathing using screw fas­ proportional to their length [36,37]. In these conditions, the coordinates
teners. The minimal length of a wall is conditioned by the distance be­ of the stiffness centre can be written as:
tween two consequent studs Lwmin = 60 cm (24 in.) and the height to ∑ ∑
lwyi xi lwxi yi
width (h/w) ratio of a SWP limited to 4:1. xs = ∑ andys = ∑ (11)
lwyi lwxi
h
Lwmin = max(60cm, ) (7) Where (xi, yi) are the co-ordinates of the centre of the shear wall i
4
relative to axes × and y respectively, lwxi and lwyi being the length of a
h is the storey height. given shear wall i in the × and y directions, respectively; subscript i
An optimal SWP length is taken as : Lwopt = h, as suggested in [34] (8) varies from 1 to the total number of SWPs in each direction.
A general rule in earthquake resisting design is to increase the
torsional stiffness that can be achieved by placing the SWPs the furthest
3.3. Total length of SWPs
possible from the centre of the building. Assuming that the stiffness of a
SWP is proportional to its length, the torsional stiffness can be expressed
In each direction, shear walls must be able to resist lateral loads
by the torsional radius along the × (or y) direction of analysis as:
provoked by the wind or the seismic action. For the design of CFS
∑ 2
structures, the seismic equivalent static load is used: (xi lwyi + y2i lwxi )
ry2 = ∑ (12)
V = Cs W (9) lwxi
∑
Where V is the total seismic lateral load in a direction, W is the (xi2 lwyi + y2i lwxi )
rx2 = ∑ (13)
effective seismic weight and the response coefficient Cs is determined as lwyi
per any seismic code [35]. The total seismic load is distributed along the
The above parameters can be used to express the multi-objective
height of the building resulting in lateral forces Fi applied at each level.
function associated to in-plan regularity criteria that need to be opti­
The minimal quantity of shear walls, expressed here as a total length
mized in CFS buildings where the equivalent static method is commonly
of SWPs (Lwtmin) at each storey level along each direction, is determined
applied.
by dividing the level shear lateral load (Vi) by the nominal SWP strength
per unit length Rnominal .
4. Performance parameters of the Multi-Objective function
Vi
Lwtmin = (10)
Rnominal To progress the individuals of a population towards an optimized
solution, each individual (solution) is assessed using a multi-objective
3.4. SWP vertical distribution and load path continuity function composed of the following criteria:

A continuous load path imposes that shear walls must be continuous • Total length of SWP: The quantity of shear walls to be placed in a
from the upper part of the wall down to the foundation in order not to CFS structure is required to satisfy both the safety and economy
disrupt vertical loads transmission. Progressive stiffness reduction from criteria. The total length of SWP must be greater than the minimum
bottom to top is permitted. This constraint prevents vertical disconti­ quantity of walls Lwtmin needed in each direction, but should permit
nuity of SWPs but it permits to stop the SWPs from bottom to any storey to fulfil some other criteria such as symmetry and torsional stiffness.
level. The following grading expression has been used effectively to opti­
mize the total length in each direction:

Lx If Lwt < Lwtmin : score = (

Lwt 3
) (14)
Lwtmin

Lwt
Else score = 1 − ( ) (15)
Lwtmin
∑
Where Lwt = lwi is the total length of SWPs along one direction.
Opening The score is the sum of scores along the × and y directions.
Loy1 Loy2
• Torsional radius: For total SWP lengths Lwtx and Lwty along the × and
y direction respectively, the torsional radius needs to be maximized
Ly in order to optimize the torsional stiffness of the building. This
parameter allows the algorithm to generate more external shear
walls. The torsional radius is normalized with respect to the theo­
Loy3 retical maximum rxmax and rymax as:
( Dx )2 ( D )2
Lox1 2
rxmax = 2
× Lwty + 2y × Lwtx
(16)
Lwty

Fig. 2. Typical rectangular cell with openings.

