See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222319477

On the folding of graphs-theory and application

Article  in  Chaos Solitons & Fractals · October 2009

DOI: 10.1016/j.chaos.2009.01.038

CITATIONS READS

5 745

2 authors:

S.I. Nada E.H. Hamouda

Menoufia University Beni Suef University
26 PUBLICATIONS   83 CITATIONS    13 PUBLICATIONS   17 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

New types of winning strategies via compact spaces View project

soft ternary semigroups View project

All content following this page was uploaded by E.H. Hamouda on 01 March 2018.

The user has requested enhancement of the downloaded file.

 ARTICLE IN PRESS
Chaos, Solitons and Fractals xxx (2009) xxx–xxx

Contents lists available at ScienceDirect

Chaos, Solitons and Fractals

journal homepage: www.elsevier.com/locate/chaos

On the folding of graphs-theory and application

S.I. Nada a,*, E.H. Hamouda b
a
Department of Mathematics, Faculty of Science, Minoufya University, Minoufya, Egypt
b
Department of Basic Sciences, Faculty of Industrial Education, Beni-Suef University, Egypt

a r t i c l e i n f o a b s t r a c t

Article history: In this article, we recall the definition of graph folding in the sense of El-kholy and El-
Accepted 30 January 2009 Esawy [El-kholy E, El-Esawy A. Graph folding of some special graphs. J Math Stat
Available online xxxx 2005;1(1):66–70.] and study the limit of folding of some graphs. Also we define the folding
number and give a formula to compute the number of folding maps f: Pn ? Pn, where Pn is a
path with n edges. We also discuss application of folding from Orgami to buckling and high
energy physics.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction and preliminaries

Most folding problems are attractive from a pure mathematical standpoint, for the beauty of the problems themselves.
The folding problems have close connections to important industrial applications. Linkage (planar graphs) folding has appli-
cations in robotics and hydraulic tube bending. Paper folding has applications in sheet-metal bending, packaging, and air-bag
folding [2]. Following the great Soviet geometer A. Pogorelov [11], M. El Naschie used folding to solve difficult problems re-
lated to shell structures in civil engineering and aero space design, namely buckling instability [9].
Consider a finite number of unit bars joined together at the ends where they are free to turn. Some meter sticks are con-
structed with this kind of structure as shown in Fig. 1. Perhaps you have seen such a meter stick and noted that it can be
folded to become
The meter stick can be regarded as a physical model of the path Pn. After folding the meter stick, it becomes a physical
model for the complete graph K2 [3]. There are graphs which cannot be folded, for example the complete graph Kn [7]. Con-
cepts of graph folding and the folding number were introduced in [3], and the authors determined the folding numbers of
some graphs. David R. Wood proved in an unpublished work that the folding number of a graph G equals the chromatic num-
ber of G [13]. The concept of folding of some types of graphs is introduced by El-Ghoul [5,6].
We follow [1] for graph theoretical terminology and notation not defined here. A graph is a topological space which con-
sists of a collection of points, called vertices, and a collection of edges. Each edge is homeomorphic to an interval of the real
line and joins two distinct vertices, i.e. a graph is a one-dimensional CW-complex. A graph is finite if it has only finitely many
vertices and edges or, equivalently, if it is a compact space. A graph is a locally contractible space: any neighborhood of any
point contains a contractible neighborhood of that point.
Throughout this paper, a graph G = (V, E) always means a finite and a simple connected graph (without loops and parallel
edges), where V = V(G) is the vertex set and E = E(G) is the edge set.
The results of this paper concern folding of one-dimensional objects in two dimensions: planar graphs. In section two, we
recall the definition of graph folding in the sense of [5] and study the limit of folding of some graphs. In section three, we
define the folding number and give a formula to compute the number of folding maps f: Pn ? Pn, where Pn is a path with
n edges.

* Corresponding author.
E-mail address: snada@qu.edu.qa (S.I. Nada).

0960-0779/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chaos.2009.01.038

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
2 S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx

Fig. 1. Folding of a meter stick.

2. The concept of folding

First of all, we recall the definition of folding of a graph in the sense of [8] and introduce examples of the folding of some
graphs.
Definition 2.1. Let G = (V, E) be a graph. The folding of G into itself is a continuous function f: G ? G such that, for each v 2 V,
f(v) 2 V and for each e 2 E, f(e) 2 E.

