A Project Report on

CLASSIFICATION THEOREM OF SIMPLE,

REGULAR POLYHEDRON IN R3

Submitted By:
Rivu Bardhan,
MSM17006,
M.Sc. Mathematics, 3rd Semester,
Tezpur University,
Tezpur, Napaam, Assam-784028.

Submitted to:

Department of Mathematical Sciences,

Tezpur University,
Napaam, Tezpur, Assam-784028.
 "Now all these bodies we must conceive as being so small that each
single body in the several kinds cannot for its smallness be seen by us at all; but when many
are heaped together, their united mass is seen..."

-Plato (424 BC-348 BC)

i
ACKNOWLEDGEMENTS

I would like to thank my project advisor Prof. Debajit Hazarika of the Department of
Mathematical Sciences at Tezpur University. The door to Prof. Hazarika’s office was
always open whenever I ran into a trouble spot or had a question about my project or
writing. He consistently allowed this project to be my own work, but steered me in the
right the direction whenever he thought I needed it.

Rivu Bardhan
September 23, 2019
 Abstract

In this project we would like to study the basic concepts of Alge-

braic Topology. In other words with the help of finite simplicial
complexes we obtain homolgy groups of the polyhedrons and prove
their topological invariance. This leads us to the proof of the path
breaking Euler Poincaré Theorem. Euler’s celebrated identity for
polyhedrons(1752) comes as a corollary of this theorem. As a fi-
nal conclusion we observe that these results finally leads us to the
classification of the simple, regular polyhedrons in R3.
Contents

1 Introduction 2

2 Finite Simplicial Complex 5

2.1 Simplicial Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Polyhedra and Triangulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Simplicial Homology 12
3.1 Orientation of Simplicial Complexes . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Simplicial Chain Complex and Homology . . . . . . . . . . . . . . . . . . . . 14
3.3 Some Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 The general structure of Hq . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4 The Euler Poincaré Theorem 21

4.1 The Euler Poincaré Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2 Euler’s Theorem and Polyhedron . . . . . . . . . . . . . . . . . . . . . . . . 22

5 Classification Theorem for Simple, Regular Polyhedrons in R3 24

5.1 Classification Theorem for Simple, Regular Polyhedrons in R3 . . . . . . . . 24
5.2 Further Generalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Bibliography 27

1
Chapter 1

Introduction

In three-dimensional space, a Platonic solid is a regular, convex polyhedron. It is constructed

by congruent (identical in shape and size), regular (all angles equal and all sides equal)
polygonal faces with the same number of faces meeting at each vertex. Five solids meet
these criteria:

Figure 1.1: Tetrahedron

Figure 1.2: Cube

2
 Figure 1.3: Octahedron

Figure 1.4: Dodecahedron

Figure 1.5: Icosahedron

Geometers have studied the Platonic solids for thousands of years. They are named for the
ancient Greek philosopher Plato who hypothesized in his dialogue, the Timaeus, that the
classical elements were made of these regular solids.

Now, let,

V = no of vertices of the polyhedron,

E = no of edges of the polyhedron and
F = no of faces of the polyhedron.

Then
For Figure 1.1 V = 4, E = 6, F = 4: V − E + F = 2
For Figure 1.2 V = 8, E = 12, F = 6: V − E + F = 2

3
For Figure 1.3 V = 6, E = 12, F = 8: V − E + F = 2
For Figure 1.4 V = 20, E = 30, F = 12: V − E + F = 2
For Figure 1.5 V = 12, E = 30, F = 20: V − E + F = 2

So it is quite obvious from the discussion that these figures follow certain homogeneous
property. Leonhard Euler (1707-1783) was the first one to observe this property in these 5
polyhedrons. In fact he has given a generalized result which is of far more importance not
only in the branch of Algebraic Topology but also in many other branches of Mathematics
(For example: Graph Theory).

Euler’s following theorem has remained a celebrated result through the centuries and
contributed a great lot to further branches of mathematics namely,

Theorem 1.0.1. (Euler’s Theorem) If S is any simple Polyhedron with V vertices, E

edges and F faces then V − E + F = 2

Euler’s Theorem Every simple polyhedron satisfies the relation V − E + F = 2

We shall try to prove this theorem using the tools of the Algebraic Topology and provide a
more generalized result which will eventually lead us in proving famous Classification
Theorem of the simple, regular Polyhedrons in R3 famously hypothesized by Plato.

But, in order to do that we first turn our attention in developing the tools of Algebraic
Topology which will provide our foundation to prove the result.

