Origami and Partial

Differential Equations
Bernard Dacorogna, Paolo Marcellini, and Emanuele Paolini

O
rigami is the ancient Japanese art of
folding paper. Even if origami is mainly
an artistic product, it has received a
great deal of attention from mathe-
maticians, because of its interesting
algebraic and geometrical properties.
We present a new mathematical model of
origami with a double purpose. On one hand y |y|
Figure 1. The map u(x, y) = (x, √2 , √2 )
we give an analytical approach which provides describes a 90-degree folding in R3 of the
a new perspective to the existing algebraic and square paper Ω in the left-hand side. The
geometrical models. On the other hand we use gradient of u is a 3 × 2 orthogonal matrix.
origami as a tool to exhibit explicit solutions to
some systems of partial differential equations.
Mechanical properties of paper, as a material, are
simple. A sheet of paper is rigid in tangential direc-
tions. Indeed, it cannot be stretched, compressed,
or sheared. If a sheet of paper is constrained on
a plane, it would only be possible to achieve rigid
motions, i.e., rotations and translations of the
whole sheet. On the other hand, in the normal
direction it can be easily folded. This is due to
the infinitesimal thickness of the paper and to Figure 2. The map
the elastic properties of the fibers composing the u(x, y) = (x, sign(y)(1−cos y), sin |y|), is not
material. piecewise linear. However, the gradient is a
The physical properties we have described so far 3 × 2 orthogonal matrix.
can be defined from a mathematical point of view.
To this aim we let Ω ⊂ R2 be a two-dimensional
domain (usually a rectangle), which is the reference configuration for the sheet of paper. An origami is
a suitable immersion of the sheet of paper in the
Bernard Dacorogna is professor of mathematics at the three-dimensional space. Hence it can be identified
École Polytechnique Fédérale de Lausanne, Switzerland. with a map u (see Figures 1 and 2)
His email address is bernard.dacorogna@epfl.ch.
u : Ω ⊂ R2 → R3 .
Paolo Marcellini is professor of mathematics at the Uni-
versity of Firenze, Italy. His email address is marcellini@ The tangential rigidity can be expressed by
math.unifi.it. requiring that u is a rigid map of Ω into R3 .
Emanuele Paolini is assistant professor of mathematics In other words the gradient Du(x) of the map
at the University of Firenze, Italy. His email address is u is an orthogonal 3 × 2 matrix, meaning that,
paolini@math.unifi.it.

598 Notices of the AMS Volume 57, Number 5

Figure 3. On the right: the crane is the most famous origami; on the left: the corresponding
singular set.

