12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal

he Royal Society A…

򯌤 Open Access

Review articles

Mathematics meets the fashion industry on path to product

innovation and sustainability
M. Magri and P. Ciarletta 
Published: 21 June 2023 https://doi.org/10.1098/rspa.2022.0715

Abstract
Mathematics remains an endless source of inspiration for both the conception and the fabrication
of clothes and fashion products. Over the last century, the prominent role of mathematics in the
fashion industry has evolved together with the progress in computational resources and
fabrication technologies. Nowadays, mathematics not only provides useful theoretical and
computational tools to assist design and fabrication, but also acts as a catalyst of creativeness.
This survey discusses the latest advances on mathematical models and methods in the fashion
industry, restricting our attention to clothing design and multi-scale modelling, and their
applications for digital manufacturing of textiles and garments. Particular focus is given to the
usage of mathematics to drive product innovation and to reduce the environmental impact of the PDF
fashion industry in order to meet the Sustainable Development Goals of the United Nations. Help

1. Introduction
Mathematics has always been an endless source of inspiration for both the conception and the
fabrication of clothes and fashion products. For example, the systematic ability of geometry to
quantify the arrangements and the interrelationship of shapes and patterns in space has
especially proved crucial to complement the creative process in clothing. In the eighteenth
century, outstanding mathematicians, such as Euler [1] and Monge [2], solved the challenge to
draw useful mappings of curved surfaces leading to fundamental results and techniques of
differential geometry. These mathematical principles immediately found applications in fashion
related problems, with the pioneering contributions of Chebyshev [3] and Bianchi [4], analysing


how to design woven fabrics in order to wrap onto a generic surface. Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 1/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

These seminal works have established the prominent role of mathematics in the fashion industry,
that has evolved together with the progress in computational resources and fabrication
technologies. Nowadays, mathematics not only provides useful theoretical and computational
tools to assist design and fabrication, but also acts as a catalyst of new conceptual ideas.

In the following, we provide a state of the art and a survey on the latest advances in mathematical
models and methods in the fashion industry, restricting our attention to clothing design and its
applications for digital manufacturing of textiles and garments. Particular focus will be given to the
usage of mathematics to drive product innovation and to reduce the environmental impact of the
fashion industry [5], in order to meet the Sustainable Development Goals of the United Nations.1

This article is organized into three sections. Section 2 deals with the mathematics of clothing
accounting for purely geometric models of cloth motion, as well as for mechanics-based
approaches in the context of virtual simulation of garments and multi-scale modelling of textiles.
Section 3 deals with mathematics-inspired design of fabrics and clothes. This part includes a
discussion of clothes style and patterns followed by some examples of iconic fashion collections
inspired by mathematical concepts. Finally, §4 examines the role played by physics-based
mathematical modelling for digital manufacturing of garments and smart textiles.

2. Mathematics of clothing
Garments and fashion accessories can be seen as surfaces wrapped onto the three-dimensional
domain occupied by the human body. The wrapping morphology depends on a number of
physical interactions including gravitational, elastic, contact and self-collision forces. Thus,
clothing is intimately related to the mathematical branches of differential geometry and continuum
mechanics. In this section, we will review various mathematical theories and simulation tools that
have been developed to predict the mechanical behaviour of clothes and fashion items.

(a) Clothes as developable surfaces

Most textiles and fabrics can easily deform out-of-plane but are constrained not to deform in-
plane. In mathematical terms, they can be effectively modelled as developable surfaces, i.e. they
can be flattened onto a plane without any stretching. Under this hypothesis, it is possible to
decompose any garment as the union of patches that can be unfolded onto a plane without PDF
distortion. This methodology is generally employed in fashion design software to define the Help
geometry of sewing patterns from three-dimensional garment models [6]. Since the mathematical
characterization of a developable surface is not unique, different modelling approaches have
been proposed. A first approach is based on ruled surfaces. The latter are generated by sweeping
a straight line, called ruling, along a curve, called directrix. Mathematically, a ruled surface is
defined through the parametrization x(v, t) = α(t) + v w(t), where α(t) is the directrix while w(t)
is the direction of the rulings. Such a surface is developable if the condition (w × α ) ⋅ w = 0 ′ ′

holds [7]. Examples of ruled developable surfaces, referred to as warped, are cylinders, cones,
tangent surfaces and union of pieces of these three. Using this definition, continuous developable
surfaces can be generated by approximating a directrix using Bézier or B-spline polynomials [8,9].
In such a scenario, a developability condition can be imposed analytically resulting in a nonlinear
problem for the control points of the boundary curves. Interacting modelling tools for generating


smooth developables were also proposed based on this approach [10]. However, despite Backtheir
to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 2/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

mathematical elegance, the resulting nonlinearities have generally prevented continuous models
from being used in industrial design.

More efficient approaches rely on discrete approximations of developables, generally using

polygonal meshes. A widespread methodology is based on constructing discrete ruled surfaces
by boundary triangulation [11]. Operationally, the boundary curve is approximated with line
segments and the surface with triangles whose vertices are the boundary points. Differently from
continuous methods, the developability constraint results in a geometric condition that can be
readily imposed. Mitani & Suzuki [12] demonstrated that discrete developables based on
boundary triangulation can be used for approximating complex surfaces by the union of
developable strips. Rose et al. [13] extended the use of boundary triangulation to automatic
generation of developables on arbitrary sketched boundaries. This algorithm was employed for
the development of a interactive design tool for garments. Discrete developables can also be
constructed using different methods. A considerable number of works focused on transforming a
given (non-developable) polygonal surface into a developable one with minimum distortion from
the original shape. Such a transformation is guided by a particular selection of the mathematical
condition for developability. For instance, developable approximation of discrete surfaces can be
obtained by modifying the position of the vertices such that a quasi-vanishing Gaussian curvature
is obtained [14,15]. Alternative approaches apply the condition that a developable surface has a
one-dimensional Gauss image, i.e it is a network of curves in the Gauss sphere [16]. Based on this
definition, the recent work of Binninger et al. [17] proposed a novel algorithm based on iterative
deformation of a given surface by ‘thinning’ its Gauss image.

Other methods for generating discrete developables were proposed in the realm of the so-called
discrete differential geometry [18]. Here, a discrete counterpart of a smooth parametrization is
obtained by representing the surface through quadrilateral meshes, also known as nets. In this
way, local isometry reduces to geometrical constraints on the discrete mesh. Similarly to smooth
manifolds, a given surface can be represented by different discrete nets. Therefore, various nets
have been proposed due to special geometrical properties. For instance, Liu et al. [19] modelled
developable surfaces via conjugate nets consisting of rulings and their conjugate directions.
Furthermore, developable manifolds have been modelled using discrete orthogonal geodesic
nets [20], discrete parallel geodesic coordinates [21] and discrete isometric mappings [22]. These
methods are generally more flexible than the ones based on conjugated developable nets since
PDF
rulings are not directly encoded into the mesh.
Help

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 3/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Figure 1. Plot of the solution of the clothing problem (2.2) for a paraboloidal obstacle of given height: resulting shape of the initially
plane Chebyshev net (a) and deformation of the textile applied to the obstacle (b). Capital letters A, B and C mark corresponding
points in the undeformed and deformed textile.

