Algebraic Topology Solutions 3

Ian Vincent
February 14, 2013
Exercise 1
Compute the simplicial homology groups of the two-sphere, directly from the denitions, using
the -complex structure coming from the boundary of a tetrahedron. Provide clearly labelled
gures.
Answer:
In the above diagram, I have labelled the tetrahedrons vertices by w
0
, w
1
, w
2
, w
3
, the edges
as:
a = |w
0
, w
1
|, b = |w
1
, w
2
|, c = |w
0
, w
2
|, d = |w
0
, w
3
|, e = |w
1
, w
3
|, f = |w
2
, w
3
|
and the faces of the tetrahedron as:
P = |w
0
, w
1
, w
2
|, Q = |w
0
, w
1
, w
3
|, R = |w
1
, w
2
, w
3
|, S = |w
0
, w
2
, w
3
|
This means that
2
= P, Q, R, S Z
4
,
1
= a, b, c, d, e, f Z
6
,
0
= w
0
, w
1
, w
2
, w
3
Z
4
,
giving us the chain complex
0

3
Z
4

2
Z
6

1
Z
4

0
0
Being a homomorphism from 0, we have im
3
= 0 and so H

2
(S
2
) =
ker
2
im
3
ker
2
.
For ease of computation, we write the boundary maps
2
and
1
as matrices:

2

P Q R S
a
b
c
d
e
f

1 1 0 0
1 0 1 0
1 0 0 1
0 1 0 1
0 1 1 0
0 0 1 1

1

a b c d e f
w
0
w
1
w
2
w
3

1 0 1 1 0 0
1 1 0 0 1 0
0 1 1 0 0 1
0 0 0 1 1 1

Now, to compute the kernels and images of the boundary maps, it is much easier to convert to
Smith Normal Form (remembering to work over Z, not R). Either doing it by hand, or putting
1
these matrices into your favourite computer algebra package, you will get:

1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 0
0 0 0 0
0 0 0 0

1 0 0 0 0 0
0 1 0 0 0 0
0 0 1 0 0 0
0 0 0 0 0 0

Therefore rank(
2
) = 3 so ker(
2
) Z. Similarly, rank(
1
) = 3 so ker(
1
) Z
3
. As
0
is the zero
map, we have ker(
0
) = Z
4
. As the diagonal entries of the Smith Normal Form of
1
are all 1,
we conclude that:
H

0
(S
2
) =
Z
Z

Z
Z

Z
Z
Z
43
Z
Similarly, the diagonal entries of the Smith Normal Form of
2
are all 1, we conclude that
H

1
(S
2
) =
Z
Z

Z
Z

Z
Z
Z
33
0
As H

2
(S
2
) = ker(
2
), we just have that
H

2
(S
2
) = Z
For the rest, when n 3 we have
n
= 0, so ker
n
= im
n+1
= 0 and so H

n
(S
2
) 0 for
n 3. and we are done.
Remark 1. It is not sucient to simply say Oh look!, ker
1
=
2
P,
2
Q,
2
R Z
3
, and
im
2
=
2
P,
2
Q,
2
R Z
3
, so H

1
Z
33
= 0.
While it works in this case, this is not necessarily true all the time. A simple counterexample
is the following: Let H = a and G = 2a. Certainly, H G Z but
H
G

Z
2Z
, not 0. The
important distinction is that the diagonal entries in the matrices give the Torsion Coecients
of the homology group. It just so happens that in this example all those torsion coecients
were 1.
Remark 2. It is entirely possible to do this question without Smith Normal Form, and simply
comparing generating sets for all the kernels and images directly. However, this needs to be
done with great care to avoid issues like those present in the rst remark. The whole reason
why Smith Normal Form works is because it does all the work in comparing generators for you
by changing to a nicer basis.
Not only that, but it is denately a lot quicker and easier to do Smith Normal form than to
compute kernels and images directly!
2
Exercise 2
Suppose that X is a non-empty, path-connected space, equipped with a -complex structure.
Show, directly from the denitions that H

0
(X) Z. (You may assume without proof that the
one-skeleton is connected.
Answer
Recall that H

0
(X) =
ker
0
im
1
. Remembering what a quotient group actually is, this means two
elements of H