for every x ∈ Ω, We will see (cf. Theorem 4.1) that such condition is
3×2 t enough to guarantee, on a simply connected open
Du(x) ∈ O(3, 2) := {M ∈ R : M M = I}.
set Ω, that a given union of segments Σ is actually
t
Since Du Du = I, the rank of Du is maximal; the singular set Σu of some rigid map u.
hence the immersion is locally one-to-one. Thus, Our general definition allows us, in a natural way,
if u is sufficiently regular, it is a local isometry to consider in the higher-dimensional case a map
between R2 and the manifold u(Ω) ⊂ R3 with the u : Ω ⊂ Rn → Rm , where the gradient Du(x) is an
metric induced by the ambient space R3 . Therefore orthogonal m × n matrix. That is, Du(x) ∈ O(m, n),
the Gauss curvature of u(Ω) coincides with the where
Gauss curvature of Ω, which is zero, and u(Ω) is a
developable surface (see [15]). O(m, n) := {M ∈ Rm×n : M t M = I}.
Since origami is a folded paper, the map u cannot When m = n we simply write O(n). In this case
be everywhere smooth; it is only piecewise smooth. our mathematical global approach also allows
Folding creates discontinuities in the gradient. us to consider flat origami in higher dimensions
Since we do not allow cutting the sheet of paper, (u : Ω ⊂ Rn → Rn , Du(x) ∈ O(n)), which will also be
u is, however, a continuous map. The singular set called hyper-origami. As an example, for n = 3, we
Σu ⊂ Ω, which is the set of discontinuities of the will describe, in the section “Fractal Construction”,
gradient Du, is called crease pattern in the origami how to “fold” a cube.
context. Usually this set is composed by straight Closely related is the more difficult problem of
segments (as in Figure 3), but it is also possible to finding a map with orthogonal gradient satisfying
make origami with curved folds (see Figure 4). some boundary condition. For example, when
This formulation of an origami as a map
n = m = 2, we consider the Dirichlet problem
u : Ω ⊂ R2 → R3 also allows us to consider modular (
origami, which corresponds to a domain Ω which Du(x) ∈ O(2) for a.e. x ∈ Ω
is not connected (several sheets of paper), as for (1)
u(x) = 0 for x ∈ ∂Ω.
instance in Figure 5.
A special attention will be given to the so-called At first sight one could think that, by means of
flat origami. A flat origami is defined as a map repeated folding, a sheet of paper could in principle
whose image is contained in a plane; it can be be crunched to a single point, say the origin. To
represented, up to a change of coordinates, as a convince oneself that this is not possible it is
map enough to look at the equation. An origami folded
u : Ω ⊂ R2 → R2 . to a single point y0 (= 0) is in fact represented
by the constant map u ≡ y0 , which, however, has
In this case it can be proved that the folding lines
gradient Du ≡ 0 6∈ O(2).
must be straight segments which satisfy some
In order to solve (1), we have to take advantage
angle condition at every vertex of the singular set
of the fact that the boundary does not have full
Σu . More precisely, at any vertex of the singular
dimension. Anyway, the construction cannot be
set Σu , exactly 2k consecutive angles α1 , · · · , α2k
that simple; in fact, it is impossible to find a map
meet with the property that
solving the system and which is smooth on some
α1 + α3 + · · · + α2k−1 = α2 + α4 + · · · + α2k = π . neighborhood of a point of the boundary. Indeed,

May 2010 Notices of the AMS 599

 Using this property of the singular set we are able
to construct some examples of fractal origami which
satisfy some “nonisometric” boundary conditions.
In general, solutions to (1) may be very irregular, up
to the point that the singular set can be everywhere
dense.
In conclusion, origami is seen here as a tool for
finding simple solutions in a context of systems
Figure 4. A circular crease can be folded to of partial differential equations where infinitely
obtain a cone. many solutions exist, none of them being of class
C 1 , most of them having very irregular gradient.
General existence theories have been developed by
means of convex integration and by Baire category
method to obtain Lipschitz continuous solutions.
Simpler scalar problems can be treated with the
viscosity theory, obtaining explicit representation
formulas. This theory cannot, however, be applied
in the context considered here, due to the lack of a
maximum principle. Nevertheless, origami allows
us to present particularly simple solutions with
a recursive structure. It is a challenging issue to
adapt the method proposed here to more general
systems of first-order differential systems such as
those encountered in geometry, optimal design,
and nonlinear elasticity (see the section: “Implicit
Partial Differential Equations: Conclusion”).

Figure 5. A picture of a modular origami

representing an icosahedron, built with 30
identical copies of a folded square, linked
together in a symmetrical way.

let u be a smooth map which satisfies Du ∈ O(2)

or equivalently
2
|Du| = 2 |det Du| = 2.
Since det Du = ±1, by continuity the Jacobian has Figure 6. On the right-hand side a solution to
a sign, say det Du = 1. Therefore the equation the Dirichlet problem (1) actually folded with
|Du|
2 paper; it gives rise to a fractal shape. In the
= 2 det Du, with the notation u (x, y) =
left-hand side the singular set of the same map

v(x,y)
w (x,y)
, can be easily transformed into the Cauchy-
is represented in a region near the boundary.
Riemann system
(
vx − wy = 0
vy + wx = 0
Some Other Topics About Origami
with furthermore |Dv| = |Dw | = 1. Denoting by
Many mathematicians interested in geometry or
τ, ν, respectively, the tangent and normal unit
algebra (for example, in group theory, Galois theory,
vectors on ∂Ω, we thus have, up to sign,
graph theory…) have studied origami constructions.
hDv; τi = hDw ; νi , hDw ; τi = hDv; νi . An important issue is the geometrical construction
Since v = w = 0 on ∂Ω, we also obtain Dv = Dw = of numbers. In this aspect origami turns out to be
0, which contradicts the fact that |Dv| = |Dw | = 1. more powerful than the classical rule and compass
Thus any solution to the differential problem constructions (see [1], [5], [6] or [17]). In order
(1) is Lipschitz continuous but not of class C 1 near to determine what can be constructed through
the boundary; therefore it assumes in a fractal origami, it is important to formalize the rules.
way the homogenous boundary datum ϕ = 0 (see These are known as Huzita axioms and have been
Figure 6). The map u will be explicitly defined at proposed by Hatori, Huzita, Justin, and Lang. Here
every (x, y) ∈ Ω, and it will be piecewise affine, are the seven axioms.
with infinitely many pieces, in accord with its Axiom 1: Given two points P1 and P2 , there is a
fractal nature near the boundary of Ω. unique fold passing through both of them.