Download figure Open in new tab Download PowerPoint

(b) Models for deformable clothes

Modelling garments as isotropic inextensible manifolds might be too restrictive for various fabrics.
Some level of deformation is indeed possible depending on the specific fabric morphology. For
instance, knitted fabrics readily adapt to a given surface as their fibres can undergo significant
stretching. Local deformation in woven fabrics is instead granted by the alteration of the angle
between warp and weft threads. The first attempt to model cloth deformation dates back to the
work of Chebyshev [3]. He was interested in determining the possibility of woven fabrics to
conform to a generic surface. To this end, Chebyshev developed a purely geometrical model
where the fabric was idealized as a network of orthogonal inextensible fibres. Such a surface
model was later referred to as Chebyshev net. By selecting the surface coordinates u and v as the
fibres directions, a smooth parametrization r(u, v) : R → R is a Chebyshev net if r ⋅ r = 1 and
2 3
u u

r ⋅ r = 1 (subscripts denote partial differentiation). The metric properties of such a surface are
v v

determined by the first fundamental form PDF

Help

2 2 2
ds = du + 2 sinγ du dv + dv ,

where ds is a line element on the surface while γ refers to the deformed angle between fibres.
The Gaussian curvature K is linked to the angle between fibres through the differential equation
2.1
ϕ uv + K (u, v) sinγ = 0.

Note that in the case of a constant Gaussian curvature, equation (2.1) reduces to the well-known
Sine-Gordon equation. In practice, the angle between fibres has to change in order to cover
surfaces that are not developable (see figure 1). Chebyshev was able to find the shape of a net

covering a hemisphere by numerical approximation of equation (2.1). The problem for the full
Back
sphere was solved only recently by Ghys [23], highlighting the existence of a singularity to top
of the
https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 4/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

cusp type. The existence of Chebyshev nets were extensively studied in mathematical literature.
On one hand, Bianchi [4] and Bieberbach [24] proved the local existence of Chebyshev net
clothing any regular surface. On the other hand, sufficient conditions for the global existence are
much more difficult to work out. In this context, Hazzidakis [25] showed that a rectangular
Chebyshev patch can fit a given regular surface as long as the magnitude of its total Gaussian
curvature is less than 2π. Despite this limit, Voss [26] showed that smooth Chebyshev patches can
be generated on surfaces of revolutions that do not intersect their rotation axis. Other recent
contributions regarding the global existence of Chebyshev nets include [27,28].

A rigorous approach for the problem of clothing a generic surface with an initially plane
Chebyshev net was proposed by Servant [29,30]. Given a surface, specified by its parametrization
r(x, y) , the clothing problem reduces to express x and y as a function of the coordinates of a

Chebyshev net, i.e. x = x(u, v) and y = y(u, v). Accordingly, the inextensibility condition for the
fibres reads

2 2
E xu + 2F xu yu + G yu = 1

2 2
} 2.2
and E xv + 2F xv yv + G yv = 1,

where E = r ⋅ r , F = r ⋅ r and G = r ⋅ r are the given coefficients of the first fundamental

x x x y y y

form of the target surface. Servant’s equations are then obtained by cross differentiation of
equation (2.2), resulting in a second order quasi-linear hyperbolic system with the fibres as
characteristics. Boundary conditions can be either specified along two intersecting fibres or along
a generic curve. It should be noted that, even in the case of a regular surface, solutions of the
Servant’s equations might present singularities due to hyperbolic behaviour. A pathological case
is found when γ → 0, i.e. in correspondence to an envelope of fibres where the fabric turns back
on itself. This case is representative of collapsed regions where the solution of Servant’s equation
fails to cover completely the target surface. Moreover, discontinuities across characteristics can
frequently arise. The latter case is connected with the appearance of folds in the fabric.

Practical design of Chebyshev nets is largely affected by the regularity of the target surface.
Analytical solutions for the clothing problem (2.2) can only be obtained for simple smooth
manifolds, such as surfaces of revolutions [31,32]. Nevertheless, even in the case of regular PDF
obstacles, the solution of the clothing problem generally involves singularities [23]. The situation
Help
gets even worse in the general context of piecewise smooth surfaces. When singularities appear,
the covering Chebyshev net is typically built by the union of individual patches sewn together.
Aono et al. [33] proposed an algorithm for generating Chebyshev nets on arbitrary smooth
obstacles. The method couples a numerical solver for the solution of the Servant’s equations with
an algorithm for preventing anomalies (e.g. gaps and wrinkles) by the insertion of darts (i.e. folds
stitched into the material). Alternative numerical strategies for the problem of mapping a
Chebyshev net on arbitrary surfaces were proposed using the finite-element method [34]. The
clothing problem was also studied on generic manifolds using discrete Chebyshev nets [35]. The
latter is a discrete analogue of a smooth Chebyshev net consisting of a quadrangular mesh
formed by inextensible edges. In this context, Garg et al. [36] presented a computational
approach for the clothing of a given surface. The method integrates shape optimization 
tools and
allows for the interactive insertion of darts by the user. Finally, an algorithm for the generation of
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 5/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

free-form surfaces composed by Chebyshev net with singularities placed automatically was
proposed in [37].

Studying the clothing problem through a purely geometrical model might give incorrect results in
many practical situations. First, an erroneous prediction of collapsed regions can arise from the
fact that the aforementioned model does not include any shear resistance. Second, geometrical
models work well as long as the fibres are subjected to tensile stresses. Indeed, the presence of
compressive loads along fibres might give rise to out-of plane bucking. The appearance of
wrinkles is then related to the bending stiffness of the cloth. Therefore, rigorous modelling of cloth
deformation, including the analysis of instabilities, shall include the formulation of suitable
constitutive laws describing the stress/strain behaviour of textiles. Generally, the response of
textiles and fabrics is highly anisotropic and nonlinear. It is also characterized by large in-plane
deformations and low-bending stiffness. Inelastic deformations are frequently observed even for
small stresses and at room temperature [38]. The mechanical characterization of fabrics can be
pursued through experimental tests. For instance, the Kawabata Evaluation System [39] provides
a well-established texting procedure for textile mechanics. A qualitative stress/stress response of
woven fabrics is reported in figure 2 for in-plane tensile and shear deformations. Because of
decrimping phenomena, an increasing tensile stiffness is observed for increasing strain along the
yarn directions. As a result, the tensile stiffness of the fabric is generally lower than the
constituting yarns form small strains. The shear stiffness is relatively small in the first part of the
curve. Subsequently, as the shear angle becomes larger, the shear rigidity increases significantly.
Large shear angles might imply yarn contacts potentially leading to wrinkling effects known as
‘shear-locking’ [40].

PDF
Help

Figure 2. Plots of the typical stress/strain response of woven fabrics subjected to (a) tensile deformation along the yarn directions
and (b) shear deformation.