4
Chapter 2

Finite Simplicial Complex

2.1 Simplicial Complex

The origin of the word "Simplicial" is Simplex which means fundamental not only in the
subject to topology but for many branches of mathematics. For example, Graph Theory.

Let Rn be the Euclidean space with the usual topology which is also eventually a vector
space of n dimension over R. A set S = {xi |i ∈ {1, 2, . . . , k}} ⊆ R is linearly independent
iff for arbitrary αi , i ∈ {1, 2, . . . , k} in R 3 ki=1 αi xi = 0 ⇒ αi = 0 ∀i ∈ {1, 2, . . . , k}.
P

We modify the the notion of the linear independence to define geometric independence.

Definition 2.1.1. A set A={ai |i ∈ {0, 1, 2, . . . , k}}, k > 1 of Rn is said to be geometrically

independent iff the set S={ai − a0 | i ∈ {1, 2, . . . , k}} is linearly independent.

Example 2.1.1. Let A = {a0 , a1 , a2 } ⊆ R2 is a geometrically independent iff the set S =

{a1 − a0 , a2 − a0 } is linearly independent.(F igure 2.1,F igure 2.2)

Definition 2.1.2. A subset H of Rn is said to be a hyperplane of dimension k 3 k 6 n if ∃

a k-dimensional vector subspace V k ∈ Rn and a vector x ∈ Rn such that H = x + V k

Proposition 2.1.1. Let A={ai |i ∈ {0, 1, . . . , k}}, k > 1 of Rn

(i) A is geometrically independent iff all the points of A do not lie on a hyperplane of
dimension k − 1.

5
 Figure 2.1: Geometrically independent
Figure 2.2: Geometrically dependent

(ii) A is geometrically independent iff for arbitrary αi ∈ Rn ,

Pk
(i) i=0 α i ai = 0
Pk
(ii) i=0 αi = 0

implies αi = 0 ∀ i∈ {0, 1, 2, . . . , k}.

(iii) If A is geometrically independent then there exists a hyperlane which passes through
all of the points in A.

(iv) Let A be geometrically independent then each point in the hyperplane passing through
all of the elements of A can be expressed uniquely as h= ki=0 αi ai where ki=0 αi = 1.
P P

Proof. Proofs are almost straight forward from the definitions.

Definition 2.1.3. A={ai |i ∈ {0, 1, . . . , k}}, k > 1 of Rn be a geometrically independent

set and h be a point in the hyperplane passing through elements of A then We can write
h = ki=0 αi ai where ki=0 αi = 1 (Clear from Proposition 2.1.1 ).
P P

The real numbers {αi |i ∈ {0, 1, 2, . . . , k}} which are are uniquely determined by A are called
barycentric coordinates of h with respect to A.

Definition 2.1.4. A={ai |i ∈ {0, 1, . . . , k}}, k > 1 of Rn be a geometrically independent set.

Then the k-dimensional geometric complex or k-simplex spanned by the set A, denoted
as σ k = {x = ki=0 ai αi | ki=0 αi = 1, αi > 0}.
P P

Example 2.1.2. The 0-simplex is trivially a point.

The 1-simplex generated by the set {a0 , a1 } is simply a line.

6
 Figure 2.3: 1-simplex

The 2-simplex happen to be a triangle.

Figure 2.4: 2-simplex

Proposition 2.1.2. A={ai |i ∈ {0, 1, . . . , k}}, k > 1 of Rn be a geometrically independent

set. Then the simplex σ k =< a0 , a1 , . . . , ak > is the convex hull of A.

Proof. Please Refer page 92 of [1] Proposition 3.1.4.

Definition 2.1.5. Let σ q ,σ p , p ≤ q ≤ n be two simplexes in Rn . We say that σ p is a

p-simplex of σ q if each vertex of σ p is also a vertex of σ q .

Definition 2.1.6. A Simplicial Complex K is a finite collection of simplexes of Rn , n

sufficiently large, which satisfies the following conditions:

(i) If σ ∈ K ⇒ all the faces of σ are in K.

T T
(ii) If σ, τ ∈ K ⇒ either σ τ = ∅ or σ τ is a common face to both σ and τ .

We also define the dimension of a simplicial complex K denoted by dim K=-1 if K=φ or
dim K=max{n ∈ N|σ n ∈ K}.

Example 2.1.3. K be a simplical complex shown in the F igure 2.5.

7
 Figure 2.5: Simplicial Complex

2.2 Polyhedra and Triangulation

S
Definition 2.2.1. Let K be a simplicial complex. Let |K| = σ∈K σ be the union of all
simplexes of K. Then |K| ⊆ Rn for some n ∈ N will be a topological subspace induced from
Rn . This space is called a geometric carrier of K. A subspace of Rn which is a geometric
carrier of some simplicial complex, is called a rectilinear polyhedron.