600 Notices of the AMS Volume 57, Number 5

 Figure 7. Trisecting an angle. The resulting red angle is a third of the given green angle.

Axiom 2: Given two points P1 and P2 , there is a One of the main points turns out to be that of
unique fold placing P1 onto P2 . finding the best possible packing of given disks
Axiom 3: Given two lines L1 and L2 , there is a fold into the paper rectangle (see the left-hand side in
placing L1 onto L2 . Figure 8); in fact, we notice that every origami is a
Axiom 4: Given a point P and a line L, there is a short map, namely
unique fold perpendicular to L passing through P .
Axiom 5: Given two points P1 and P2 and a line |u (x) − u (y)| ≤ |x − y| , ∀ x, y ∈ Ω;
L, there is a fold placing P1 onto L and passing this means that the distance of two points in the
through P2 . folded model cannot be larger than the distance of
Axiom 6: Given two points P1 and P2 and two lines the two points in the unfolded model. This implies,
L1 and L2 , there is a fold placing P1 onto L1 and P2 for example, that the preimage of the claws of
onto L2 . the crab in Figure 8 must contain the whole circle
Axiom 7: Given a point P and two lines L1 and L2 , centered in the preimage of the tip of the claw with
there is a fold placing P onto L1 and perpendicular radius given by the length of the claw.
to L2 . Recently in [19] some numerical algorithms have
been found to reproduce three-dimensional curved
For a slightly different approach to these axioms models. Some of them have the peculiar property
and for the connection to Galois theory, we refer that the crease pattern is mostly composed of
to the book by Cox [6]. curved lines, whereas classical origami only uses
Origami constructions, contrary to rule-and- straight lines. Curved creases have also been
compass ones, allow one to trisect any angle (see considered by Huffman [15].
Figure 7), to double a cube, or to construct a regular Another interesting application comes from
heptagon. aerospace engineering. A team of scientists at the
Another topic, more closely related to what we Lawrence Livermore National Laboratory in the
discuss in this paper, is the problem of finding an United States are building a large space telescope.
algorithm to decide whether a given crease pattern To put it into action in space, the telescope will be
can actually be folded to generate a physical origami, folded and unfolded following the rules of origami.
i.e., a rigid map without interpenetration. The The picture in Figure 9, taken from [14], represents
problem turns out to be NP-hard (nondeterministic a prototype of the telescope.
polynomial-time hard), see [4].
Other interesting research is devoted to con-
structing crease patterns to fold to some given
three-dimensional models with elaborated shape.
With suitable software, Lang [21] has been able to
fold realistic models of different kinds of animals.
(Courtesy/LLNL.)
(Image(s) courtesy of Robert J.
Lang, www.langorigami.com.

Figure 9. The unfolded space telescope and, on

the left, the corresponding singular set.

Figure 8. On the left-hand side the “ball Mathematical Origami as a Rigid Map
packing technique” to create the crease In this presentation an origami will be modeled by
pattern of the model (crab) shown on the right the notion of rigid map. A map u : Ω ⊂ Rn → Rm is
(by R. T. Lang). said to be a rigid map if it is Lipschitz continuous
and Du(x) ∈ O(m, n) for a.e. x ∈ Ω.