Download figure Open in new tab Download PowerPoint

Rivlin [41] formulated the equilibrium of a Chebyshev net subjected to in-plane deformation. Such
a formulation was later extended to the general deformation of a curved net by application of
forces [41]. In both cases, the mechanics was considered only in terms of tension across fibres.
The effect of shear resistance was later included in [42,43] for both planar and general
deformations. In particular, a comprehensive analysis of possible singularities arising from the
solution of the resulting boundary value problem was discussed therein. Subsequently, 
general
theoretical frameworks for a network of inextensible fibres including bending [44] and Back
twisting
to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 6/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

[45] were also proposed. For completeness, it should be mentioned that similar models were also
developed for elastic nets, i.e. by getting rid of the characteristic fibre inextensibility of Chebyshev
nets (see [46,47]).

(c) Virtual simulations of textiles and garments

A cornerstone in the digitalization of fashion industry is the ability to reproduce the behaviour of
clothes in operational conditions. Virtual garment simulations are at the core of the emerging
virtual-try-on technology, in which customers can virtually try garments in a digital environment.
The final goal is animating the shape of garments worn by a digital mannequin including aesthetic
details such as drapes, folds and wrinkles. Accurate and reliable virtual modelling is a challenging
task since it involves complex interactions between garment deformation, mannequin motion and
external forces. Some of the main difficulties are summarized next. First, fabric is a deformable
solid with a relatively small flexural rigidity compared to the in-plane stiffness. As a result,
garments undergo large displacements with the formation of peculiar drapes and wrinkles
triggered by out-of-plane instability. Second, the resulting garment shape depends on the
application of different forces such as gravity, contact with the wearer, self-collisions and frictions.
Among them, modelling repulsive forces due to cloth–cloth collisions and preventing cloth from
interpenetrating solid objects require involved mathematical treatments. Finally, suitable models
for virtual-try-on should be computationally fast in order for the customer to perform real-time
interactive cloth manipulation.

Early approaches for cloth animation employed physics-based continuous models in which the
fabric is idealized as an elastic surface moving in a three-dimensional space. Continuous models
allow one to accurately describe cloth mechanics by the usage of engineering properties such as
elastic modulus, fabric thickness and density. In their seminal paper, Terzopoulos et al. [48]
demonstrated the potential of continuum mechanics to properly simulate cloth motion subjected
to gravitational forces and collisions with impenetrable objects. Subsequently, a number of
publications investigated the application of rigorous continuous descriptions based on theories of
elastic plates and shells [49–52]. These works employed the finite-element method to numerically
solve the resulting mechanical problem. Since the computational time of continuous models is
generally high, numerical simulations were limited to the draping of relatively simple cloth
patches. Continuous models were soon abandoned in favour of faster to solve discrete models.
PDF
However, the recent rise of isogeometric analysis has opened new perspectives in continuum
Help
clothing animations [53]. As an alternative to the finite-element method, isogeometric analysis is
based on NURBS (non-uniform rational B-spline), a standard technology used in CAD systems.
This numerical method is, therefore, prone to be directly coupled with existing CAD methods for
the design of garments.

Examples of discrete models are the so-called particle methods and mass-spring systems. The
former category idealizes fabrics as a collection of particles that are subjected to specific forces
that model internal cloth structure and the surrounding environment. Different approaches have
been adopted for modelling particle interactions and then computing cloth motion. Breen et al.
[54] treated cloth deformation as an energy minimization problem with total energy function
expressed as

2.3
Back to top
U = Ucoll + Ugra + Ustretch + Ubend + Ushear ,

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 7/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

where U is a collision energy and U is the gravitational potential, while

coll gra

U + U
stretch
+ Ubend
accounts for cloth deformation in terms of stretching, bending and
shear

shearing. Realistic expression for the energy of deformation can be estimated from empirical data
based on the Kabawata testing procedure. Alternatively, Eberhardt et al. [55] employed the
Lagrange equation

d ∂L ∂L
= , 2.4
dt ∂ v i ∂ xi

to determine the trajectories of particles, where v and x refer to velocity and position of the ith
i i

particle, respectively, while L = E − V is the Lagrange function with V being a cloth-specific

kin

potential and E the kinetic energy of the system. The resulting mathematical problem of
kin

particle methods consists of a system of ordinary differential equations in time which is solved by
numerical approximation. The usage of explicit integration schemes generally requires very small
time steps because of the poor conditioning of the problem. To overcome this limitation, Baraff &
Witkin [56] proposed a numerical method based on the combination of implicit integration and the
usage of penalty forces to handle cloth–cloth collisions.

Differently from particle methods, mass-spring systems idealize cloth by means of virtual masses,
each of which is connected to its neighbours by elastic springs [57]. To describe fabric
deformation, three kinds of connective springs were generally considered: structural, shear and
flexural. The dynamics of cloth is governed by Newton’s second law F = m a , with F being the i i i i

total force, m the mass and a the acceleration of each virtual mass i . The applied force splits
i i

between internal and external contributions, F = F + F . Internal forces account for the
i i
int ext
i

resulting tension of the connected springs. External contributions collect forces applied to the
garment as well as self-collisions and interactions with other objects. Due to their reasonable
trade-off between accuracy and simulation speed, mass-spring models have been extensively
adopted and enhanced in different aspects. Traditional linear-elastic models for the mechanics of
fabrics have been replaced with more realistic constitutive laws. For instance, Wang et al. [58]
modelled the typical nonlinear behaviour of fibres using a data-driven piecewise linear
approximation while Volino et al. [59] used the Saint Venant–Kirchoff hyperelastic law to model in-
plane cloth mechanics. Significant efforts have also been made in deriving more accurate andPDF
robust treatments for collisions, contact and frictions. This specific topic is crucial for simulationHelp
speed since it can take most of the computational time. In addition, the prediction of realistic
fabric drapes and wrinkles is largely affected by the accuracy of contact and collision handling. In
this regard, Bridson et al. [60] derived an efficient model based on the combination of fail-safe
geometric collision and a fast repulsion force method. A novel algorithm for incremental collision
detection based on spatial hashing coupled with a GPU-based nonlinear impact zone solver has
been recently proposed by Tang et al. [61]. An alternative fast numerical method for solving the
dynamics of mass-spring systems was presented by Liu et al. [62]. The latter made use of a
coordinate block descent method coupled with a nonlinear Newton solver.

Besides physics-based continuum and discrete methods, cloth simulations are also carried out
using geometrical models. These approaches rely on a purely geometrical descriptionof cloth
instead of incorporating fundamental physics equations. The computational load is, therefore,
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 8/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

much lower than physics-based models at the expense of accuracy due to lack of physical
background. Garments are idealized as the union of cylindrical patches in the geometrical model
for cloth folds proposed by Decaudin et al. [16]. The formation of folds was described by means of
the ‘buckling mesh’ procedure, a solution strategy that combines the three characteristic buckling
modes of cylinders: vertical axis align folds, diamond folds pattern and twist buckling. Finally,
Chen & Tang [63] formulated an alternative geometrical approach by modelling cloth deformation
through a minimization problem including collision handling.