1336
N. Bourahla et al. Structures 43 (2022) 1334–1341

( Dx )2 ( D )2
× Lwty + 2y × Lwtx Input an IFC format model of an architectural
2
rymax = 2
(17) configuration from a BIM platform
Lwtx

rx
Scorex2 = (18) Optimization parameters: CFS
rxmax weights of the objective function Structural
constraints
ry
Scorey2 = (19)
rymax
Define potential positions of
The score including the two directions is given by: SWPs
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∑ 2 ̅
√
rx ry √ (xi lwyi + y2i lwxi )
Score2 = + = √2 ( )2 ( D )2 (20)
rxmax rymax Dx
× Lwty + y × Lwtx
2 2
Define floor joists orientation

• Floor torsional eccentricity (Symmetry): The torsional eccentricities Generate randomly an initial
e0x and e0y, are normalized to 0 for a perfectly symmetric solution population of structural layouts
and 1 refers to a given limit eccentricity or an eccentricity equal to
the maximum allowed by the regulations or the in-plan dimension of
the slab along the considered direction. Evaluate Fitness
e0x
Scorex3 = (21)
emax
Selection
e0y
Scorey3 = (22)
emax

e0x + e0y Best fit Yes Limit of

Score3 = (23) layout
emax iterations
The fitness value of a layout solution is defined as a weighted linear No
combination of the different score criteria.
Crossover and mutation
∑
3
fv = wi scorei (24)
i=1 Fig. 3. Automation and optimization algorithm flowchart.

The weights are determined empirically according to the criteria

importance in evaluating a solution. Using the same weights equal to building models are selected among conventionally designed CFS
unity gives in general good results. However, the weights can be structures with different geometrical dimensions in terms of surface,
changed for particular cases if required. height, opening locations and interior architectural partitions. Model 1
consists of a two-story medical office building with a one-storey annex
framed with cold-formed steel, Model 2 is a two-storey building with one
5. Genetic algorithm optimization
dominant dimension in-plan and Model 3 is also a two-storey building
with a compact form in-plan.
The optimization is carried out using a genetic algorithm with the
All the structural members used in the design of the structures are
defined constraints and objective function. The input to the GA is an
cold-formed steel material only, except for the strengthening frames
architectural configuration in IFC format which defines the slab con­
with a hot rolled material member that is not considered in the design of
tours, the position of partitions at each floor. The GA outputs, after a
the structures treated by the optimization algorithm. In order to make it
limited number of iterations, an optimized CFS SWPs layout which best
possible for the algorithm to determine automatically the bearing di­
satisfies the constraints and optimizes the objective function. The
rection of the joists while taking into account the dimensions of the
flowchart in Fig. 3 illustrates the main tasks of the GA as defined below:
rooms and the openings, all rooms of the structures are considered as
rectangular cells. It should be noted that, in CFS low rise buildings,
• Random generation of first population: The algorithm generates an
generally the minimum total length of shear walls required to ensure the
initial population of 80 individuals with different SWP lengths
stability of the building does not change drastically from a level to
positioned randomly within the architecture partitions.
another. For simplicity purpose, the minimum total SWP length is kept
• Selection: The selection of the layouts for the next population is
unchanged for all levels.
simply based on fitness ranking of all the individuals after adding
those of the cross-over and mutation.
7. Determination of the floor joists direction
• Crossover function: This operation consists of generating a new
layout by combining SWPs from two or more existing configurations,
The algorithm scans first the rooms of the building, subdivides the
this ensures that the new solution inherits features of both parents
floor into rectangles and identifies the spans and openings in order to
and keeps it under the same constraints (Fig. 4-a).
propose the orientation of the joists in accordance with the constraints
• Mutation function: Consists in either inserting a new SWP, removing,
given in section 3.1. This first step ensures that the architectural layout
shortening or extending an existing one to introduce new possibil­
satisfies the preliminary CFS constraints (maximum span and openings)
ities of SWP lengths (Fig. 4-b).
for the gravity load transmission using an adequate joists orientation.
For the considered layouts, the floor joists orientations generated by the
6. Description of CFS building samples algorithm are in most cases identical to the conventional joist directions
plan, except for two specific spaces – as shown if Fig. 6. For area 1, it is
The performance of the GA in generating optimized layouts is eval­ caused by errors encountered while parsing the architectural
uated using three different architectural configurations (Fig. 5). The

1337
N. Bourahla et al. Structures 43 (2022) 1334–1341

Extending length Inserting new wall

Shortening length Removing wall

(a) Crossover function (b) Mutation function of walls

Fig. 4. Crossover and mutation functions.