May 2010 Notices of the AMS 601

 From the definition it follows that, given any and the sign of the determinant gives a coloration
two vectors σ , τ ∈ Rn , we have to Ω \ Σu .
hDu(x)σ , Du(x)τiRm = h(Du(x))t Du(x)σ , τiRn In particular we see that if M1 is given, M2 is
uniquely determined. From these considerations it
= hσ , τiRn , follows that, if Ω is also connected, once we know
where h·, ·iRn is the scalar product in Rn . In the affine map u in a single region, we can recover
particular u leaves invariant the metric on the it in every other region.
tangent space. However, not every polyhedral complex Σ is
Since u is Lipschitz continuous it is, by the clas- the singular set of some rigid map u. Notice, in
sical Rademacher theorem, differentiable almost particular, that if e is a (n − 2)-dimensional edge of
everywhere in Ω. We call the singular set of u the Σ, then the number of (n − 1)-dimensional facets
set Σu where the map u is not differentiable. We containing e must be even. This is due to the
now quote two classical results. The first one says 2-coloration property of the complement of Σu . So
that isometries are rigid maps with empty singular we denote with f1 , · · · , f2k the faces containing e. If
sets. we denote with αi the angles between faces fi and
fi+1 , by means of condition (2) we can show that
Theorem 3.1 (Cartan-Dieudonné). Let Ω ⊂ Rn be the following angle condition holds (see Figure 10):
an open connected set and u : Ω → Rm be an isome-
k k
try: X X
(3) α2i−1 = α2i = π .
|u(x) − u(y)| = |x − y|, ∀x, y ∈ Ω. i=1 i=1

Then m ≥ n and u is affine. This condition is well known in the study of

flat origami. When n = 2 this condition was first
The second result applies to the case m = n. In observed by Kawasaki [18].
this case a rigid map corresponds to a flat origami.
The theorem states that a flat and smooth origami
can only be trivial.
Theorem 3.2 (Liouville). Let Ω be a connected open
subset of Rn . If u : Ω → Rn is of class C 1 and satisfies
Du ∈ O(n), then u is affine.
From now on we will only consider piecewise C 1
rigid maps, which are maps u : Ω → Rn satisfying
the following properties: (i) Σu is relatively closed in
Ω; (ii) u is of class C 1 in every connected component
of Ω \ Σu ; (iii) every compact set K ⊂ Ω intersects
only a finite number of connected components of
Ω \ Σu .

Reconstruction from the Crease Pattern

As above, when n = m we are in the flat case, and Figure 10. The angle condition. The sum of
we consider flat origami, i.e., piecewise C 1 rigid angles with the same color is equal to π .
maps u : Ω ⊂ Rn → Rn . In this context we can prove
that, if Ω is a convex polyhedral set, then every
Hence, up to now, we know that the singular set
connected component of Ω \ Σu is also a convex
Σu of a piecewise C 1 rigid map u is a polyhedral
polyhedral set. The singular set Σu is the union of
set such that Ω \ Σu is 2-colorable; moreover, every
the faces bounding the convex regions, hence it is
(n − 2)-dimensional edge is contained in an even
itself a (n − 1)-dimensional polyhedral complex.
number of (n − 1)-dimensional facets spaced so
Every (n − 1)-dimensional facet of Σu is the
that the sum of angles on the regions of one color
interface between two different regions, say A1
is equal to the sum of angles for the regions of the
and A2 . On each region the map u is affine,
other color.
u(x) = Mi x + qi for x ∈ Ai , and it is not difficult to
It is interesting to know that the reverse impli-
show that the two affine maps are related to each
cation, stated in the following theorem, holds (see
other by
[10]).
(2) M2 = M1 S, M1 = M2 S,
n×n
Theorem 4.1 (Recovery theorem). Let Ω be a simply
where S ∈ R is the matrix representing a connected open subset of Rn . Let Σ ⊂ Ω be a locally
symmetry with respect to the (n − 1)-dimensional finite polyhedral set satisfying the angle condition
plane containing the facet we are considering (thus (3) on every (n − 2)-dimensional facet. Then
det S = −1). In particular we have (i) there exists a rigid map u such that Σ = Σu is
det M2 = − det M1 , the singular set of Du;