(d) Multi-scale modelling of textiles

The composite industry has in the last decades seen a significant advancement in the mechanical
modelling of textiles [64]. Textile reinforced composites are structural materials combining a
matrix, usually a polymer, and a textile-based reinforcement, typically made of carbon/glass fibres.
Their manufacturing is characterized by a two-step process. First, a planar textile patch is
deformed into the target shape via the so-called forming or preform stage. Second, a liquid resin
is injected within the reinforcement and later consolidated at high temperature. In particular, the
forming process not only determines the fibres alignment of the reinforcement but may also drive
the formation of flaws and defects in the composite. In order to improve the fabrication processes,
a large amount of published works and software focused on the optimization of the forming, as
well as the mechanical characterization of textiles at different scales [65].

The mechanical response of engineering textiles is inherently multi-scale, since it involves an

interplay of (i) macro-, (ii) meso- and (iii) micro-scale effects (see figure 3).

PDF
Figure 3. Sketch of the multi-scale composition of an engineering textile (adapted from [65]).
Help

Download figure Open in new tab Download PowerPoint

The macro-scale is the composite part whereby the dry reinforcement can be described as a
continuous medium. This description is generally preferred to predict textile draping and wrinkling
in the preform stage [66]. Early approaches used purely geometrical approaches similar to those
described in §2(b) (see [67,68] for instance). Subsequently, more accurate approaches were
developed using solid shell theories to include the elastic response of textiles. In this regard, a
number of constitutive laws—generally of hyperelastic [69,70] and hypoelastic [71,72] type—have
been proposed to capture the peculiar mechanical behaviour of these engineering fabrics.

Meso-scale approaches aim at simulating the response of textiles at the intermediate  scale of
interlacing yarns [73]. The main benefit of this approach is the explicit description of the topology
Back to top
of a given textile (e.g. woven, knitted, braided). Meso-scale models can be classified into large
https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 9/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

meso- and unit-cell meso-models [74]. In the first category, the overall preform stage is simulated
with full representation of textile yarns. Therefore, reinforcement deformation can be predicted
precisely, especially at the locations characterized by high distortion, where continuous macro-
scale models generally fail [75]. On the other hand, unit-cell meso-models are used to determine
the homogenized properties of textiles through the simulation of a representative unit-cell of the
textile [76].

The accuracy of meso-scale simulations depends mainly on the digital reconstruction of the textile
structure and on the constitutive laws used for the yarns. A significant number of meso-scale
models considers yarns as continuous media whose interlacing structure is generated by either
geometric methods or microscanning reconstruction [77]. In geometric methods, virtual textiles
are generated according to the given interlacing structure (typically of woven/knitted type) and
yarn properties. More advanced geometric routines also account for the bending/compressive
forces applied to the yarns in order to mimic the actual textile production [78,79]. Geometric
methods offer the possibility of exploring novel interlacing structures and patterns in a fully digital
environment [80]. This capability is nowadays implemented in a variety of commercial and open-
source platforms such as TexGen [81] and WiseTex [82]. Alternatively, the fabric structure can be
reconstructed from real composites through their microscanning [83]. This methodology has two
main benefits: first, it is possible to reconstruct the exact meso-structure of a textile composite;
second, the material parameters of the yarns can be calibrated via comparison between virtual
tests and sample microscanning.

Discrete meso-scale models have also been proposed in order to reduce the computational
burden of continuous meso-models. Here, threads are modelled using either elastic springs or
one-dimensional structural elements such as trusses and beams [84–86]. Compared to continuum
meso-models, discrete elements enhance the treatment of yarns–yarn contact thus avoiding
spurious yarn interpenetration [87].

Micro-scale models consider textiles at the scale of a single constituting fibre [88–90]. In
conventional and engineering textiles, each yarn consists indeed of a bundle made by micrometre
fibres. In this approach, fibres shall be distinguished in order to evaluate their relative motion
within a yarn. The main difficulties in fibres modelling are related to their geometry, mechanical
behaviour as well as their mutual contact/friction interactions. Numerical simulations at such small
scales can be used to predict the geometry of the yarn cross-section as a function on the waving PDF
process, as well as the mechanical properties of a fibre bundle or a textile unit-cell. The potential
Help

of micro-scale simulations has also been exploited in virtual simulations of cloth. For instance,
Kaldor et al. [91] proposed a yarn-level model for the animation of knitted cloth, simulating its
mechanical response using different interlocking patterns. The treatment of collision was later
enhanced by Otaduy and co-workers using a persistent contact method for both woven [92] and
knitted fabrics [93].

3. Mathematics-inspired fashion style

Clothing can serve different functions in everyday life. Besides insulating humans, it has a
remarkable social impact, being a sort of non-verbal communication which may identify many
socioeconomic factors. These aspects are strictly connected to the clothing style, which

encompasses a combination of aesthetic features, such as fabric type, design, fitting, colouring
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 10/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

and patterning. This section deals with the mathematical principles proposed in the creative
design of fabrics and clothes. It pertains to the so-called mathematical art, a practice in which
artists and designers conceive visually attracting shapes and motifs inspired by mathematical
concepts.

(a) Tilings and patterns for fabrics

Tilings and patterns are forms of surface decoration that are often used to make pleasant and
visual attracting designs. Specifically, a tiling, or tessellation, is the covering of the plane or the
space by closed shapes, called tiles, that have disjoint interiors. Patterns are created by the
repetition of some motifs in a more or less systematic manner. Tilings and patterns originated very
early in human history for architectural and artistic purposes. Such geometric decorations have
also been extensively experimented in textile design. In this regard, tilings and patterns can be
created in two different ways. First, the production of fabrics itself leads inevitably to the formation
of a geometrical structure. Moreover, fabrics are often subjected to embellishment through
colouring and drawing of geometrical shapes and motifs.

Some classic examples of tilings and patterns created through weaving are reported in figure 4.
Textile fabrication through waving is generally performed with three distinct techniques: plain
weave, satin and twill. Matt (figure 4a) is a variant of plain wave that creates a checkerboard effect
similar to a tessellation by square tiles. Twill fabrics are characterized by diagonal lines as typically
shown in herringbone and diamond twill fabrics (4b,c, respectively). Houndstooth fabric (known
also as pied-de-poule) is a further example of classic tessellation created by weaving. In this case,
the tiling is characterized by the assembly of equal tiles in the shape of broken checks.

Figure 4. Some examples of geometric patterns arising from fabric weaving: (a) matt fabric, (b) herringbone fabric, (c) diamondPDF
twill
fabric and (d) houndstooth fabric.
Help

Download figure Open in new tab Download PowerPoint

Figure 5 collects some iconic patterns and tailing styles in the fashion industry. The Argyle and
Paisley styles (figure 5a,b) are traditional examples of the usage of tessellation and pattern
concepts, respectively. The former reproduces a tessellation by diamonds. Paisley is Persian
ornamental pattern created by the repetition of the boteh, a stylized teardrop shaped motif. Textile
tilings and patterns are also used by fashion brands to mark their clothes. In this regard, Prada’s
triangle pattern and Burberry’s tartan motif are among the most iconic and recognizable shapes in
the fashion world (see figure 5c,d).