Definition 2.2.2. A topological space X said to be a polyhedron if there exists a simplicial

complex K such that |K| is homeomorphic to X. In this case X is said to be triangulable
and K is called a triangulation of X.

Example 2.2.1. (Cylinder) Consider the hollow prism (Fig 2.6), we break each rectangles
into two triangles.

Figure 2.6: Triangulation of cylinder

8
K= {< a0 >, < a1 >, < a2 >, < b0 >, < b1 >, < b2 >, < a0 , a1 >, < b0 , b1 >, < a1 , a2 >, <
b1 , b2 >, < a0 , a2 >, < b0 , b2 >, < a0 , b1 , b0 >, < a0 , a1 , b1 >, < a1 , b1 , b2 >, < a1 , a2 , b2 >, <
a0 , a2 , b2 >, < a0 , b0 , b2 >}

is a triangulation of the cylinder.

Proposition 2.2.1. Let U be a open set of Rn , n > 1 which is bounded and convex and
w∈ U . Then

(i) . Each half line L starting from w intersects the boundary of U at precisely one point.

(ii) . There is a homeomorphism of Ū with the unit disk Dn which carries the boundary of
U onto the unit sphere Sn−1 .

Proof. The proof is lengthy and out of our scope. Therefore we would like to skip it. An
out line of the proof given in page 96 of [1].

Example 2.2.2. (Discs) Consider D2 = {(x, y) ∈ R2 |x2 + y 2 6 1}.

Consider Cl(σ 2 ) where σ 2 =< a0 , a1 , a2 > is 2-simplex. Then |Cl(σ 2 )| is the ∆ with vertices
a0 , a1 , a2 . So D2 is homeomorphically equivalent to ∆ and ∆ is a triangulation of D2 . More
generally Dn (the n dimensional unit disc) is a simplicial complex of dimension n i.e Dn is a
polyhedron. It now follows from P roposition 2.2.1 every compact convex subset of Rn is a
polyhedron.

Definition 2.2.3. Let K be a k-dimensional simplicial complex. For each r, 0 6 r 6 k, let

K r denote the set of all simplexes of dimension6 r. Then K r is a simplicial complex which
is known as r-dimensional skeleton of K. The space |K r | will be a rectilinear polyhedron of
|K|.

Example 2.2.3. (Sphere) Let σ k be a k-simplex, k > 1 and K = Cl(σ k ) and K 0 be the
(k − 1) dimensional skeleton of K. Then K 0 consists of all proper faces of σ k . It can be
shown that |K 0 | is homeomorphic to Sk−1 = {(x0 , x1 , . . . , xk−1 ) ∈ Rk | k−1 2
P
i=0 xi = 1} which is

a (k − 1) dimensional unit sphere. This proves that every k-sphere Sk is a polyhedron and
the collection of a all proper surface of a (k + 1)-simplex gives a triangulation of Sk .

9
Example 2.2.4. (Möbius Band)

Figure 2.7: Triangulation of Möbius Band

As in Möbius Band the opposite edges of the rectangular sheet are joined reversely. Thus
the labelling of the f ig 2.7 indicates the reverse identification.

Example 2.2.5. (Torus)

Figure 2.8: Triangulation of Torus

This gives us the triangulation of a torus. From the labeling of the f ig 2.8 clearly states
that several 0-simplexes of the given simplicial complex into a new 0-simplex and several
1-simplex will collapse into new 1-simplex. So it will give us the triangulation of a torus.

Although triangulation of every topological space is often difficult and

in some cases impossible. There are some more interesting topological spaces with their
triangulations (for example: Klein’s Bottle, Real protective Plane) that are described in
[2],[1],[3],[4].

Definition 2.2.4. Let K be a simplicial complex. A simplicial Complex L is said to be a

subcomplex of K if L ⊂ K. The boundary of a simplicial complex K, denoted by ∂(K)
is defined by
∂(K) ={ τ | τ is a face of σ k ∈ K which belongs to some (k + 1) simplex of K}

10
For example, let K= Cl < a0 , a1 , a2 > so
∂K = {< a0 >, < a1 >, < a2 >, < a0 , a1 >, < a1 , a2 >, < a0 , a2 >}.

11
Chapter 3

Simplicial Homology

3.1 Orientation of Simplicial Complexes

Let σ n =< v0 , v1 , . . . , vn > be a n-simplex. Let the vertices of σ n are ordered by the
following:
v0 < v1 < · · · < vn .
This ordering determines certain direction among the vertices of σ n . We say that σ n is
positively oriented and write σ n as +σ n if we take σ n with the even permutation of the
given ordering. Otherwise we say σ n is negatively oriented and write it as −σ n .

