Alg Sol 2016

Preliminary Exam 2016
Solutions to Afternoon Exam
Part I.
Solve four of the following five problems.
Problem 1. Let W be the plane x + 2y + 3z = 0 in R3 . Find a basis for W
which is orthonomal relative to the usual dot product.
Solution: The vectors w1 = (2, −1, 0) and w2 = (3, 0, −1) are a basis for W . We
use the Gram-Schmidt process to obtain an orthogonal basis {v1 , v2 } for W : Put
v1 = w1 and
w1 · w2
v2 = w2 − w1 = w2 − (6/5)w1 = (3/5, 6/5, −1).
w1 · w1
Finally, an orthonormal basis is {u1 , u2 }, where
√ √
u1 = v1 /|v1 | = (2/ 5, −1/ 5, 0)
and
√ √ √
u2 = v2 /|v2 | = (3/ 70, 6/ 70, −5/ 70).
Of course there are infinitely many correct answers.
Problem 2. Consider the matrix (with real coefficients)

5 −2
A= .
3 0
Find an invertible matrix C such that C −1 AC is diagonal.
Solution: The characteristic polynomial of A is x2 − 5x + 6 = (x − 2)(x − 3),
and one easily computes that (2, 3) and (1, 1) are eigenvectors corresponding to the
eigenvalues 2 and 3. So the matrix

2 1
C=
3 1
has the property that C −1 AC is the diagonal matrix with diagonal entries 2 and 3.
Problem 3. Let A be a 2 × 2 matrix with real coefficients, and suppose that
A4 = I but A2 6= I, where I is the 2 × 2 identity matrix.
(a) Find tr (A).
Solution: Since A4 = I the eigenvalues of A are contained in the set {±1, ±i},
and since the equation x4 − 1 = 0 has 4 distinct roots A is diagonalizable. But since
A2 6= I, either i or −i is an eigenvalue. In fact if one i or −i is an eigenvalue then
both are eigenvalues, because A – and therefore also the characteristic polynomial
of A – has real coefficients. Therefore the eigenvalues are precisely i and −i, and
tr (A) = i + (−i) = 0.
(b) Give an example of a 2 × 2 matrix B with complex coefficients such that
B 4 = I and B 2 6= I but tr (B) 6= tr (A).
Solution: Let

1 0
B= ,
0 i
so that tr (B) = 1 + i 6= tr (A).
1
2
Problem 4. Let ϕ : Q → Z be a group homomorphism. Show that ϕ(x) = 0

for all x ∈ Q.
Solution: Suppose that for some x ∈ Q we have ϕ(x) 6= 0. Then there is a prime
number p not dividing ϕ(x). Since ϕ is a homomorphism, we find
ϕ(x) = ϕ(p(x/p)) = pϕ(x/p) = pn
where n = ϕ(x/p) ∈ Z, a contradiction.
Problem 5. A sequence of row operations transforms the matrix
   
1 4 ∗ ∗ 1 0 1 −3
A = 3 2 ∗ ∗ into the matrix B = 0 1 2 5  .
5 3 ∗ ∗ 0 0 0 0
Let T : R4 → R3 be the linear transformation which has A as its matrix relative to
the standard bases of R4 and R3 . Find a basis for the kernel and image of T , and
determine the third and fourth columns of A.
Solution: Because row operations correspond to left-multiplication by an invert-
ible matrix, B is the matrix of T relative to the standard basis {e1 , e2 , e3 , e4 } for
R4 and a new basis {v1 , v2 , v3 , } for R3 . By inspection, T (e3 ) = T (e1 ) + 2T (e2 ) and
T (e4 ) = −3T (e1 ) + 5T (e2 ), and consequently the vectors
e1 + 2e2 − e3 = (1, 2, −1, 0)
and
−3e1 + 5e2 − e4 = (−3, 5, 0, −1)
are in the kernel of T . Since the nullity of T is visibly 2 and these vectors are
linearly independent, they form a basis for the kernel of T . Using B and A one can
also read off a basis for the image of T : A basis is given by
v1 = T (e1 ) = (1, 3, 5)
and
v2 = T (e2 ) = (4, 2, 3)
Finally, the third and fourth columns of A are the transposes of
v1 + 2v2 = T (e1 ) + 2T (e2 ) = (1, 3, 5) + 2(4, 2, 3) = (9, 7, 11)
and
−3v1 + 5v2 = −3T (e1 ) + 5T (e2 ) = −3(1, 3, 5) + 5(4, 2, 3) = (17, 1, 0)
respectively.
Part II.
Solve three of the following six problems.
Problem 6. Let A be a 3 × 3 matrix such that (A2 − I)(A − I) = 0 but
2
A − I 6= 0. What are the possibilites for the Jordan normal form of A? (You may
take A to have complex coefficients.)
Solution: Let f (x) be the minimal monic polynomial of A. As (A−I)2 (A+I) = 0
but (A − I)(A + I) 6= 0 we see that f (x) divides (x − 1)2 (x + 1) but does not divide
3
(x − 1)(x + 1). So the possibilities for f (x) are (x − 1)2 (x + 1) and (x − 1)2 . Given
that A is a 3 × 3 matrix, the corresponding Jordan normal forms are
   