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 11/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Figure 5. Some examples of textile tilings and patterns: (a) Argyle style, (b) Paisley style, (c) Prada pattern and (d) Burberry tartan
pattern.

Download figure Open in new tab Download PowerPoint

Despite the fact that tilings and patterns are well-established forms of art and design, the study of
their mathematical properties is relatively recent. Extensive investigations on the subject started
only at the beginning of the twentieth century. A comprehensive mathematical treatment of the
topic is contained in the classic book by Grünbaum & Shephard [94]. During the last century, the
beauty and mathematical structure of tilings and patterns stimulated the production of fascinating
styles. Escher, Fathauer and Penrose are among the most notable tessellation artists and
mathematicians. Some of their iconic creations inspired the production of innovative fabric styles
as well. For instance, artist Sam Kerr collaborated with fashion designer Paul Smith and
accessories brand Marwood to design tilings for T-shirts, ties, bow ties and pocket squares [95].
Creative tessellations can be designed by adopting a well-established methodology that consists
in replacing the edges of polygonal tiles by suitable curves. An example of this geometric method
is Escher’s famous lizard, whose construction procedure is sketched in figure 6a. In this case, a
lizard is obtained from a regular hexagon by cutting shapes of some sides of the hexagon that are
subsequently rotated and reapplied to other sides. To ensure that the created shapes tessellate
(as shown in figure 6b), precise geometrical rules must be followed. Using a similar strategy, Park
[97] explored the use of this technique to create various patterns based on traditional Korean
motifs.

PDF
Help

Figure 6. (a) Graphical construction of famous Escher’s lizard tile out of a regular hexagon. (b) Reproduction of monohedral Escher
tessellation by lizard shaped tiles. (c) Representation on the Poicaré disc of the pentapentagonal tessellation of the hyperbolic
plane used by Heleman Ferguson for crafting an hyperbolic quilt (courtesy of Jeffrey R. Weeks2). (d) Hyperbolic geometry made via
crocheting [96] (courtesy of Daina Taimina).

Download figure Open in new tab Download PowerPoint

It is also possible to tessellate non-Euclidean geometries, such as the hyperbolic plane. In the
Euclidean plane the only regular edge-to-edge monohedral tessellations by regular polygons

are
the ones that have as prototiles an equilateral triangle, a square and a regular hexagon.
BackIf one
to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 12/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

gives up on the requirements of Euclidean geometry, many more options and styles are available.
As a form of mathematical art, Escher created five works based on tessellations of the hyperbolic
plane. A practical way to craft hyperbolic planes is by sewing curved edges polygons made of
fabric. Heleman Ferguson, an American artist, used this technique to make a quilt that
approximates a uniform pentapentagonal tiling of the hyperbolic plane (figure 6c). Crocheting is
another building technique that can be used to create hyperbolic geometrical shapes as
illustrated by mathematician Daina Taimina [98] (see figure 6d).

Mathematical theories can inspire the generation of alternative graphical shapes and motifs. For
instance, Neves et al. [99] pioneered the usage of computer generated fractals for printed textile
patterns. The term fractal refers to geometric motifs that can replicate their shapes at smaller, or
larger, scales due to inherent self-similarity [100]. By the usage of fractal geometry it is possible to
generate visual attracting motifs that resemble the shape of nature, that found many applications
in the fashion industry [101]. At first, individual fractal motifs are generated using a combination of
fractal theory and computer graphic technology. Once a motif has been generated, a fabric
pattern is created by repeating individual motifs in a structured manner. Finally the designed
pattern is reported on fabrics by means of advanced printing technology. Figure 7 reports a fractal
pattern generated in [101] to be printed on a silk scarf. Further contributions in the research of
fractal art for textile design include [102,103].

Figure 7. Scarf fractal pattern designed in [101]. (a) Computer generated fractal motif. (b) Fabric pattern obtained by the spatial
repetition of individual motifs. (c) Picture of the three-dimensional configuration of a silk scarf prototype digitally printed with the
obtained fractal pattern.

Download figure Open in new tab Download PowerPoint PDF

Help
2
Not surprisingly, the beauty of fractal art for textile design was also researched by notable
fashion designers. Starting from the early 1990s, American stylist Jhane Barnes made extensive
use of fractal patterns generated by an in-house computer software developed by mathematician
William Jones and Dana Cartwright [104]. Her creations demonstrated the capability of
mathematical art to readily inspire an endless spectrum of colourful textile designs.

(b) Geometrical ideas for creative clothes

Besides textile embellishment through unconventional patterns, fashion designers have also
devised creative garment shapes inspired by mathematics. In particular, some complex ideas
developed in geometry and topology became sources of inspiration for designing garment
 forms
and shapes. One of the most iconic examples is the collaboration between Japanese stylist Issey
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 13/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Miyake and American mathematician William Thurston in 2010 [105]. The latter gave a major
breakthrough in the study of topology in two and three dimensions for which he received the
Fields Medal in 1982. In his famous geometrization conjecture, Thurston claimed that all 3-
manifolds result from the combination of eight types of geometrical structures, prevalently of
hyperbolic type [106]. The conjecture was formally proved in 2002 by Perelman [107]. The
connection between Miyake and Thurston was established by creative director Dai Fujiwara. He
was intrigued by Thurston’s treatment of geometry and convinced that it could provide new ideas
for Miyake's Prêt-à-porter fashion line. This collaboration gave rise to the 2010 Autumn/Winter
collection entitled ‘8 Geometry Link Models as Metaphor of the Universe’.3 The concept clothes
therein took inspiration from Thurston’s drawings of the eight 3-manifolds geometry (see figure
8a). These geometrical artefacts were rearranged by Fujiwara to create garments formed by
coloured linked scarves as shown in figure 8b. It is worth mentioning that there is much more than
mere beauty in this collection. It is indeed the attempt to capture the underlying mathematical
concepts as expressions of space.

Figure 8. (a) Thurston’s illustration of orbifolds representative of the eight geometries (Courtesy of Kelly Delp). (b) Issey Miyake’s
2010 Autumn/Winter collection on the runway at Paris Fashion Week (Photograph by Frédérique Dumoulin).

Download figure Open in new tab Download PowerPoint

A further example within the Miyake fashion house is the 2010 brand entitled ‘132 5’. This
collection was inspired from the work on origami by Japanese computer scientist Jun Mitani. PDF
Origami is a form of art in which three-dimensional forms are created by folding a paper sheet. In
Help
a research paper [108], Mitani presented a novel method for designing origami starting from a
target three-dimensional shape. The method automatically generates the crease pattern of the
target shape by inserting appropriate flaps between the constituent polygonal faces. An important
property of this construction is that the resulting surface is developable. Within a collaboration
between Mitani and Miyake, this technique was adapted to develop a new production method for
foldable clothes. The idea, similarly to paper origami, is to create garments by folding a single
piece of cloth as shown in figure 9a. The mathematical foundation of this collection is well
described by the brand name ‘132 5’. A single piece of one-dimensional cloth takes a three-
dimensional form and then is folded into a flat surface (2). (5D) refers to the way that wearing
finally transforms it.


Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 14/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Figure 9. (a) Origami dress from the Issey Miyake ‘123 5’ collection. The unfolding mechanism is illustrated next to the mannequin
(photographer Hiroshi Iwasaki) and (b) Issey Miyake’s ‘Minaret Dress’ presented at the 1995 Spring/Summer collection in Paris [109]
(Powerhouse collection, photographer Sue Stafford).

Download figure Open in new tab Download PowerPoint

Miyake’s continuous research on paper-like developability had been experienced even before the
launch of the ‘132 5’ brand. Concepts of cloth folding and creasing are indeed the pillars of his
famous brand ‘Pleats Please’. In this collection, the crafting of garments was revolutionized by the
usage of a single piece of permanently pleated fabric that is folded with a specific process. The
result is a dress with a sophisticated structure similar to origami clothes. An iconic, yet
representative, example of Miyake’s pleating style is the concept dress entitled ‘Minaret’ (see
figure 9b). Presented at the 1995 Spring/Summer collection, this lantern-shaped dress
demonstrated the ability of the Japanese stylist to master the geometry of ruled surfaces. The
dress consists of the union of developable ruled patches generated from circular rulings. As a
result of this geometrical construction, the garment is completely foldable. As a formidable
example of the interplay between art, technology and science, the dress is exposed at the PDF
Museum of Applied Art and Science of Australia.
Help

4. Mathematics of digital manufacturing and smart textiles

Digital manufacturing is one of the pillars of the current industrial revolution, rapidly changing the
way we design, produce and distribute nearly all products. It relies on the integration of computer-
aided design and simulation tools to simultaneously create products and manufacturing
procedures. This section discusses the prominent role played by physics-based mathematical
modelling for digital manufacturing of garments and for the design of smart textiles.

(a) Additive manufacturing of textiles and garments

Additive manufacturing is a digital methodology of making three-dimensional objects by 
adding
material in a layer-by-layer fashion [110]. After creating a three-dimensional model withBack
CADto top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 15/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

programmes, this manufacturing technique is realized through the usage of proper three-
dimensional printers and applies to a wide range of materials, including polymers, ceramics and
metals. Available three-dimensional printing techniques depend on the phase of the material used
in the process such as powder, solid and liquid-based methods. Widespread examples of three-
dimensional printing methods include fused deposition modelling (FDM), selective laser sintering
(SLS), stereolithography (SLA) and inject binding [111], offering inherent advantages over
conventional manufacturing [112].

Not surprisingly, additive manufacturing is revolutionizing the fashion industry. For instance,
fashion brands Cheng4 and Timberland5 partnered with three-dimensional printing companies to
produce shoes with printed plastic midsoles. This technological evolution will soon enable the
production of personalized on-demand fashion items as already established in the jewellry
industry [113]. However, additive manufacturing of textiles and garments is still in its infancy as it
has mainly explored for the production of concept clothes. For example, Dutch fashion designer
Iris van Herpen made use of SLA to craft a semitransparent dress with a highly complex
geometrical structure.6 Using the same printing technology, a flexible floor-length nylon gown was
created by Michael Schmidt and Francis Bitonti for burlesque dancer Dita Von Teese7. In such a
case, the printed cloth embodies a sophisticated grid-like structure that provides flexibility and
readily adapts to the wearer's body shape.

The research on three-dimensionally-printed textiles and garments has lagged behind other
applications because of some intrinsic difficulties. The main one is recreating the essential
properties of conventional textiles such as flexibility, softness, strength and porosity [114].
Flexibility of printed fabrics depends on the combination between the material employed and
textile structure. Looking at the body of published literature, three-dimensionally-printed textiles
are generally produced by replicating conventional interlocking and interlacing structures. For
example, Bingham et al. [115] designed and printed a conformal garment using a particular chain
mail structured textile that can be folded and bent like conventional fabrics. Using a similar
methodology, Bloomfield & Borstrock [116] created a printed textile from individual links joined
together. The printing of textiles with weft-knit structure was investigated by Melnikova et al. [117]
using both SLS ans FDM. Single-faced and double-faced knit structured sheets were also
fabricated in [118]. In the realm of woven structures, Partsch et al. [119] focused on three-
dimensional printing of fabrics with various woven configurations, yarn cross-sections and yarn
PDF
densities. More recently, novel strategies for printed woven and quasi-woven flexible textiles have
Help
been proposed in [120,121]. Flexible three-dimensionally-printed garments can also be designed
using non-conventional textile structures. For instance, the kinematic dress by Nervous System
consists of rigid triangular panels linked by hinges [122]. Thanks to the kinematic concept
underneath, the garment drapes like conventional cloth and can be folded to a smaller shape for
efficient fabrication.

Theoretical and numerical modelling can play a valuable role in the rational design of textiles with
tailored functionality. By contrast to trial-and-error methodologies, the use of predictive tools
allows for virtual prototyping and rapid design exploration. Their potential is emphasized in the
case of additive manufacturing, whereby the printed product reproduces the feature of a digital
model with high fidelity. Three-dimensionally-printed textiles are a special class of flexible
mechanical metamaterials [123]. Their structure thus consists of periodically arranged  building
blocks whose collective interaction determines the overall textile behaviour. Suitable modelling
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 16/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

approaches of such structured materials rely on homogenization theories, since direct numerical
simulations tend to be computationally hard to solve. This particular strategy provides a way of
replacing a small scale structure with an average, homogeneous medium. Among the various
developed methodologies, computational homogenization represents one of the most powerful
techniques [124]. The averaged mechanical properties therein are computed through the
numerical solution of a boundary value problem of an adequate digital representation of the
micro-structure.

A general modelling strategy applicable to three-dimensionally-printed textiles is the two-scale

model proposed by Geers et al. [125]. As schematized in figure 10, the method relies on the
nested solution of two problems at different scales. At the so-called macro-scale the textile is
idealized as a homogeneous solid shell with no assumptions on its constitutive behaviour. The
macro-scale constitutive response, for each material point, is computed from the averaged
solution of a micro-scale problem. The latter consists of a boundary value problem on a
representative volume element (RVE) of the structured sheet. At this scale, the geometry of the
structure is fully described and suitable closed-form constitutive laws model the behaviour of the
constituting material phases. Loading and boundary conditions for the micro-scale problem
depend on the kinematics of the corresponding macroscopic points. In the simple case of first-
order computational homogenization, periodic boundary conditions are generally employed.

PDF
Help

Figure 10. Schematic of a shell-based computational homogenization scheme redrawn from [125]. The macro-scale kinematics
provide boundary and loading conditions for the micro-scale boundary value problem (BVP). Symbols εαβ , γα , καβ refer to shell
in-plane strains, transverse shear strains and curvature components, respectively. After solving the micro-scale BVP, generalized
shell forces (Nαβ , Qα , Mαβ ) and tangents are extracted and sent back to the macro-scale.

Download figure Open in new tab Download PowerPoint

The selection of a proper RVE for the micro-scale problem is not a trivial task. In the case of
perfectly periodic structures, the RVE usually coincides with the unit-cell, i.e. the smallest tileable
unit of the micro-structure. However, in the case of slender structures, compressive loads may
give rise to out-of-plane buckling that can alter the initial periodicity of the micro-structure
 [126]. To
Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 17/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

predict such a morphological transformation, the computational homogenization scheme can be

complemented with a Bloch-type stability analysis [127].