Figure 3.1: Orientation of σ 2

Example 3.1.1. Let us take σ 2 =< v0 , v1 , v2 >. The figure 3.1. clearly indicates two
different orientation of σ 2 .

Definition 3.1.1. A simplicial complex is said to be oriented if each of its simplexes is

assigned to an orientation.

Definition 3.1.2. Let K be a simplicial complex and σ p , σ p+1 ∈ K such that their dimen-
sion differ by 1. With each such pair (σ p , σ p+1 ) we assign its Incidence Number denoted by

12
[σ p+1 , σ p ] as follows: If σ p is not a face of σ p+1 then [σ p+1 , σ p ] = 0. If +σ p =< v0 , v1 , . . . , vp >
and let v be the additional vertex in the σ p+1 then < v, v0 , v1 , . . . , vp > is either +σ p+1 or
−σ p+1 . That is we define

1, if < v, v0 , v1 , . . . , vp >= +σ p+1

p+1 p
[σ ,σ ] =
-1, if < v, v0 , v1 , . . . , vp >= −σ p+1


Example 3.1.2. Let the Cl(σ 2 ) be ordered with v0 < v1 < v2 . Then,
[σ 2 , < v0 , v1 >] = +1
[σ 2 , < v0 , v2 >] = −1
[< v0 , v1 >, < v2 >] = 0

Theorem 3.1.1. Let K be an oriented simplicial complex. If σ p−2 is a p − 2 face of σ p in

K, then
P p p−1 p−1 p−2
[σ , σ ][σ , σ ] = 0

Proof. Let v0 , v1 ..., vp−2 be the vertices of σ p−2 so that +σ p−2 =< v0 , v1 ..., vp−2 >. There are
only two p − 1 simplexes of K which yield the nonzero term in the summation above. They
are as follows:
σ1p−1 =< a, v0 , v1 , . . . , vp−2 > , σ2p−1 =< b, v0 , v1 , . . . , vp−2 >. Now we have four disjoint
cases:

Case 1 Suppose +σ1p−1 =< a, v0 , v1 , . . . , vp−2 > and +σ2p−1 =< b, v0 , v1 , . . . , vp−2 >
Then it is obvious that,
[σ p , σ1p−1 ] = −1 , [σ1p−1 , σ1p−2 ] = +1
[σ p , σ2p−1 ] = +1 , [σ1p−1 , σ2p−2 ] = +1
⇒ The summation is zero.

Case 2 Suppose +σ1p−1 =< a, v0 , v1 , . . . , vp−2 > and −σ2p−1 =< b, v0 , v1 , . . . , vp−2 >
Then it is obvious that
[σ p , σ1p−1 ] = −1 , [σ1p−1 , σ1p−2 ] = +1
[σ p , σ2p−1 ] = −1 , [σ1p−1 , σ2p−2 ] = −1
⇒ The summation is zero.

13
Case 3 Suppose −σ1p−1 =< a, v0 , v1 , . . . , vp−2 > and +σ2p−1 =< b, v0 , v1 , . . . , vp−2 >
Then it is obvious that,
[σ p , σ1p−1 ] = +1 , [σ1p−1 , σ1p−2 ] = −1
[σ p , σ2p−1 ] = +1 , [σ1p−1 , σ2p−2 ] = +1
⇒ The summation is zero.

Case 4 Suppose −σ1p−1 =< a, v0 , v1 , . . . , vp−2 > and −σ2p−1 =< b, v0 , v1 , . . . , vp−2 >
Then it is obvious that,
[σ p , σ1p−1 ] = +1 , [σ1p−1 , σ1p−2 ] = −1
[σ p , σ2p−1 ] = −1 , [σ1p−1 , σ2p−2 ] = −1
⇒ The summation is zero.

Hence the conclusion follows.

3.2 Simplicial Chain Complex and Homology

Let S˜q be the set of all oriented q-complexes in K.

Definition 3.2.1. Let 0 6 q 6 dim K and Z be an additive group of integer. Any map
f : S˜q −→ Z 3 f (−σ q ) = −f (σ q ) for each σ q ∈ S˜q is called a q-chain of K. Cq (K) =
{f | f is a q-chain} is the collection of all q-chains. If q > dim K then we define Cq (K) = 0

Remark 3.2.1. It is a trivial observation that Cq (K) is an abelian group under addition.