Overall 3D View Typical floor 3D View Plan View

M odel 1

M odel 2

M odel 3

Fig. 5. Architectural IFC Models.

configuration from IFC files. As for the area 2, the algorithm generates 8. Sensibility analysis of the weights of the Multi-Objective
the directions strictly according to the pre-defined conditions function
(maximum allowed span, openings…etc) and does not take into
consideration any constructional conditions, such as harmonization, The three architectural IFC models are used to investigate the effect
which in this case, the designer kept the same orientation for the adja­ of the variation of the weights of the multi-objective function on the
cent areas. Constructional provisions can be integrated in the algorithm performance of the GA in generating optimized structural layouts.
for more advanced versions.

1338
N. Bourahla et al. Structures 43 (2022) 1334–1341

1
2

(a) GA generated orientation (b) designer orientatio n

Fig. 6. Floor joists orientation.

8.1. Effect of the eccentricity parameter

Table 2
Variation of torsional radius scaling factor’s value.
The eccentricity parameter is expected to control the degree of the
torsional eccentricity for a given floor in a building. To determine the Model Scaling Factor Percentage of peripheral walls Eccentricity
influence of this parameter, the weight has been assigned four different e0x e0y
values 0, 1, 2 and 3, while keeping all other scaling factors equal to 1 0 27% 0.6% 0.3%
unity. The obtained eccentricities of the floors of the optimized layouts, 1 18% 1.8% 0.9%
varied according to the values of the weight of the eccentricity param­ 2 21% 2.8% 1.9%
eter (Table 1). For a weight value equal to 0 (deactivated parameter), the 3 24% 4.3% 3.5%
2 0 26% 1.6% 0.2%
percentage of the eccentricity ratio ranges from 1.3% to 24.4% and
1 28% 3.0% 4.0%
decreases to a range of 0.1%–0.3% for a scaling factor equal to 3 for all 2 30% 0.4% 2.0%
the configurations. 3 30% 4.0% 7.4%
3 0 49% 0.1% 0.0%
8.2. Effect of the torsional radius parameter 1 41% 1.3% 2.0%
2 46% 2.5% 0.9%
3 58% 3.0% 1.0%
The torsional radius parameter is intended to control the SWP dis­
tribution to optimize the torsional stiffness of the building. The weight of
this parameter has been assigned four different values 0, 1, 2 and 3.
Depending on the architectural configuration, the results show that
the SWP quantity in exterior walls tend to increase with higher values of
the weight, with a side effect on the eccentricity as shown in Table 2.
Note that the scaling factor of the eccentricity parameter is fixed to 1.

8.3. Effect of the minimum required SWP length weight

The minimum required SWP length is a key parameter to ensure a

good performance of the structure under a given level of a seismic
loading. This parameter is crucial to the process, since without defining
the minimum required length of the SWP, the resulting layouts will be
either cluttered (full) of SWPs or having less than the required quantity.
An example of a layout (Fig. 7.) resulting from a multi-objective function
with disabled score1 (weight of the minimum required SWP length equal
to zero), shows that only 5 m length SWP has been laid along the lon­ Fig. 7. Minimum SWP length parameter disabled.
gitudinal direction representing 17% of the needed SWPs in that direc­
tion. On the contrary, the layout comprises more than twice the needed 8.4. Robustness and stability of the GA
length of SWPs in the transverse direction. Using a weight value equal to
unity yield in all cases a total length very close to the specified minimum A GA optimized structural layout for a given architectural configu­
SWP length. ration is not a unique solution. The GA can provide different variants
which are consistent with the predefined criteria and the weights of the
multi-objective function. In this section the stability and robustness of
the algorithm is investigated. To this end, three optimization runs are
carried out for the same architectural model with the same constraints
Table 1 and weights. The layout solutions (Fig. 8) are compared in terms of the
Eccentricity values relative to different weight values. eccentricity ratios, the torsional radius rx and ry, and the ratio of the total
Weights wi Model 1 Model 2 Model 3 SWP over the corresponding minimum length in each direction. As can
be seen in Table 3, the variation of the characteristics of the variants is
e0x e0y e0x e0y e0x e0y
insignificant in most cases. Although, some small variation in the ec­
0 3.6% 24.4% 7.1% 1.9% 11.0% 1.3% centricity ratios appeared along the X and Y axis, but in absolute term,
1 1.8% 0.9% 3.0% 4.0% 0.9% 1.3%
2 0.4% 0.8% 3.5% 3.0% 0.3% 0.3%
the eccentricities are all less than 1%. Similarly, the torsional radius rx
3 0.1% 0.1% 0.2% 0.3% 0.1% 0.2% and ry are very stable in the three variants. The total SWP length and its

1339
N. Bourahla et al. Structures 43 (2022) 1334–1341

Variant 1 Variant 2 Variant 3

Fig. 8. GA layout variants for the same architectural configuration.

algorithm in order to enhance the practicality of the optimized solutions.