0
(X) are equal if and only if the dierence between (their representatives) is a
boundary.
Also note that
0
is the zero map, so ker
0
=
0
, which is by denition the free abelian group
generated by the vertices v
i
of X. Pick any two vertices v
i
, v
j
. As the 1-skeleton X
1
of X is
connected, we can choose a nite sum of edges =
k
e
k
that geometrically form a continuous
path connecting the vertices v
i
and v
j
. But then
1
() = v
j
v
i
so the homology classes
represented by v
i
and v
j
are equal, meaning H

0
(X) = v Z for any vertex v.
Exercise 3
Suppose X is a nite, path-connected, one-dimensional -complex: that is, a nite connected
graph. Suppose that X has E edges and V vertices. Compute the simplicial homology groups
of X.
Answer
X gives rise to the following chain complex
0

2
Z
E

1
Z
V

0
0
If X = , then immediately from the denitions we have that H

n
(X) = 0 for all X. For the
rest of the assignment, assume X . I think it is easier to think of
1
as represented by some
matrix with V rows and E columns, so from now on I think of
1
as a linear map over Z.
We know that as X is path-connected, by assignment 3.2 we know that H

0
(X) = Z. As
H

0
(X) =
Z
V
im
1
= Z, this implies that the rank of
1
is V 1. Therefore by the rst isomorphism
theorem, the dimension of the kernel of
1
is E (V 1) = E V + 1 and so ker
1
Z
EV +1
.
But now as
2
is the zero map, we have H

1
(X) =
ker
1
im
2
= ker
1
Z
EV +1
.
As there are no simplices in dimensions 2 or higher, all other boundary maps are zero and hence
H

n
(X) 0 for all n 2.
Exercise 4
Compute the reduced singular homology groups of a point, directly from the denitions.
Answer
By denition [Hatcher, page 110], the reduced homology groups

H
n
(X) of X are the homology
groups of the following chain complex:

3
C
2
(X)

2
C
1
(X)

1
C
0
(X)

Z 0
where is the map dened by (
i
n
i

i
) =
i
n
i
for n
i
Z and singular 0-simplices
i
.
Note, for each n, there is a unique map which we call
n

n
X = pt. Therefore C
n
(X) is
3
the free abelian group generated by one element
n
, hence isomorphic to Z.
Now, by denition,
n
(
n
) =
n
i=0
(1)
i

n
||v
0
, ..., v
i
, ..., v
n
| but
n
restricted to any (n 1)-face
of
n
is simply
n1
. Therefore

n
(
n
) = _
n

i=0
(1)
i
_
n1
=

n1
if n is even and nonzero
0 if n is odd
In particular, if n is even and non-zero then
n
is an isomorphism (it sends the generator to a
generator). Now, we know Z Z just maps n
0
n which is also an isomorphism. Hence
we get the singular chain complex
Z

Z
0
Z

Z
0
Z

Z 0
We have an alternating sequence of 0 maps and isomorphisms, so it is now very easy to compute
the singular homology:
If n is odd then

H
n
(X) =
ker
n
im
n+1

Z
Z
0. If n is even then

H
n
(X) =
0
0
0 so the reduced
homology of a point is trivial in all dimensions.
Exercise 5
Compute the singular homology groups of the circle S
1
, directly from the denitions.
Reaction
The exercise is labelled Challenge. Therefore, in order for it to remain a Challenge you should
try it yourself!
Exercise 6
Compute the singular homology groups of the plane R
2
, minus n points.
Reference any theorems from Hatcher that you use.
Answer
Let X be R
2
x
1
, ..., x
n
. We may assume that the n points x
1
, ..., x
n
are equally spaced along
the x-axis. The idea is to homotope the whole of X down to a wedge sum of n circles, each
one S
1
i
centered about the removed point x
i
. Such a map is in general hard to write down, but
geometrically the situation is as follows:
The idea is to gradually contract X in the direction of the arrows through the grey shape,
and down to the above circles in a continuous way.
4
Note that X is path connected and non-empty, so by Proposition 2.7 of Hatcher on page 109,
we have H
0
(X) Z. For the others, we now apply Corollary 2.11 from page 111 in Hatcher:
Corollary. The maps f

H
n
(X) H
n
(Y ) induced by a homotopy equivalence f X Y are
isomorphisms for all n.
Therefore we now need to compute H
n
{

n
i=1
S
1
). Flicking through Hatcher to Corollary
2.25 on page 126:
Corollary. For a wedge sum

, the inclusions i

induce an isomorphism

(i

H
n
(X

)