For each positively oriented σ q ∈ K define a q-chain σ¯q as follows:


if τq = +σ q



1,


σ¯q (τ q ) = -1, if τq = −σ q




0,

otherwise

Let Sq be the set of all positively oriented q-simplexes in K.

Zσ¯q , σ¯q ∈ Sq
L
Proposition 3.2.1. For each q > 0, Cq (K) =

14
 nq σ¯q for some nq ∈ Z clearly we are
P
Proof. If we show that each f ∈ Cq (K), f = σ q ∈Sq

done.
Suppose f (σ q ) = nq for some σ q ∈ K.
⇒ f (−σ q ) = −nq
Consider σq ∈Sq nq σ¯q ∈ Zσ¯ . Now suppose σ q ∈ K
P L q

If σ q ∈ Sq ,
⇒ ( σq ∈Sq nq σ¯q )(σ q ) = nq σ¯q (σ q ) = nq = f (σ q ).
P

If σ q ∈
/ Sq ,
⇒ −σ q ∈ Sq .
⇒ f (σ q ) = f (−(−σ q )) = −f (−σ q ), by Def inition 3.2.1.
⇒ f (σ q ) = − nq σ¯q )(−σ q ), from above proof.
P

⇒ f (σ q ) = ( nq σ¯q )(σ q )
P

This proves our claim.

nq σ¯q = mq σ¯q ⇒ ( (nq − mq )σ¯q )(σ q ) = 0 ⇒

P P P
To prove uniqueness suppose f =
nq − mq = 0 ⇒ nq = mq .
Since σ q is arbitrary,our proof is complete.

Definition 3.2.2. For each q, 0 < q < dimK, a homomorphism ∂q : Cq (K) −→ Cq−1 (K),
called as boundary homomorphism as follows:
Suppose σ q be a generator of Cq (K). Then,
q
X
∂q (σ q ) = [σ q , σiq−1 ]σiq−1 (3.1)
i=0

where σiq−1 runs through all positively oriented (q − 1) faces of σ q . Then We extend it over
whole Cq (K) linearly, i.e. We set
q
X X
q
∂q ( nq σ ) = nq ∂q (σ q ) (3.2)
i=0

where ∂q (σ q ) is defined by 3.1.

Lemma 3.2.2. For each q the composite homomorphism ∂q−1 ◦ ∂q : Cq (K) −→ Cq−2 (K) is
zero map.

15
Proof. Let σ q be a generator of Cq (K).
It is enough to prove that ∂q−1 ◦ ∂q (σ q ) = 0.
Now,
∂q−1 ◦ ∂q (σ q ) = ∂q−1 ( σq−1 ∈K [σ q , σiq−1 ])
P
i

= σq−1 ∈K [σ q , σiq−1 ]∂q−1 (σiq−1 )

P
i

= σq−1 ∈K [σ q , σiq−1 ]( σq−2 ∈K [σiq−1 , σjq−2 ])

P P
i j

Changing the order of summation and collecting the coefficients of σjq−2 shows that the above
term is equal to
q
, σiq−1 ][σiq−1 , σjq−2 ])σjq−2
P P
σjq−2 ∈K ( σiq−1 ∈K [σ

but, by theorem 3.1.1, ( σq−1 ∈K [σ q , σiq−1 ][σiq−1 , σjq−2 ]) = 0 for each σjq−2 , so ∂q−1 ◦ ∂q (σ q ) =
P
i

Let K be an oriented complex.

Definition 3.2.3. Zq := {zq ∈ Cq (K)|∂q (zq ) = 0} is the kernel of ∂q . zq ∈ Zq is called

q-cycle of K. Now Zq is clearly a subgroup of Cq (K).

Definition 3.2.4. Bq := {bq ∈ Cq (K)| ∃c0 ∈ Cq+1 (K) 3 ∂q+1 (c0 ) = bq } is the image of ∂q+1 .
bq ∈ Bq is called q-boundary of K. Now clearly Bq is image of ∂q+1 . So it is trivially a
subgroup of Cq (K).

Note 1. If dim K = n it is notable that there is a sequence

∂n+1 n ∂ ∂q+1 ∂q
C(K):. . . 0 −−−→Cn (K) −→ · · · →Cq+1 (K) −−→Cq (K) −
→Cq−1 (K) → · · · → C0 (K) → 0 →
...

of free abelian group homomorphism in which composite of any two consecutive homomor-
phism is zero (Lemma 3.2.2). This long sequence is called the oriented simplicial chain
complex of K.

Since ∂q ◦ ∂q+1 = 0 for each q, Im∂q+1 ⊆ Ker∂q , i.e Bq (K) ⊆ Zq (K). Consequently we can
talk about quotient group.