Table 3
However, in terms of the performance parameters, the eccentricities e0x
Structural characteristics of different GA optimized variants of an architectural
and e0y of the layouts generated by the GA are lower than those of the
configuration.
conventionally designed (Table 4). The SWP minimum lengths Lwtx , Lwty
Variant e0x e0y rx ry Lwtx/Lwtmin Lwty/Lwtmin
are almost identical in all cases and the torsional radii rx and ry are very
1 0.04% 0.31% 17.33 17.31 1.0 1.0 close or slightly higher in the layouts generated by the GA.
2 0.22% 0.97% 17.53 17.52 1.0 1.0
3 0.10% 0.25% 17.20 17.18 1.0 1.0
10. Conclusion

distribution are also unchanging. Within the limits of this investigation, The construction process of cold-formed steel structures is well
the GA proves to be stable and robust in delivering optimized variants suited to industrialization. In an effort to promote the concept of sys­
with constant characteristics fulfilling the predefined criteria. tematization of the interface between the disciplines, especially the
architectural and structural designs, this paper applies a genetic algo­
9. Performance of the optimized layouts generated by the GA rithm to generate optimized cold-formed SWP layouts compliant with a
given early architectural design of a multi-story building extracted from
The weights or the scaling factors values of the different score a BIM platform in IFC format. The most controlling seismic code-based
criteria used to define the objective function and the constraints of the criteria for CFS overall conceptual design are identified and integrated
genetic algorithm are kept equal to unity. as constraints and performance parameters of a multi-objective function
The output layouts of the GA respect totally the architecture interior in a genetic algorithm. The latter defines first the floor joists orientation
partitions at all levels. The SWPs were positioned in agreement with all and all the possible geometric positions of SWPs. Then it generates an
the openings (doors). The load path condition of the walls has also been initial population of layouts with different SWP lengths placed randomly
satisfied for all the configurations. within the architecture partitions. The initial population evolves
A comparison between the GA generated layouts and the corre­ through series of crossover and mutations to a best solution in a limited
sponding conventional SWP plans – manually designed and calculated number of iterations. The performance of the algorithm in controlling
by expert structural engineers – are presented in Fig. 9. some features of the optimized layouts is investigated by varying the
The degree of similarity between the generated SWP distribution weights of the multi-objective function for different CFS prototype
plans and the reference layouts (conventional design) depends on the buildings and comparing the best GA generated solutions with conven­
architectural configuration and the density of the SWPs. The differences tional designs. The efficiency and robustness of the proposed protocol
in SWP lengths, where most SWPs are shorter or subdivided into mul­ provides a springboard for further research work to improve the design
tiple walls, because the algorithm tends to optimize the total SWP length automation of CFS structures.
by assigning a length close to the optimal length (Lwopt) to the majority
of generated SWPs. This ensures an optimal resistance and behaviour for Declaration of Competing Interest
individual SWP, but may not be an ideal solution in practice. Future
development can introduce some construction provisions into the The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence

Fig. 9. Conventional and GA generated layouts.

1340
N. Bourahla et al. Structures 43 (2022) 1334–1341

Table 4
Performance parameters of the GA and the conventionally designed layouts.
GA generated layouts Conventionally designed layouts

e0 (%) Lwt r e0 (%) Lwt r

x y x y x y x y x y x y

1 0.30 0.34 29.3 33.5 22.9 24.5 0.75 2.63 29.2 33.5 20.3 21.8
2 0.32 1.00 43.0 46.5 20.9 21.7 2.47 4.87 43.2 46.5 18.2 18.9
3 0.42 1.30 41.6 35.0 17.8 17.8 0.39 3.46 41.8 35.2 18.9 16.9