In recent years, various articles made use of computational homogenization for the mechanical
characterization of conventional and three-dimensionally-printed textiles. To reduce the
computational cost, the constitutive behaviour at the macro-scale is generally computed offline by
means of a large number of numerical simulations at the micro-scale for a prescribed deformation
space. For instance, Sperl et al. [128] performed yarn-level numerical simulations using the theory
of elastic rods to build a model for the potential energy of the cloth at the macro-scale. The
resulting homogenized model therein was used to efficiently simulate draping of various knitted
and woven fabrics. Using a different approach, the homogenized response of woven fabrics was
calculated through micro-scale analyses based on standard continuum mechanics in [129].
Focusing on three-dimensionally-printed textiles, Schumacher et al. [130] characterized the macro-
scale response of printed structured sheets represented as a network of thin Kirchoff rods. Finally,
Li et al. [131] developed a dedicated computational framework for computing the mechanical
response of printed textile-like materials with sliding connections that resemble the yarn structure
of conventional fabrics.

(b) Smart textiles and four-dimensional printing

Over the past few decades, there have been a tremendous development of smart textiles for the
production of garments with enhanced functionalities [132]. In a broad sense, smart textiles are
unconventional fabrics that are either fully or partly made with stimuli-responsive materials. The
great advantage of this class of textiles is their ability to change aesthetic and functional
properties as a function of environmental conditions [133]. Temperature and moisture regulation,
self-motion, self-cleaning and colour change are some of the functional feedbacks that have been
developed so far [134]. The behaviour of the employed stimuli-responsive materials, as well as the
smart fabric production process, are key factors for the emergence of such advanced
functionalities.

Shape-change and colour-change responses are the stimuli that have found a wider application in
the textile and clothing industry. Shape-memory-alloys (SMAs) and shape-memory-polymers
(SMPs) are two important examples of smart materials. Their peculiar behaviour comes from the
PDF
shape-memory effect, i.e. the ability to recover predetermined, programmed, shapes in response
to the right stimulus [135]. The shape-memory effect in SMAs typically stems from a thermally Help
activated phase transformation between the high temperature austenite and low temperature
martensite phase. Depending on the programming procedure, SMAs can be distinguished
between one-way SMAs, in which only the high temperature shape is recovered, and two-way
SMAs, where both high and low temperature shapes can be remembered. On the other hand,
shape-memory effects in SMPs can be programmed thanks to the glass transition mechanism.
Polymer based smart materials are generally more versatile than metal based [136]. Some of their
advantages include higher recoverable strains, easy shape programming and multi-shape-
memory effect. Besides thermal activation, a shape-change response in polymers can be
triggered by water/moisture stimuli as well, as in the case of hydrogels. Other examples of smart
materials for potential applications in textiles include photo-chromic [137] and electro-responsive
materials. 

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 18/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

One of the first documented uses of smart textiles in the fashion world is the ‘Oricalco’ shirt
designed by Italian brand Corpo Nove.8 This smart shirt actively reacts to temperature changes
thanks to the use of Nilitol yarns, a shape-memory alloy made of nickel and titanium. The sleeves
can be programmed to shorten as the temperature increases and no ironing is needed as the
original shape can be recovered by just a flux of hot air. A Nilitol-based active textile was also
developed by Bezowska & Coelho [138] to fabricate two heat-responsive concept garments.
Simple aesthetic functionalities were programmed therein in the form of animated felt flowers and
moving hemlines. Chan Vili [139] investigated the incorporation of shape-memory-materials in
various woven structures to enhance the functionality of interior textiles. Smart curtains were
programmed to open and close their woven structure depending on sunlight conditions. Water-
responsive polymers can find application in the sportwear sector to improve body thermo-
regulation. For instance, sport brand Nike launched a smart T-shirt named ‘Sphere React Shirt’ that
embodies small rear vents that can open up when the wearer sweats [140]. Colour-changing
fashion items have been commercialized as well. For example, a colour-changing fabric
technology named ‘ChroMorphus’ was developed by researchers at the University of Central
Florida.9 Thanks to the usage of smart wires waved into the clothing, the fabric can be
programmed to change its coloured pattern. Finally, the emergence of new conductive fabrics and
soft sensors has enabled the rapid development of electronic-based textiles, or e-textiles [141].
Widespread applications of e-textiles include smart garments that can monitor physiological
parameters such as heart-rate, breathing-rate and muscle activity [142].

The combination between smart materials and additive manufacturing gives rise to the so-called
four-dimensional printing10. In essence, four-dimensional printing can be defined as the targeted
evolution of four-dimensional printed structures, in terms of shape, property and functionality
[143]. The fourth dimension thus refers to the time scale needed by smart, three-dimensionally-
printed, structures to develop a targeted shape dynamics in the presence of a tailored stimulus.
Multiple shape-morphing behaviours can be obtained though four-dimensional printing. Typical
examples include bending, folding, twisting and the generation of surface topological features
such as wrinkles, creases and buckles. Programmed self-assembly of such structures enhances
significantly the potential of additive manufacturing. Indeed, complicated geometries can be first
printed on the plane and then actuated in the location of use. This aspect can impact positively on
storage and transportation of goods. On the other hand, current limitations of four-dimensional
printing are related to the printability of smart materials and the development of multimaterial PDF
printers [144]. Help

Except for a few examples, four-dimensional printing has not yet been explored in the fashion
industry. A comprehensive discussion on the potential of four-dimensional printing for textiles and
fashion goods is contained in [145]. One notable prototype of four-dimensionally-printed textiles is
the ‘Space Fabric’ designed by NASA Jet Propulsion Laboratory.11 The printed textile structure
looks like a chain mail with small silver squares joined together. Thanks to its particular structure,
the space fabric can fold and adapt to shapes while maintaining tensile strength. In addition, the
use of stimuli-responsive materials allows for thermal regulation through combined light
reflectivity and passive heat management.

Tailored shape-morphing of four-dimensionally-printed structures is generally realized using a

combination of smart and conventional materials. Mathematical modelling plays a keyrole
in optimal design of smart structures in order to achieve the desired change in shape, Back
property
to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 19/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

and functionality [143]. A simple idea of shape-shifting material is based on kirigami tessellations,
regular planar patterns formed by cut portion of thin sheets, whose number, size and orientation
can be inversely defined to conform approximately to any prescribed target shape [146].
Combined with the possibility to three-dimensionally-printed materials with controlled physical
and geometrical properties at the micro-scale, it is now possible to solve the geometric inverse
problem for a complex structured material to grow into a any target shape [147,148]. Another
crucial feature regards the implementation of suitable physical models for the mechanical
response of the smart materials. Such an aspect includes the development of realistic constitutive
laws, the estimation of the relevant material parameters and the description of the stimulus
properties (such as type, intensity and duration).