Definition 3.2.5. Let K be an oriented complex. Then q-dimensional homology group

of K, denoted by Hq (K), is defined by the quotient group

16
 Hq (K) = Zq (K)/Bq (K)

If we consider the chain complex C∗ (K; G) where G is any coefficient group then similarly
define

Hq (K; G) = Zq (K; G)/Bq (K; G)

This will be called the homology group of K with G as coefficients.

Remark 3.2.2. It is a important observation that if we take G = Q then Cq (K; G) is

a vector space over Q and therefore Hq (K; G) is also a vector space over Q which is the
quotient space of the vector subspaces Zq (K; G) and Bq (K; G) and more importantly the
dimension of Hq (K; G) is the same as the rank of the free group Hq (K) with Z as coefficient
group.

Theorem 3.2.3. Let K1 and K2 denote the same simplicial complex K with different
orientation. Then Hq (K1 ) ∼
= Hq (K2 ).

Proof. Let σ q ∈ K
Let σ q = i σ q ∈ Ki for i=1,2.
But 1 σ q = Ψ(σ)2 σ q such that Ψ(σ) = ±1 depending upon the orientation of σ q . Thus we
get a mapping Ψ : K −→ {−1, 1}.
Now we define a sequence fq : Cq (K1 ) −→ Cq (K2 ) homomorphisms by
X X
fq ( gi .1 σ q ) = gi .Ψ(σ q ).2 σ q (3.3)

gi .1 σ q ∈ Cq (K1 ).
P
where c=
Now we claim that f ◦ ∂ = ∂ ◦ f

Now,
fq−1 (∂(1 σ q )) = fq−1 ( ki=0 [1 σ q , 1i σ q−1 ]1i σ q−1 )
P

⇒ fq−1 (∂(1 σ q )) = ki=0 Ψ(σ q ).Ψ(σ q−1 ).Ψ(σ q−1 ).[2 σ q , 2i σ q−1 ]2i σ q−1
P

⇒ fq−1 (∂(1 σ q )) = ki=0 Ψ(σ q ).[2 σ q , 2i σ q−1 ]2i σ q−1

P

⇒ fq−1 (∂(1 σ q )) = Ψ(σ q ) ki=0 [2 σ q , 2i σ q−1 ]2i σ q−1

P

⇒ fq−1 (∂(1 σ q )) = ∂(fq (1 σ q ))

17
 ∂
Cq (K1 ) Cq−1 (K1 )
fq fq−1

∂
Cq (K2 ) Cq−1 (K1 )

Now, let zq ∈ Zq (K1 ) ⇒ ∂(zq ) = 0 ⇒ fq−1 (∂(zq )) = 0

⇒ ∂(fq (zq )) = 0 ⇒ fq (zq ) ∈ Zq .
So, since zq is arbitrary
fq (Zq (K1 )) ⊆ Zq (K2 ). (3.4)

Similarly exchanging the role of K1 and K2 we get that,

fq (Zq (K2 )) ⊆ Zq (K1 ) (3.5)

from 3.4 and 3.5 We get,

Zq (K2 ) ∼
= Zq (K1 ) (3.6)

In similar way we can obtain,

Bq (K2 ) ∼
= Bq (K1 ) (3.7)

From 3.6 and 3.7 we get Hq (K1 ) ∼

= Hq (K2 ).[3]

3.3 Some Examples

Example 3.3.1. Let K be the 1-skeleton of the complex K=cl(σ 2 ). Then K has 3 vertices
namely v0 , v1 , v2 and 3 edges < v0 , v1 >, < v1 , v2 >, < v0 , v2 > (fig 3.2) and K is the
triangulation of the space |K| which is homeomorphic to S1 .

Figure 3.2: Positively oriented 1-skeleton of 2-simplex

18
We orient K by the ordering v0 < v1 < v2 . Hence the nontrivial part of the complex C(K)
looks like,

0 −→ C1 (K) −→ C0 (K) −→ 0 (3.8)