the work reported in this paper. [18] Kechidi S, Bourahla N, Castro JM. Seismic design procedure for cold-formed steel
sheathed shear wall frames: Proposal and evaluation. J Constr Steel Res 2017;128:
219–32. https://doi.org/10.1016/j.jcsr.2016.08.018.
Acknowledgements [19] NIST, Seismic design of cold-formed steel lateral load resisting systems: A guide for
practicing engineers, GCR 16-917-38, NEHRP Seismic Design Technical Brief No.
The financial support of the Ministry of higher education MESRS in 12, produced by ATC and CUREE for the NIST, Gaithersburg, MD. https://doi.org/
10.6028/NIST.GCR.16-917-38.
Algeria (Grant CNEPRU J0400420140001) for conducting this study is [20] Boudreault FA, Blais C, Rogers CA. Seismic force modification factors for light-
greatly acknowledged. gauge steel-frame - wood structural panel shear walls. Can J Civ Eng 2007;34(1):
56–65.
[21] J. Becque, Optimization of cold-formed steel products: achievements, challenges
References and opportunities, NORDIC STEEL 2019, The 14th Nordic Steel Construction
Conference, September 18–20, 2019, Copenhagen, Denmark.
[1] He R, Li M, Gan VJL, Ma J. BIM-enabled Computerized Design and Digital [22] Parastesh H, Hajirasouliha I, Taji H, Sabbagh AB. Shape optimization of cold-
Fabrication of Industrialized Buildings: A Case Study. J Clean Prod 2021;278(1). formed steel beam-columns with practical and manufacturing constraints. J Constr
https://doi.org/10.1016/j.jclepro.2020.123505. Steel Res 2019;155:249–59. https://doi.org/10.1016/j.jcsr.2018.12.031.
[2] Chen Q, García de Soto B, Adey BT. Construction automation: Research areas, [23] Carbas S, Artar M. Optimum design of cold-formed steel frames via five novel
industry concerns and suggestions for advancement. Autom Constr 2018;94:22–38. nature-inspired metaheuristic algorithms under consideration of seismic loading.
https://doi.org/10.1016/j.autcon.2018.05.028. Structures 2021;33:4011–30. https://doi.org/10.1016/j.istruc.2021.06.096.
[3] Pătrăucean V, Armeni I, Nahangi M, Yeung J, Brilakis I, Haas C. State of research in [24] J. Leng, Optimization techniques for structural design of cold-formed steel
automatic as-built modelling. Adv Eng Inf 2015;29(2):162–71. https://doi.org/ structures. In: Cheng Yu (Ed), Recent Trends in Cold-Formed Steel Construction,
10.1016/j.aei.2015.01.00. Woodhead Publishing, 2016, pp 129-151, doi:10.1016/b978-0-08-100160-
[4] Khan A, Sepasgozar S, Liu T, Yu R. Integration of BIM and Immersive Technologies 8.00006-2.
for AEC: A Scientometric-SWOT Analysis and Critical Content Review. Buildings [25] Perampalam G, Poologanathan K, Gunalan S, Nagaratnam B, Tsavdaridis KD, Ye J.
2021;11(3):126. https://doi.org/10.3390/buildings11030126. Structural behaviour of optimised cold-formed steel beams. Steel Constr 2020;13.
[5] Chi H, Wang X, Jiao Y. BIM-Enabled Structural Design: Impacts and Future https://doi.org/10.1002/stco.201900024.
Developments in Structural Modelling, Analysis and Optimisation Processes. Arch [26] Phan DT, Lim JBP, Sha W, Siew CYM, Tanyimboh TT, Issa HK, et al. Design
Computat Methods Eng 2015;22:135–51. https://doi.org/10.1007/s11831-014- optimization of cold-formed steel portal frames taking into account the effect of
9127-7. building topology. Eng Optim 2013;45(4):415–33.
[6] Panteli C, Kylilia A, Fokaides PA. Building information modelling applications in [27] Phan DT, Mojtabaei SM, Hajirasouliha I, Ye J, Lim JBP. Coupled element and
smart buildings: From design to commissioning and beyond A critical review. structural level optimisation framework for cold-formed steel frames. J Constr Steel
J Clean Prod 2020;265:121766. https://doi.org/10.1016/j.jclepro.2020.121766. Res 2020;168:105867. https://doi.org/10.1016/j.jcsr.2019.105867.
[7] Mangal M, Cheng JCP. Automated optimization of steel reinforcement in RC [28] Mojtabaei SM, Hajirasouliha I, Ye J. Optimisation of cold-formed steel beams for