602 Notices of the AMS Volume 57, Number 5

 (ii) u is uniquely determined once we fix the value represented by a flat origami. However, the Dirichlet
y0 = u(x0 ) and the Jacobian J0 = Du(x0 ) at a point boundary condition is, in principle, incompatible
x0 ∈ Ω \ Σ. with the first-order differential system, as already
The theorem means that if we construct a mentioned in the introduction.
set Σ satisfying the conditions just mentioned, We first discuss the case n = 2. As we already
there always exists a map u such that Σ = Σu . said, only a fractal construction can ensure the
Moreover, the map u is uniquely determined up to boundary condition u = 0. Figure 6 gives an idea
compositions (on the left) with rigid motions of of the construction that we are going to develop. If
the whole space. Ω is a square, we can divide it in infinitely many
In the literature about flat folding (see, for squares, which are smaller and smaller, while we
instance, [2], [4]), the problem of reconstructing approach the boundary of Ω. Then it is enough to
an origami from its singular set (which is also consider a base map u0 defined on one of these
called crease pattern) has already been considered. squares. This map will be translated, rotated, and
However, a more strict definition of origami should rescaled to fit any other square. To assure that the
be used to treat physical origami. In a physical gluing of the squares gives a continuous map, we
origami the interpenetration problem is addressed, need the base map to have prescribed recursive
i.e., the problem that physically a noninjective map boundary conditions. More precisely we require
cannot be folded without cutting the paper. To be that, on the right-hand side of the base square,
more precise, since the thickness of the paper can the map is defined so that it reproduces twice the
anyway be neglected, we admit noninjective maps values of the left-hand side, rescaled by half; i.e.,
as long as they are the limit of injective ones.
u0 (1, y) = u0 (0, 2y) for y ∈ [0, 1/2],
This restriction makes the angle condition not
sufficient to guarantee the global reconstruction u0 (1, y) = u0 (0, 2y − 1) for y ∈ [1/2, 1];
of a physical origami. An example is given by the while on the upper and lower sides we only need
singular set in Figure 11 which satisfies the angle periodic boundary conditions:
condition, hence it is the singular set of some map
u which, however, cannot be folded without cutting u0 (x, 0) = u0 (x, 1) for x ∈ [0, 1].
the paper. In this more strict setting it is shown If the map assumes at least once the value 0 on
(see [4]) that the problem of deciding whether a every square in the net, then by its Lipschitz conti-
crease pattern can be flat-folded is NP-hard. nuity (every rigid map is 1-Lipschitz continuous)
it can be extended to the boundary ∂Ω with the 0
value.
The base module can thus be found by using the
angle condition (3) and the recovery theorem 4.1. In
this way we are able to find solutions to the above
problem. More generally we might consider linear
boundary conditions, such as u(x, y) = (λx, µy)
on ∂Ω. The corresponding singular set is shown in
Figure 12.
Notice that we can easily pose the problem in
higher dimension by considering n-dimensional flat
origami, namely hyper-origami. Again we subdivide
the domain in smaller and smaller cubes (or
parallelepipeds). On the base domain we need
Figure 11. A singular set of a rigid map which to find an n-dimensional origami which fulfills
cannot actually be folded without cutting the the recursive boundary condition. For n = 3
paper. we can imagine this operation as the folding
of a rigid three-dimensional box. In Figure 13,
we present the “instructions” for folding the
base parallelepiped to obtain the zero boundary
Fractal Construction condition. By gluing together the base modules we
We now turn our attention to the more difficult obtain the 3-dimensional singular set.
problem of finding a solution to the following Finally we note that one can solve (4) with
implicit partial differential system symmetric rigid maps, i.e., maps u which are
gradients of functions v : Ω ⊂ R2 → R. The problem
(
Du(x) ∈ O(n) for a.e. x ∈ Ω
(4) (4) when n = 2 then becomes
u(x) = 0 for x ∈ ∂Ω
( 2
in the case when Ω is a rectangle in R2 (or D v(x) ∈ O(2) for a.e. x ∈ Ω
parallelepiped in Rn ). Now we know that a solution is Dv(x) = 0 for x ∈ ∂Ω.

May 2010 Notices of the AMS 603

 Figure 12. A fractal construction for the linear boundary data.