Where C1 (K) = Z < v0 > ⊕Z < v1 > ⊕Z < v2 >

and C0 (K) = Z < v0 , v1 > ⊕Z < v1 , v2 > ⊕Z < v0 , v2 >.
Now trivially Z0 (K) = Z ⊕ Z ⊕ Z
Suppose c = m0 < v0 , v1 > +m1 < v1 , v2 > +m2 < v0 , v2 >∈ C1 (K)
⇒ ∂1 (c) = −(m0 + m2 ) < v0 > +(m0 − m1 ) < v1 > +(m1 + m2 ) < v2 >
Let,h0 = −(m0 + m2 ), h1 = (m0 − m1 ), h2 = (m1 + m2 )
⇒ h0 = −(h1 + h2 )
then, clearly ⇒ B0 (K) = Z ⊕ Z.
Hence H0 (K) = Z.
Trivially B1 (K) = 0.
Now let,z1 ∈ Z1 (K).
⇒ ∂1 (z1 ) = 0
⇒ −(m0 + m2 ) < v0 > +(m0 − m1 ) < v1 > +(m1 + m2 ) < v2 >= 0
From P roposition 2.1.1,
−(m0 + m2 ) = (m0 − m1 ) = (m1 + m2 ) = 0
⇒ m0 = −m2 , m1 = m0 = −m2
⇒ Z1 (K) = Z
H1 (K) = Z
So, 



 Z, if q=0


Hq (K) = Z, if q=1




0,

otherwise

Example 3.3.2. Let σ 3 =< v0 , v1 , v2 , v3 > be a 3-simplex and L be the 2-skeleton of K

(fig 3.3). Then |L| is homeomorphic to S2 . We orient L by v0 < v1 < v2 < v3 .
As described in example 3.3.1 we can obtain the homology group and get the following:

19
 Figure 3.3: Positively oriented 2-skeleton of 3-simplex


if q=0,2

Z,
Hq (K) =
0,

otherwise

There are few more examples described in [1],[2],[3],[4].

3.4 The general structure of Hq

Since we are dealing with finite simplicial complexes so for each q > 0, Hq (K) for any
simplicial complex K will be finitely generated.
So, from the Structure theorem for finitely generated abelian groups, we get,

Hq (K) ∼
= Z ⊕ · · · ⊕ Z ⊕ Z/m1 Z · · · ⊕ Z/mn Z (3.9)

where mi |mi+1 ∀ i ∈ {1, 2, . . . , n − 1}.

Definition 3.4.1. The number of copies of Z in Hq (K) (refer 3.9) is called q-th Betti
Number (In the honour of E. Betti (1823-1892)).

20
Chapter 4

The Euler Poincaré Theorem

In 1752 Leonard Euler proved the famous theorem for simple regular polyhedrons:

No. of vertices-No. of edges+No. of faces=2

Later Poincaré (1854-1912) used the term Betti number in his paper Analysis Situs[3] and
assuming the Euler Characteristics as a topological invariance and generalizes the formula.
Actually he has conjectured a more stronger result of which Euler’s theorem comes as a
corollary. Later J.W Alexander(1915) and Oswald Verben(1922)[3] gave a rigorous proof of
the Euler Poincaré theorem and the topological invariance of Euler characteristics.

Definition 4.0.1. Let K be an oriented Complex of dimension n and Rq (K) be the (Betti
Numbers), rank of the abelian group Hq (K), q ∈ {0, 1, . . . , n}. Then the alternate sum
Pn q
q=0 (−1) Rq (K), denoted by χ(K), is called the Euler Characteristics of K.

Remark 4.0.1. From Remark 3.2.2 Hi (K; Q) will be a vector space over Q and it is an
important observation that rank (Hi (K; Z)) = dim(Hi (K; Q)).

Let K and K 0 be two triangulation of a space X, then by topological invariance of

simplicial homology Hi (K) ∼
= Hi (K 0 ), ∀i > 0.Hence, the Euler characteristics depends
upon the topology or in other words the Euler characteristics of a compact polyhedron is a
topological invariance.

21
4.1 The Euler Poincaré Theorem
Theorem 4.1.1. Let K be an oriented complex of dimension n. Suppose for each q ∈
{0, 1, 2, . . . , n}, αq denotes the number of q − simplexes of K. Then χ(K) = n0 (−1)q αq .
P

Proof. Consider the following chain complex:

∂q+1
0 −→ Cn (K) −→ · · · −→ Cq+1 −→ Cq (K) · · · −→ C0 (K) −→ 0 (4.1)

Now αq = dimCq (K), q > 0. Also we know that Bq = Im(∂q+1 ) ∼

= Cq+1 /Ker(∂q+1 ) =
Cq+1 /Zq+1 .
⇒ ∀q > 0,
dimBq = dimCq+1 − dimZq+1 (4.2)

Again Hq ∼
= Zq /Bq ⇒ dimHq = dimZq − dimBq
⇒ dimCq+1 = dimZq − dimHq + dimZq+1 , from equation 4.2
Putting, q = −1, 0, 1, . . . , n and taking alternate sum We get that,
Pn q
Pn q
0 (−1) dimCq = 0 (−1) dimHq .