H
n
_

_
provided that the wedge sum is formed at basepoints x

such that the pairs (X

, x

) are
good.
Note that the above shape is homotopy equivalent to a wedge sum of circles (we need that
the circles are all attached at one point rather than in a series). Note that in our case, about
any point y S
1
there exists a neighbourhood U of which y is a deformation retract of U (that
is, U is contractible) so we are allowed to apply the corollary.
Recall that the reduced singular homology of S
1
is

H
k
(S
1
) =

Z if k = 1
0 otherwise
Therefore the two corollaries do all the work for us and so we get
H
k
(X) =

Z if k = 0
Z
n
if k = 1
0 otherwise
Exercise 7
Suppose f
#
C

is a chain map of chain complexes; that is f = f. Show that this

induces a well dened homomorphism f

H
n
(C) H
n
(D) on homology.
Reaction
The induced map is dened by the formula f

|c| = |f
#
(c)| where the square brackets denote
the homology class of. From here on, just treat H
n
(X) as you would any other quotient of
abelian groups. See Algebra II for example.
Exercise 8
1. Suppose that X is a -complex. Let i

(X) C

(X) be the inclusion homomorphism.

Show that i is a chain map.
2. Suppose that f X Y is a map of topological spaces. Show that the induced function
f
#
C

(X) C

(Y ) is a homomorphism, and a chain map.

5
Answer
1. This follows immediately from the fact that i is a homomorphism:
(i()) =
i
(1)
i
i()|v
0
, ..., v
i
, ..., v
n
|
= i _
i
(1)
i
|v
0
, ..., v
i
, ..., v
n
|_ = i(())
2. Go read Hatcher, pages 110-111.
Exercise 9
Show that chain homotopy of chain maps is an equivalence relation.
Answer
We say two chain maps f, g are chain homotopic, and write f g if there exists a homomorphism
P (called the Prism Operator) so that P + P = f g.
Now, note that if P and Q are homomorphisms, then so are P and P + Q. Now note that
for any f, f f = 0 = 0 + 0 so is reexive. If f g then there exists P so P + P =
f g hence g f = (f g) = (P) + (P) which means g f so is symmetric. If f g
and g h then there exists P and Q so f g = P + P and g h = Q + Q. Therefore
f h = (f g) + (g h) = (P + Q) + (P + Q) so f h and is transitive.
Exercise 10
We say two chain complexes C

and D

are chain homotopy equivalent if there are chain maps

f C

and g D

so that g f Id
C
and f g Id
D
. Let C

be the chain complex

with C
1
= C
0
= Z and all other chain groups trivial, and with (m) = 2m. Let D

be the chain
complex with D
1
= D
0
= Z
2
, and all other chain groups trivial, and with

(x, y) = (xy, x+y).

Prove that C

and D

are homotopy equivalent.

Answer
We need to nd maps f
1
, g
1
, f
0
and g
0
that t into the following diagram, and linear maps
P
1
, ..., P
4
so that:
0P
1
+ P
2
= g
1
f
1
1
P
2
+ 0 = g
0
f
0
1
0P
3
+ P
4

= f
1
g
1
1

P
4
+ 0 = f
0
g
0
1
0
//
0
//
C
1
= Z

//
f
1

C
0
= Z
//
f
0

0
0
//
0
//
D
1
= Z
2

//
g
1
OO
D
0
= Z
2
//
g
0
OO
0
Note that we may as well let P
1
and P
3
be zero since the boundary operators C
2
C
1
and
D
2
D
1
are zero.
Dene f
1
(m) = (m, m) and f
0
(m) = (0, m). Then f is a chain map as

f
1
(m) = (0, 2m) =
6
f
0
(m). Dene g
1
(x, y) = x and g
0
(x, y) = x + y. Then g is a chain map as g
1
(x, y) = 2x =
g
0

(x, y).
Thus g
1
f
1
(m) = g
1
(m, m) = m and g
0
f
0
(m) = g
0
(0, m) = m so we can use P
2
= 0.
On the other hand, f
1
g
1
(x, y) = f
1
(x) = (x, x) and f
0
g
0
(x, y) = f
0
(x + y) = (0, x + y) is
very much not the identity!
We use the prism operator P
4
(x, y) = (0, x) and check it works:
f
1
g
1
(x, y) Id(x, y) = (x, x) (x, y) = (0, x y) = P
4
(x y, x + y) = P
4

(x, y) and:
f
0
g
0
(x, y) Id(x, y) = (0, x + y) (x, y) = (x, x) = (0, x) =

P
4
(x, y) as required.
7