It turns out (see [11]) that the corresponding angle Lipschitz solution may exist. Curiously enough, if
condition (3) is, in this case, much more restrictive there is a solution of (4) or more generally of (5),
in the sense that at every vertex of Σu = ΣDv exactly then, in general, there are infinitely many solutions.
four consecutive angles α+ , α− , β+ , β− meet with Let us review some of the techniques for solving
π such equations and see how the constructions we
α + + β+ = π and α − = β− = .
2 have proposed for origami may shed light on some
We call this property “second-order angle condition”. issues related to implicit PDEs.
The singular set represented in Figure 12 satisfies Let us start with the scalar case m = 1 which
the second-order angle condition, in contrast with has received considerable attention since the work
the singular set of Figure 6. of E. Hopf, notably by Kruzkov, Lax, and Oleinik.
In this case the equation
Implicit Partial Differential Equations: F(x, u, Du) = 0
Conclusion
What we have done so far is to find explicit solutions is sometimes called Hamilton-Jacobi equation.
to Dirichlet problems, as in the section “Fractal Crandall-Lions further developed in this context
Construction”. This kind of problem belongs the important tool of viscosity solution. The beauty
to a wider class, called implicit partial differential and the importance of the viscosity method is that,
equations. Namely we look for Lipschitz continuous at the same time, it gives existence of solutions
solutions to and a way of selecting one among the infinitely
(
F (x, u (x) , Du (x)) = 0 for a.e. x ∈ Ω many solutions of (5). It is, however, of crucial
(5) importance in this context that the Hamiltonian
u (x) = ϕ (x) for x ∈ ∂Ω,
F be convex in the variable Du. For example, the
where u, ϕ : Ω ⊂ Rn → Rm and F : Ω × Rm × Rm×n → eikonal equation
R (also called Hamiltonian). (
|Du (x)| = 1 for a.e. x ∈ Ω
Note that, as in (4), the problem is apparently ill
posed, since we want to solve a first-order PDE and u (x) = 0 for x ∈ ∂Ω
at the same time prescribe full Dirichlet condition
has, for convex domains Ω, the viscosity solution
on the boundary. Indeed, except for very special
cases, no C 1 solution is to be expected, and only u (x) = dist (x, ∂Ω) ,

604 Notices of the AMS Volume 57, Number 5

Figure 13. A “hyper-origami”, precisely a three-dimensional flat origami which solves the Dirichlet
problem (4) for n = 3 with zero boundary condition. At the top an elementary brick is represented.
It has a self-similar property which allows recursively gluing it together with half-sized
homothetic copies of itself to obtain the pattern represented at the bottom.

although, as said earlier, there are infinitely many (ii) The Baire category method is the other tool
other solutions, one of them being and has been introduced by Cellina, Bressan-Flores,
and De Blasi-Pianigiani for the scalar case and
u (x) = − dist (x, ∂Ω) .
generalized to the vectorial context by Dacorogna-
However, as soon as the Hamiltonian F is Marcellini.
nonconvex (with respect to the last variable) and We also point out that, in the scalar case
still in the scalar case m = 1, or in the vectorial m = 1, and for Hamiltonians F, not necessarily
context n, m ≥ 2 (as for origami), the viscosity convex, that do not depend on lower-order terms
theory does not apply. Two other main theories (i.e., F(Du) = 0), explicit solutions can easily be
have been developed to deal with these cases, and, obtained and are known as “pyramids”, following
although different, they give essentially the same the work of Cellina and Friesecke. This was a source
results (for detailed references see [9]). However, of inspiration to construct explicit solutions of
both theories are purely existential and do not give (4) to Cellina-Perrotta and to the work presented
a simple criterion for selecting one solution among here. Contrary to the finite element method used in
the infinitely many as did the viscosity theory. Let numerical analysis to build approximated solutions,
us briefly discuss both of them. here, by means of piecewise affine maps with
(i) The theory of convex integration proposed appropriate grid, we are able to exhibit exact
by Gromov is one of them. It was originally solutions.
introduced to give a new proof of the celebrated Since the two approaches described above are
Nash-Kuiper theorem on isometric embedding, only existential, it is then an important issue to
which is closely related to the problem discussed in single out one solution among the infinitely many.
the present paper. The theory of convex integration It is what we achieved by means of origami for the
has been developed in the context of implicit partial special problem (4). There are many other problems
differential equations by Müller-Sverak and others. related, for example, to geometry, optimal design,