As dimCq = αq and dimHq = Rq (K), so

Pn q
Pn q
0 (−1) Rq (K) = 0 (−1) αq

⇒ χ(K) = n0 (−1)q αq .
P

4.2 Euler’s Theorem and Polyhedron

Definition 4.2.1. By a rectilinear polyhedron P in R3 we mean a solid in R3 which is
bounded by properly joined 2-dimensional convex polygon. The convex polygons are can be
any n − gon, are called the faces of P . The intersection of any two faces are called an edge
of P and any intersection of two edges is called a vertex of P . The polyhedron is said to be
simple if its boundary is homeomorphic to S2 . A rectilinear polyhedron is said to be regular
when its faces are all regular plane polygons and all polyhedral angles are congruent.

Proof of Theorem 1.0.1

22
Proof. Since S is simple then,

|S| ∼
= S2 ⇒ χ(S) = χ(S2 ) (4.3)

Let L be a hollow tetrahedron.

⇒ |L| ∼
= S2 .
From Example 3.3.2 
if q=0,2

Z,
Hq (L) =
0,

otherwise

⇒ χ(L) = 1 · (−1)0 + 1 · (−1)2 = 2, from Euler Poincaré Theorem ( Theorem 4.1.1).

⇒
χ(S2 ) = 2 (4.4)

Again clearly V, E, F are the number of 0-simplexes, 1-simplexes, 2-simplexes of |S|.

⇒ χ(S) = V − E + F. (4.5)

From 4.3,4.4,4.5, we get that

V − E + F = 2.

23
Chapter 5

Classification Theorem for Simple,

Regular Polyhedrons in R3

5.1 Classification Theorem for Simple, Regular Polyhe-

drons in R3
Theorem 5.1.1. There are only 5 simple, regular polyhedrons in R3 .

Proof. Let S be a simple, regular polyhedron in R3 with V vertices, E edges and F faces.
Let m be the number of edges incident on each vertex of S, and n be the number of edges
in every face.
As S is resided in R3 , n > 3.
From Euler’s Theorem (T heorem 1.0.1)

V −E+F =2 (5.1)

Again We know that

mV = 2E = nF (5.2)

from 5.1 and 5.2 we get

F (2n − mn + 2m) = 4m (5.3)

⇒ 2n − mn + 2m > 0
⇒ 2m > n(m − 2) > 3(m − 2) = 3m − 6, [∵ n > 3]

24
⇒ m < 6.
This geometrically mean that there are at most 5 edges meeting at any vertex of S.
From 5.3 and the above inequality and n > 3, we get

F (2n − mn + 2m) = 4m, n > 3, m < 6. (5.4)

But it is clear that 5.4 permits the following five solution for (m,n,F), viz.

Figure 5.1: Tetrahedron Figure 5.2: Cube

Figure 5.3: Octahedron

Figure 5.4: Dodecahedron Figure 5.5: Icosahedron

(m, n, F ) = (3, 3, 4), (4, 3, 8), (3, 4, 6), (3, 5, 12), (5, 3, 20). Therefore there are 5 simple regular
polyhedrons in R3 namely Tetrahedron, Octahedron, Cube, Dodecahedron, Icosahedron (fig
5.1, 5.3, 5.2, 5.4, 5.5 respectively) each satisfying the respective tuples.

5.2 Further Generalization

Definition 5.2.1. A polyhedron P is defined to be a topologically regular if each of its
vertex has same number of edges are incident on them but the faces are not necessarily
congruent although they are homeomorphically equivalent.

Theorem 5.2.1. There are 5 classes of topologically regular polyhedrons in R3

25
Proof. Let S be topologically regular polyhedron with V vertices , E edges and F faces.
It is obvious that S is convex and bounded in R3
⇒ |S| ∼
= S2 , from proposition 2.2.1
⇒ χ(S) = χ(S2 )
but χ(S2 ) = 2, from equation 4.4
⇒ χ(S) = 2
but from T heorem 4.1.1 χ(S) = V − E + F
So,
V −E+F =2 (5.5)

but from theorem 5.1.1 we get that there are 5 solution to 5.5 in R3 and therefore 5 distinct
polyhedrons.
Therefore there are 5 simple, topologically regular polyhedrons in R3 upto homeomorphism.
Hence the conclusion follows.

26
Bibliography

(1) Deo, S., Algebraic Topology A primer ; Hindustan Boook Agency: 2006.

(2) Hatcher, A., Algebraic Topology; Cambridge University Press: 2002.

(3) Croom, F. H., Basic Concepts of Algebraic Topology; Springer-Verlag: 1978.

(4) Massey, W. S., Algebraic Topology: An Introduction; Cambridge University Press: 1977.

27