Scribd
Upload
45 views3 pages

Math 8202 Homework 3 Spring 2015: Tianyu Tao February 11, 2015

This document contains solutions to 7 problems from a homework assignment on abstract algebra. Problem 1 proves that a module M cannot be generated by a single element or decomposed into a direct sum of cyclic modules. Problem 2 finds the elementary divisors and invariant factors of a module. Problem 3 shows that a cyclic module over a principal ideal domain is irreducible if and only if its annihilator is prime.

Uploaded by

Tianyu Tao
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
45 views3 pages

Math 8202 Homework 3 Spring 2015: Tianyu Tao February 11, 2015

This document contains solutions to 7 problems from a homework assignment on abstract algebra. Problem 1 proves that a module M cannot be generated by a single element or decomposed into a direct sum of cyclic modules. Problem 2 finds the elementary divisors and invariant factors of a module. Problem 3 shows that a cyclic module over a principal ideal domain is irreducible if and only if its annihilator is prime.

Uploaded by

Tianyu Tao
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Math 8202 Homework 3 Spring 2015

Tianyu Tao
February 11, 2015
Problem 1:

p. 188 Exx 3.

Proof: First we show M = (2, x) cannot be generated by a single element: indeed,


if M = Zp for some polynomial p = p(x), then 2 = p(x)q(x) and x = p(x)r(x)
comparing degree, one sees p(x) has to be 1, but 1
/ M.
Now if M is generated by two cyclic Z[x] modules, say M = Z[x]a + Z[x]b
where a, b Z[x], then ab Za Zb already shows the sum cannot be direct. If
M cannot be written as direct sum of two cyclic modules, it cannot be written as
direct sum of any number of cyclic modules.
Problem 2:

p.193 Exx 1.

Proof. The elementary divisors are:


( 1)3 , (2 + 1)2 , ( 1), (2 + 1)4 , ( + 2), (2 + 1)2 ,
the invariant factors are: (1)3 (2 +1)4 (+2), (2 +1)2 (1), and (2 +1)2 .

Problem 3:

p.193 Exx 2.

Proof: We shall use exercise 8 on page 170: M is irreducible if and only if M ' D/I
where I is a maximal ideal in D. If p is a prime and annz = (p), then M ' D/(p)
by definition of cyclic module, since p is prime and D is p.i.d. we know (p) is
maximal, hence M is irreducible; conversely, if M is irreducible, then M ' D/I
for some I D maximal, since D is p.i.d we know I = (p), and we know (p) is
prime.
To show exercise 8, first suppose M is irreducible, let : R M be the map
r 7 rm for some m 6= 0 fixed, this is surjective since (m) = M by irreduciblity. By first isomorphism theorem we see M ' R/ ker , by the correspondence
1

Math 8202 Homework 3

Tianyu Tao

theorem (4th isomorphism theorem/Lattice theorem for modules on e.g. Dummit


and Foote), ker is maximal ideal, since any submodule of M will corresponds to
R/I 0 for some I 0 containing I; conversely, if M ' R/I for some I maximal, then
correspondence theorem shows M contains no submodule.
Now suppose M is indecomposable, by structure theorem M is not direct sum
of more than 1 element, so either M ' D/(0) = D (which is free of rank 1) or
M ' D/(pe ) for some prime p; conversely, if M = Dz ' D/(0), clearly M is
not direct sum of two non-zero submodules because nothing annilates z except 0,
otherwise if M ' D/(pe ), then by correspondence theorem, any submodule of M
corresponds to D/I where I contains (pe ), hence I = (pl ) where l < m (since I is
p.i.d. so I = (k), since pe I we see k divides pe , but p is prime). But then the
sum of any two submodules of M cannot be direct, so M is indecomposable.

Problem 4:

p.194 Exx 5.

Proof: I can
P only think of the following approach: grant we have the formula
[AB]i,j = [A]i,k [B]k,j where [A]i,j denotes the i, jth minor of the matrix A, we
can then use the formula on page 185 to compute the diagonal entries and notice
a1 = gcd(aij ) and b1 = gcd(bi j), then clearly a1 divides the g.c.d of (AB)ij since the
entries are sums and products each involves a factor from aij , similar considerations
holds since the are computed using g.c.ds.

Problem 5:

p.201 Exx 1.

Proof: Note 360 = 23 32 5. By the structure theorem the number of nonisomorphic abelian groups of order 360 equals the number of ways we can write
23 , 32 , 5 as product of different prime powers, there are the following cases: 3 =
1 + 2 = 1 + 1 + 1 = 3 + 0, 2 = 1 + 1 = 2 + 0, 5 = 5 + 0, so there are 3 2 1 = 6
different abelian groups of order 360, these are:
Z/(23 ) Z/(32 ) Z/(5);
Z/(2) Z/(22 ) Z/(32 ) Z/(5);
Z/(2) Z/(2) Z/(2) Z/(32 ) Z/(5);
Z/(23 ) Z/(3) Z/(3) Z/(5);
2

Math 8202 Homework 3

Tianyu Tao

Z/(2) Z/(22 ) Z/(3) Z/(3) Z/(5);


Z/(2) Z/(2) Z/(2) Z/(3) Z/(3) Z/(5).

Problem 6:

p.201 Exx 2.

Proof: By a standard algorithm we convert A = {aij } to its Smith normal form


S, and we note all the elementary matrix in MZ has determinant 1 because they
are invertible and 1 are the only units in Z. So | det A| = | det S|. Now say
S =diag(1, . . . , 1, d1 , . . . , ds ) where the di s are the invariant
factors, we know by
Ls
(n)
the structure theorem M = Z /K is isomorphic to i=1 Zzi where ann zi = (di ).
Each Zzi is isomorphic to Z/(di ), so has order |di |, the order of M is thus the
product of the absolute value of di s, which is equal to | det S| since S is diagonal,
which in turn equals |d| = | det A|.
Problem 7:

p.202 Exx 5.

Proof: Let V = F (n) which is both a F vector space and a F []module.


Suppose I A = A and I B = B are equivalent, that is, there exist
P1 , Q1 invertible matrix in Mn (F []) such that P1 A Q1 = B , then in particular,
they can be converted to the same diagonal matrix D =diag(1, . . . , 1, d1 , . . . , ds )
where di s are the invariant factors of V seen as a F []module, this is easy to see
since if P, Q are such that P B Q = D, then P P1 A Q1 Q = D and P P1 , Q1 Q are
invertible. Having this in mind, let us review the description on page 198, which
says we can transform A to its rational canonical form via a matrix which is the
relation matrix between the F vector basis and the F []module basis of V , but
the rational canonical form only depends on the set of invariant factors, since A
and B are equivalent, this say A and B can be converted to the same rational
canonical form C via similarity: A = PA CPA1 and B = PB CPB1 for some PA , PB
invertible, then clearly A and B are similar.
Conversely, if A and B are similar, then in particular they are both similar
to a common rational canonical form, hence they have the same invariant factors,
then, applying the algorithm of converting a matrix to its Smith normal form will
leads A and B to be equivalent to the same diagonal form, which says A is
equivalent to B .

You might also like