Example 2.2

(i) Let G be a simple graph as in Fig. 2. Let f: G ? G be a folding map defined as follows:
f : v 1 ! v 4 ; ðv 1 ; v 2 Þ ! ðv 2 ; v 4 Þ;
ðv 1 ; v 3 Þ ! ðv 3 ; v 4 Þ
Where any omitted vertex (edge) will be mapped onto itself.
(ii) Let H be a graph with a loop as in Fig. 2. One can define the folding map g: H ? H by:
g : v 1 ! v 2 ; ðv 1 ; v 2 Þ ! ðv 2 ; v 2 Þ;
ð v 1 ; v 3 Þ ! ðv 2 ; v 3 Þ
(iii) A complete graph G has no non-trivial folding [6].
(iv) Any tree T can be folded into itself by a sequence of foldings onto an edge, i.e. the limit of folding of T is an edge [7].

Lemma 2.3. Limit of the folding of an n-cycle depends on the number of edges n.

Proof. Let Cn be an odd cycle, and n > 2. If n = 3, then C3 is isomorphic to a complete graph. As shown in 2.2.
(iii), C3 has no non-trivial folding. So the limit of foldings is isomorphic to C3.
Consider an odd cycle Cn, n > 3, as in Fig. 3. We can define a folding map f: Cn ? Cn as follows:

Fig. 2. Folding of some graphs.

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx 3

Fig. 3. n-cycle (odd) graph.

f ðv i Þ ¼ v i ; i ¼ 1; 3; 5; . . . ; n:
f ðv i Þ ¼ v iþ1 ; i ¼ 2; 4; 6; . . . ; n  1:
and
f : ðv 1 ; v 2 Þ ! ðv 1 ; v 3 Þ;
ðv 2 ; v 4 Þ ! ðv 3 ; v 5 Þ;

ðv n3 ; v n1 Þ ! ðv n2 ; v n Þ:
It is clear that the edge (vn1, vn) must be folded to a loop; this means that the cycle Cn has a loop (contradiction). Hence an n-
cycle has a trivial folding when n is odd. If n is even, one can define a folding map f: Cn ? Cn so that f(Cn) is a tree with n/2
edges. The graph f(Cn) can be subjected to a series of folding until we obtain an edge, as shown 2.2(iv). Therefore limit of
foldings of a cycle C2n is an edge. h

Lemma 2.4. Let G be a graph contains no odd cycles. Then the limit of folding of G is an edge.

Proof. Consider G is a graph contains m even cycles, and f1: G ? G is a folding map, such that f1(G) has no terminal edges, i.e.
f1(G) consists of even cycles. By Lemma 2.3, we can define a series of folding maps by:

f2: f1(G) ? f1(G), such that f2(C2n) is a tree with n edges,

f3: f2(f1(G)) ? f2(f1(G)), such that f2(f1(G)) has no terminal edges,
–––
–––
–––
fm+1: fm(fm 1(. . .f1(G))) ? fm(fm1(. . .f1(G))), such that fm+1(fm(fm1(. . .f1(G)))) is a tree. So the limit of a series of folding of G is
an edge. h

Lemma 2.5. Let G be a multiple graph contains no C2n1 cycle. Then the limit of folding of G is a loop.

Proof. Let G be a graph with parallel edges and a loop. One can define a sequence of folding maps, so that f1(G) is a graph
without parallel edges and terminal edges. As in the proof of Lemma 2.3, we repeat this process until get a graph H, see Fig. 4.
As shown in Example 2.2, the limit of folding of H is a loop. Therefore a multiple graph contains no n-cycle, n is odd, can be
folded onto a loop. h

Fig. 4. A tree with a loop.

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
4 S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx

In the following example, we show that a multiple graph with odd cycles can be folded onto a loop if each odd cycle has a
loop at one of its vertices.
Example 2.5. Let G be a multiple graph contains n-cycles, n is odd, as in Fig. 5. One can define a sequence of folding maps of G
into itself such that the limit of folding is a loop as follows:
f1 : v 1 ! v 4 ; ðv 1 ; v 2 Þ ! ðv 2 ; v 4 Þ;
ðv 1 ; v 3 Þ ! ðv 3 ; v 4 Þ:
f2 : v 2 ! v 3 ; ðv 2 ; v 4 Þ ! ðv 3 ; v 4 Þ;
ðv 2 ; v 3 Þ ! ðv 3 ; v 3 Þ:
f3 : v 4 ! v 3 ; ðv 3 ; v 4 Þ ! ðv 3 ; v 3 Þ:

3. The folding number of Pn

In [3] the authors have defined the folding number of a graph G, denoted by f(G), by the minimum order of a complete
folding of G. They determined the folding number of wheels and fans.
In this section we define the folding number, F(G), to be the order of the set of folding maps of G into itself. Also we com-
pute the folding number of Pn, a path with n edges.
Theorem 3.1
FðP n Þ ¼ n þ 1; if n is odd;
FðP n Þ ¼ n þ 2; if n is ev en:

Proof. By the induction on the number of edges n. When n = 1, F(P1) = 2 since there are two trivial folding maps:

f0 :P1 ! P1 ; so that f 0 ðv i Þ ¼ v i ; i ¼ 0; 1
and f 1 : P1 ! P1 ; such that f 1 ðv 0 Þ ¼ v 1 and; f 1 ðv 1 Þ ¼ v 0 :
Now we consider the cases n = 2, 3. See Fig. 6.
For n = 2, there are four folding maps:

Fig. 5. Folding of a multiple graph.

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx 5

Fig. 6. Folding numbers.

 The trivial maps f0(vi) = vi and f1: P2 ? P2 such that f1(v0) = v2, f1(v2) = v0. P2 can be folded onto P1 by the folding maps f2
and f3 shown in Fig. 6. This means that F(P2) = 4, and the result is true.

Consider the case n = 3, we have the following folding maps:

 The trivial maps f0(vi) = vi and f1: P3 ? P3 so that f1(vi) = v3I for i = 0, 1, 2, 3.
 P3 can be folded onto P2 by the folding maps f2 and f3 shown in Fig. 6. This implies that F(P3) = 4.

Now we proceed to the general case and assume n is even. It follows from the induction hypothesis that
FðPn1 Þ ¼ ðn  1Þ þ 1 ¼ n:
Let Pn be obtained from Pn1 by adding an edge. Since n is even, then Pn has the symmetric vertex vn/2 (i.e. the number of
edges before vn/2 equals the number of edges after). So we can fold Pn onto Pn/2 by defining two additional maps. Thus
F(Pn) = n + 2. Similarly if n is odd, then
FðPn1 Þ ¼ ðn  1Þ þ 2 ¼ n þ 1:
Because Pn does not have a symmetric vertex, it is clear that
FðPn Þ ¼ FðPn1 Þ ¼ n þ 1:
This completes the proof. h
If n is large, the set of folding maps of Pn into itself may be huge and so we should try a relatively much smaller set. A
canonical way is to put an equivalence relation on the set of folding maps and then consider the quotient set.
Definition 3.2. Let f: Pn ? Pn and g: Pn ? Pn be folding maps of Pn into itself. The maps f and g are said to be equivalent,
written f  g, if and only if f(Pn) is isomorphic to g(Pn).
It is a trivial matter to verify that the relation f  g is reflexive, symmetric and transitive. Thus the folding maps Pn ? Pn
divide into equivalence classes. Using modulo of this relation, we obtain the quotient set, denoted by [Pn]. The class repre-
sented by f is denoted by [f].

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
6 S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx

It is clear that every class contains only two elements of F(Pn), so the number of folding maps determines the number of
classes and vice versa. We shall consider examples of the quotient set [Pn].

 The set [P1] contains only one class [f0], so that f0: P1 ? P1 is the trivial folding map f0 (vi) = vi.
 As in Fig. 6, the set [P2] consists of two classes represented by the trivial map f0 and f2: P2 ? P2, where f2(P2) is isomorphic
to P1.

From now on, [fi] denote to the class of folding maps Pn ? Pn such that fi (Pn) is isomorphic toPi. The following theorem
gives us a formula to compute the number of classes of [Pn].
Theorem 3.3. If Pn denotes the number of classes of [Pn], then for n = 1, 2, . . .
X
n X
n
P2n ¼ ½f2ni  and P 2n1 ¼ ½f2ni :
i¼0 i¼1

Proof. In case n = 1, the set [P1] contains one class [f1] represented by the trivial folding map f1, and [P2] consists of two clas-
ses [f2] and [f1]. Now assuming that the result is true for n  1, that is

X
n1 X
n1
P2n2 ¼ ½f2n2i  and P2n3 ¼ ½f2n2i :
i¼0 i¼1

By Theorem 3.1, we have

FðP2n1 Þ ¼ FðP2n2 Þ ¼ 2n:
This means that [P2n1] and [P2n2] has the same number of classes. It follows from the induction assumption that
X
n
P2n1 ¼ ½f2ni :
i¼1

It remains to show that the result is true for P2n. Since

FðP2n Þ ¼ 2n þ 2 ¼ FðP2n1 Þ þ 2:
Then [P2n] has the same classes of [P2n-1] with additional class represented by the trivial folding map f2n: P2n? P2n so that
f2n(P2n) is isomorphic to P2n. This implies
X
n X
n
P2n ¼ P2n1 þ ½f2n  ¼ ½f2ni  þ ½f2n  ¼ ½f2ni :
i¼1 i¼0

So the result is true for every n, this completes the proof. h

Now let us look at some calculations of the classes of the folding maps of Pn for n = 1, 2, . . ., 10 shown in the table below
and see what Theorem 3.3 says about the number of folding maps of Pn into itself?