Modelling of smart structures is inherently multi-physical since actuation results from the
interaction of different physical phenomena (e.g. mechanical, thermal, chemical and electrical).
Rigorous multi-physics models are derived from the application of a few fundamental laws, e.g.
the balance equations, the constitutive theory and the laws of thermodynamics, in terms of energy
balance and entropy imbalance. Moreover, in multi-scale homogenization theories, the so-called
Hill–Mandel condition imposes the conservation of energy across scales. The model governing
equations resulting from this sequence of tasks are typically nonlinear partial differential
equations. Therefore, the solution of such models is sought by numerical approximation.

In recent years, various scientific articles proposed combined simulation and experiments
procedures for the fabrication of smart structures. For example, Ge et al. [149], modelled the
actuation of four-dimensionally-printed structures made of thermally activated SMPs. As reported
in figure 11a, shape morphing was accurately predicted by means of three-dimensional thermo-
mechanical numerical simulations. The constitutive behaviour of the SMP was adapted from the
multi-branch model of Yu et al. [151]. The latter uses a multi-branch model to simulate the multi-
shape memory effect of thermo-activated polymers. Zhang et al. [150] focused on lightweight
smart sheets obtained by three-dimensional-printing of SMPs. The thermo-mechanical behaviour
of the printed structures was modelled using two different methods. An extension of the thermo-
mechanical beam model was adopted for grid-like structures. A thermo-mechanical continuum
model was instead used for predicting shape-morphing of two-dimensional lattice structures (see
figure 11b). Concerning water-responsive polymers, Gladman et al. [152] modelled and printed
composite hydrogels encoded with localized anisotropic swelling to develop complex shape-
PDF
changes. The theoretical model was developed based on a thermo-mechanical theory for
Help
anisotropic surfaces.

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 20/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Figure 11. Modelling of smart structures: (a) Comparison between three-dimensionally-printed SMPs stents and finite-element
numerical simulation of the thermo-activated shape-morphing [149]. (b) Prediction of the final configuration (right) of a thermo-
responsive lattice material with initial straight beams (left) [150].

Download figure Open in new tab Download PowerPoint

5. Concluding remarks
In conclusion, we have outlined the evolution of the role played by mathematics in the fashion
industry over the last century. On one hand, even more complex topological and geometrical
principles have permeated the creative processes of designers, inspiring the innovation of
garments and textiles. On the other hand, the radical switch to digital manufacturing has found in
mathematics a key enabling tool for product innovation. Hence, mathematical models and
methods are nowadays widely used in the fashion industry not only to assist the item conception
and the material design, but also to improve the methods of production and fabrication.

Sustainability is undoubtedly the most important fallback of such novel applications of

mathematics in fashion. From a production perspective, we have shown how mathematical
models allow us to optimize fabrication processes for reducing material waste and environmental
impact; to propose new ideas for recycling, reusing and repurposing that support the conversion
PDF
from traditional to more sustainably sourced materials; to provide software tools, such as virtual-
try-on and digital items, that have a lower carbon footprint and minimize in-shop returns; to Help

personalize the production on-demand for shortening the supply chain. Finally, although out of the
scope of this article, it is worth mentioning that math-based models are at the basis of any
blockchain-based platform that are currently used to promote a circular business model, aiming to
promote trust and transparency for customers, innovation and sustainability. These tools allow us
to trace the source of raw materials and their transformation along the whole supply chain. Thus,
customers gain not only a proof of authenticity against counterfeiting, promoting re-selling in
second-hand markets, but also awareness of the environmental impact of their choices.

Data accessibility
This article has no additional data. 

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 21/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Authors' contributions
M.M.: conceptualization, formal analysis, investigation, methodology, software, writing—original
draft, writing—review and editing; P.C.: conceptualization, formal analysis, funding acquisition,
investigation, methodology, project administration, resources, writing—original draft, writing—
review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work
performed therein.

Conflict of interest declaration

We declare we have no competing interests.
All Journals 

Funding
This work was partly supported by MUR, PRIN Research Projects 2017KL4EF3 and 2020F3NCPX,
by the National Group of Mathematical Physics (GNFM–INdAM) and by European Commission
through FSE REACT-EU funds, PON Ricerca e Innovazione.

Footnotes
1 https://unfashionalliance.org/.

2 www.ams.org/publicoutreach/ams-news-releases/thurston-miyake.

3 www.3dsystems.com/learning-center/case-studies/timberland-company/.

4 https://www.forbes.com/sites/andriacheng/2018/05/22/with-adidas-3d-printingmay-finally-see-
its-mass-retail-potential/?sh=4c12dd834a60.

5 www.materialise.com/en/inspiration/cases/iris-van-herpen/.

6 www.gzinnovation.eu/material/7/shape-memory-materials/.
PDF
7 https://www.dezeen.com/2013/03/07/3d-printed-dress-dita-von-teese-michaelschmidt-francis-
bitonti/. Help

8 www.cnet.com/tech/mobile/color-changing-smart-clothes-smart-fabrics-purse-ucf-
chromorphous/.

9 www.jpl.nasa.gov/news/space-fabric-links-fashion-and-engineering/.

10 https://www.ted.com/talks/skylar_tibbits_the_emergence_of_4d_printing.

11 www.geometrygames.org/HyperbolicBlanket/.

© 2023 The Authors.


Published by the Royal Society under the terms of the Creative Commons Attribution License
Back to top
http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original
https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 22/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

author and source are credited.

VIEW FULL TEXT | DOWNLOAD PDF

Recommended articles

A mathematical model of adult subventricular neurogenesis

J. M. A. Ashbourn et al., Interface, 2012

Introductory remarks
Stephen. K. Scott et al., Philosophical Transactions A, 1991

Macrocyclic intermediates in the biosynthesis of porphyrins

A. H. Jackson et al., Philosophical Transactions B, 1976

The speaker’s linearization problem

W. J. M. Levelt et al., Philosophical Transactions B, 1981

Fashion forward face masks on display

Alison Shepherd, The BMJ, 2021

Role of mathematical modelling in future pandemic response policy

British Medical Journal Publishing Group, The BMJ, 2022

Analysis and optimal control of a Huanglongbing mathematical model with resistant vector
Youquan Luo et al., Infectious Disease Modelling, 2021

Unsupervised Recognition of Informative Features via Tensor Network Machine Learning and Quantum Entanglement
Variations
Sheng-Chen Bai et al., Chinese Physics Letters, 2022

Powered by
PDF
Help

PROCEEDINGS A

About this journal

Author benefits

Submit


Contact information Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 23/24
12/8/23, 1:09 PM Mathematics meets the fashion industry on path to product innovation and sustainability | Proceedings of the Royal Society A…

Transformative journals

Purchasing information

Journal metrics

Recommend to your library

Search help

ROYAL SOCIETY PUBLISHING

Our journals

Historical context

Open access

Open science

Publishing policies

Permissions

Conferences

Videos

Blog

Manage your account

Terms & conditions

Privacy policy

Cookies

THE ROYAL SOCIETY

About us PDF
Help
Contact us

Fellows

Events

Grants, schemes & awards

Topics & policy

Collections

Venue hire

Back to top

https://royalsocietypublishing.org/doi/10.1098/rspa.2022.0715 24/24