May 2010 Notices of the AMS 605

 or nonlinear elasticity where an actual explicit [11] B. Dacorogna, P. Marcellini, and E. Paolini,
construction would be important. Functions with orthogonal Hessian, Differential and
Among them a problem which is significant in Integral Equations 23 (2010), 51–60.
[12] M. Gromov, Partial Differential Relations, Springer-
applications to microstructures in crystallographic
Verlag, Berlin, 1986.
models is the one of potential wells (see, for [13] K. Haga, Origamics, Mathematical Explorations
example, [3]). When we consider the case of two Through Paper Folding, World Scientific, Singapore,
wells, it consists in finding maps u whose gradient 2009.
satisfies [14] A. Heller, A giant leap for space telescopes, Science
( & Technology Review, March (2003), 12–18.
Du(x) ∈ SO(n)A ∪ SO(n)B, a.e. in Ω [15] D. Huffman, Curvature and creases: A primer on
u(x) = ϕ(x), on ∂Ω, paper, IEEE Trans. on Computers C-25 (1976), 1010–
1019.
where A and B are two given n × n matrices and [16] T. Hull, On the mathematics of flat origamis,
SO(n) is the set of special orthogonal matrices; Congressus Numerantium 100 (1994), 215–224.
i.e., those with determinant equal to +1. This is a [17] T. Hull, Project origami. Activities for exploring
natural generalization of the problem considered mathematics, A K Peters, Ltd., Wellesley, MA, 2006.
here (see (4)), since [18] T. Kawasaki, On the relation between mountain
creases and valley creases of a flat origami, Pro-
O(n) = SO(n)I ∪ SO(n)I− , ceedings of the 1st International Meeting of Origami
Science and Technology, Ferrara (H. Huzita, ed.),
where I is the identity matrix and I− = 1989, pp. 229–237.
diag(−1, 1, . . . , 1). [19] M. Kilian, S. Floery, Z. Chen, N. Mitra, A. Sheffer,
Finally, implicit partial differential equations, and H. Pottmann, Curve Folding, ACM SIGGRAPH,
because of the existence of infinitely many solu- 2008.
tions, are particularly challenging for numerical [20] N. H. Kuiper, On C 1 -isometric imbeddings I, Nederl.
Akad. Wetensch. Proc. Ser. A. 58 (1955), 545–556.
analysts. In the problem studied in the section [21] R. J. Lang, A computational algorithm for origami
“Fractal Construction”, one could imagine, as has design, Proceedings of the 12th Annual ACM Sym-
been suggested by one of the referees, finding posium on Computational Geometry, 1996, pp.
algorithms that mimic the origami construction. 98–105.
[22] J. Nash, C 1 -isometric imbeddings, Annals of
Mathematics 60 (1954), 383–396.
References
[1] R. C. Alperin, A mathematical theory of origami
constructions and numbers, New York J. Math. 6
(2000), 119–133.
[2] E. M. Arkin, M. A. Bender, M. A. Demaine, et al.,
When can you fold a map?, Computational Geometry
29 (2004), 23–46.
[3] J. M. Ball and R. D. James, Proposed experimental
tests of a theory of fine microstructure and the two
wells problem, Phil. Trans. Royal Soc. London A 338
(1991), 389–450.
[4] M. Bern and B. Hayes, The complexity of flat origami,
Proceedings of the 7th Annual ACM-SIAM Symposium
on Discrete Algorithms, 1996, pp. 175–183.
[5] B. Casselman, If Euclid had been Japanese, Notices
Amer. Math. Soc. 54 (2007), 626–628.
[6] D. A. Cox, Galois Theory, Wiley-Interscience,
Hoboken, NJ, 2004.
[7] M. G. Crandall, H. Ishii, and P. L. Lions, User’s
guide to viscosity solutions of second order par-
tial differential equations, Bull. Amer. Math. Soc. 27
(1992), 1–67.
[8] B. Dacorogna and P. Marcellini, General existence
theorems for Hamilton-Jacobi equations in the scalar
and vectorial case, Acta Mathematica 178 (1997),
1–37.
[9] B. Dacorogna and P. Marcellini, Implicit Partial
Differential Equations, Progress in Nonlinear Dif-
ferential Equations and Their Applications, vol. 37,
Birkhäuser, 1999.
[10] B. Dacorogna, P. Marcellini, and E. Paolini,
Lipschitz-continuous local isometric immersions:
Rigid maps and origami, J. Math. Pures Appl. 90
(2008), 66–81.

606 Notices of the AMS Volume 57, Number 5