1. As we stated above, the number of classes Pn determines F(Pn). For example, P10 = 6 implies that F(P10) = 12, which agrees
with Theorem 3.2.
2. If n is odd, then Pn = Pn1.
3. If n is even, then Pn = Pn+1.
4. For n > 2, there no a folding map f: Pn ? Pn such that f(Pn) is isomorphic to P1.

[Pn]
1 [f1]
2 [f2] + [f1]
3 [f3] + [f2]
4 [f4] + [f3] + [f2]
5 [f5] + [f4] + [f3]
6 [f6] + [f5] + [f4] + [f3]
7 [f7] + [f6] + [f5] + [f4]
8 [f8] + [f7] + [f6] + [f5] + [f4]
9 [f9] + [f8] + [f7] + [f6] + [f5]
10 [f10] + [f9] + [f8] + [f7] + [f6] + [f5]

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
 ARTICLE IN PRESS
S.I. Nada, E.H. Hamouda / Chaos, Solitons and Fractals xxx (2009) xxx–xxx 7

4. Conclusion

In the present short paper we considered various definitions and theorems related to the limit folding of graph. Although
the study seems to be purely mathematical, in general folding has numerous roles to play in engineering and physical appli-
cations [4,8–12]. Starting from the Japanese art of folding paper, Origami [4], Yoshimura, Pogorelov and El Naschie solved the
difficult nonlinear problem of elastic shell buckling [9]. In addition, folding is related to Coxeter polytopes and therefore to E-
infinity and E8. Consequently, it connects the theory of transfinite dimensions and Cantorian manifold to high energy physics
[10,11]. For instance we can easily understand that for n equal infinity topology we have three non-equal fundamental trans-
finite dimensions, namely the transfinite small inductive dimension, trind X, the transfinite large inductive dimension, trInd
X, and the transfinite covering dimension trdim X. Choosing our basic Cantorian blocks to be random Cantor sets, the three
dimensions can be readily calculated and found, respectively, to be

DF ¼ 4  k; DT ¼ 4; DH ¼ 4 þ U3 ;
pffiffi
where K = U3(1  U3) and U ¼ 521.
This is exactly the well known result of E-infinity theory [8,10].
For non-transfinite topology we have DF = DT = DH = 4 as should be in Euclidean smooth geometry. Note also that
DH = 4 + U3 is the covering dimension of a Hilbert four dimensional cube [8,10].

References

[1] Bondy JA, Murty USR. Graph theory with applications. New York: Macmillan; 1976.
[2] Erik D. Demaine: folding and unfolding, Ph.D. thesis, Waterloo University, Canada; 2001.
[3] Gervacio SV, Guerrero RC, Rara HM. Folding wheels and fans. Graph Combinator 2002;8:731–7.
[4] Hull T, editor. Origami papers, Massachusetts; 2002.
[5] El-Ghoul M. Folding of fuzzy graphs and fuzzy spheres. Fuzzy Set Syst 1993;58:355–63.
[6] El-Ghoul M. Folding of fuzzy torus and fuzzy graphs. Fuzzy Set Syst 1996;80:389–96.
[7] El-kholy E, El-Esawy A. Graph folding of some special graphs. J Math Stat 2005;1(1):66–70.
[8] Marek-Crnjac L. A short history of fractal – Cantorian spacetime. Chaos, Solitons & Fractals 2008. doi:10.1016/j.chaos.2008.10.007.
[9] El Naschie MS. Stress, stability and chaos in structural engineering. New York: McGraw-Hill; 1990.
[10] El Naschie MS. The theory of Cantorian space–time high energy particle physics (an informal review). Chaos, Solitons & Fractals 2008. doi:10.1016/
j.chaos.2008.09.059.
[11] Pogorelov AV. Differential geometry. Groningen, The Netherland: Noordhoff; 1959.
[12] Urysohn P. Les multiplicites cantoriennes. Fund Math 1925;30:137.
[13] Wood DR. Folding = colouring; 2008. 0802.2467v1.

Please cite this article in press as: Nada SI, Hamouda EH. On the folding of graphs-theory and application. Chaos, Solitons &
Fractals (2009), doi:10.1016/j.chaos.2009.01.038
View publication stats