building frames using building information modeling and hybrid genetic best seismic performance in bolted moment connections. J Constr Steel Res 2021;
algorithm. Autom Constr 2018;90:39–57. https://doi.org/10.1016/j. 181:106621. https://doi.org/10.1016/j.jcsr.2021.106621.
autcon.2018.01.013. [29] Liang H, Roy K, Fang Z, Lim JBP. A Critical Review on Optimization of Cold-
[8] Hamidavi T, Abrishami S, Hosseini MR. Towards intelligent structural design of Formed Steel Members for Better Structural and Thermal Performances. Buildings
buildings: A BIM-based solution. J of Build Eng 2020;32:101685. 2022;12:34. https://doi.org/10.3390/buildings12010034.
[9] Yin X, Liu H, Chen Y, Al-Hussein M. Building information modelling for off-site [30] RPA99v2003 (Algerian seismic regulations), DTR-B.C. 248, National Centre of
construction: review and future directions. Autom Constr 2019;101:72–91. https:// Earthquake Engineering (CGS) 2003.
doi.org/10.1016/j.autcon.2019.01.010. [31] CEN (European Committee for Standardization).. Design of structures for
[10] Malik N, Ahmad R, Al-Hussein M. Generation of safe tool-paths for automatic earthquake resistance, Part 1: General rules, seismic actions and rules for buildings,
manufacturing of light gauge steel panels in residential construction. Autom Constr Eurocode08, ENV 1998-1-1. Brussels: Belgium; 2004.
2019;98:46–60. https://doi.org/10.1016/j.autcon.2018.11.023. [32] Tafraout S, Bourahla N, Bourahla Y, Mebarki A. Automatic structural design of RC
[11] Manrique JD, Al-Hussein M, Bouferguene A, Nasseri R. Automated generation of wall-slab buildings using a genetic algorithm with application in BIM environment.
shop drawings in residential construction. Autom Constr 2015;55:15–24. https:// Autom in Constr 2019;106:102901.
doi.org/10.1016/j.autcon.2015.03.004. [33] Sharafi P, Teh LH, Hadi MNS. Conceptual design optimization of rectilinear
[12] Xingxing W, Jihong Y. Reversed cyclic performance of cold-formed steel shear building frames: A knapsack problem approach. Eng Optim 2015;47(10):1303–23.
walls with reinforced end studs. J Constr Steel Res 2015;113:28–42. https://doi. [35] Bourahla N. Equivalent static analysis of structures subjected to seismic actions. In:
org/10.1016/j.jcsr.2015.05.012. Beer M, Kougioumtzoglou IA, Patelli E, Au I-K, editors. Encyclopedia of Earthquake
[13] Peck Q, Rogers N, Serrette R. Cold-Formed Steel Framed Gypsum Shear Walls: In- Engineering. Berlin, Heidelberg: Springer Berlin Heidelberg; 2021. p. 1–13.
Plane Response. J Struct Eng 2012;138(7):932–41. [36] Nie S, Zhou T, Zhang Y, Zhang B, Wang S. Investigation on the Design Method of
[14] Schafer BW, Ayhan D, Leng J, Liu P, Padilla-Llano D, Peterman KD, et al. Seismic Shear Strength and Lateral Stiffness of the Cold-Formed Steel Shear Wall. Math
Response and Engineering of Cold-formed Steel Framed Buildings. Structures 2016; Prob Eng 2020;2020:1–13.
8:197–212. [37] Xu L, Martinez J. Strength and stiffness determination of shear wall panels in cold-
[15] Fiorino L, Macillo V, Landolfo R. Shake table tests of a full-scale two-story formed steel framing. Thin-Walled Struct 2006;44(10):1084–95. https://doi.org/
sheathing-braced cold-formed steel building. Eng Struct 2017;151:633–47. https:// 10.1016/j.tws.2006.10.002.
doi.org/10.1016/j.engstruct.2017.08.056.
[16] Fiorino L, Iuorio O, Landolfo R. Designing CFS structures: The new school bfs in
naples. Thin-Walled Struct 2014;78:37–47. https://doi.org/10.1016/j. Further reading
tws.2013.12.008.
[17] Fiorino L, Campiche A, Shakeel S, Landolfo R. Seismic design rules for lightweight [34] Rouaz I, Bourahla N, Kechidi S. Numerical evaluation of shear strength of CFS
steel shear walls with steel sheet sheathing in the 2nd-generation Eurocodes. shear wall panels for different height-to-width ratios. J of Mater and Eng Struct
J Constr Steel Res 2021;187:106951. 2018;vol5 No4:399–417. http://revue.ummto.dz/index.php/JMES/article/view/
1888.

1341