1 1 0 1 1 0
0 1 0  and 0 1 0
0 0 −1 0 0 1
Problem 7. Make a list of abelian groups of order 32 so that every abelian
group of order 32 is isomorphic to exactly one group on your list.
Solution: By the Elementary Divisor Theorem, there are 7 possibilities:
Z/32Z,
(Z/16Z) × (Z/2Z),
(Z/8Z) × (Z/4Z),
(Z/8Z) × (Z/2Z) × (Z/2Z),
(Z/4Z) × (Z/4Z) × (Z/2Z),
(Z/4Z) × (Z/2Z) × (Z/2Z) × (Z/2Z),
(Z/2Z) × (Z/2Z) × (Z/2Z) × (Z/2Z) × (Z/2Z).
Problem 8. Let R be a commutative ring, and let I be the set of all r ∈ R such
that rk = 0 for some positive integer k (which may depend on r). Show that I is
an ideal of R.
Solution: If r, s ∈ I then there exist integrers k, l > 1 such that rk = sl = 0.
Since R is commutative, the Binomial Theorem shows that (r + s)k+l is a sum of
integer multiples of expressions of the form rj sk+l−j with 0 6 j 6 k + l, and either
j > k or k + l − j > l. Therefore all such terms are 0 and r + s ∈ I. On the other
hand, if a ∈ R then (ar)k = ak rk = ak · 0 = 0. So I is an ideal.
Problem 9. Write Z3 /L as a direct sum (or a direct product) of cyclic groups,
where L is the subgroup of Z3 generated by (1, 1, −2), (7, 9, −14), and (5, 9, −6).
Solution: Let A be the matrix with these vectors as columns:
 
1 7 5
A= 1 9 9 .
−2 −14 −6
A series of row and column operations over Z converts A into a diagonal matrix
with diagonal entries 1, 2, and 4. So Z3 /L ∼
= (Z/2Z) × (Z/4Z).
Problem 10. Find an integer n > 1 such that the Galois group over Q of the
polynomial xn − n is isomorphic to (Z/2Z)2 .
Solution: Take n = 4. Since x4 − 4 = (x2 + 2)(x2 − 2), the corresponding
√
Galois extension
√ of Q is the compositum K = K1 K2 , where K1 = Q( −2) and
K2 = Q( 2). Note that Gal(K1 /Q) ∼ = Gal(K2 /Q) ∼= Z/2Z. Also K1 ∩ K2 = Q:
Indeed K1 6= K2 (since K2 ⊂ R and K1 6⊂ R) and the only subfields of K2 are
K2 itself and Q (since [K2 : Q] = 2). Hence the formal properties of the Galois
correspondence give Gal(K/Q) ∼ = Gal(K1 /Q) × Gal(K2 /Q) ∼= (Z/2Z)2 .
4
Problem 11. Let A be an n × n matrix, let Aij be the (n − 1) × (n − 1) matrix

obtained by deleting the ith row and jth column of A, and
Plet aij be the entry in the
n
ith row and jth column of A. For 1 6 h, i 6 n, what is j=1 ahj (−1)i+j det(Aij )?
Why? Your answer should depend on whether h = i or h 6= i.
Solution: If h = i then the given sum is just the expansion of det(A) along the
ith row, so the value of the sum is det(A). If h 6= i then the given sum is the
expansion of det(A0 ) along the ith row, where A0 is obtained from A by changing
the ith row so that it is identical to the hth row and leaving all other rows the
same. Thus in A0 the ith row and hth row are equal, and so det(A0 ) = 0. So the
given sum is 0.
Part III.
Solve one of the following three problems.
Problem 12. Factor the polynomial x3 − 2 into irreducible polynomials in Fp [x]
for p = 3, 5, and 7, where Fp is the field with p elements.
Solution: If p = 3 then x3 − 2 = x3 + 1 = (x + 1)3 .
If p = 5 then |F× 3
p | = 4, so the map u 7→ u is an isomorphism and 2 = α for
3
some α ∈ F5 . In fact 33 = 27 = 2 in F5 , so x − 3 is a factor of x3 − 2 in F5 [x]. Using

synthetic division or otherwise, one finds that x3 − 2 = (x − 3)(x2 − 2x − 1), and
f (x) = x2 − 2x − 1 is irreducible because all the values of f are nonzero: f (0) = 4,
f (1) = 3, f (2) = 4, f (3) = 2, and f (4) = 2.
Finally, if p = 7 then f (x) = x3 − 2 is irreducible, either because the values of f
are all nonzero or because 2 has order 3 in F× 3
7 (indeed 2 = 8 = 1 in F7 ) and any
cube in a group of order 6 has order dividing 2.
Problem 13. Let G be a finite group and H a subgroup, and let S be the group
of permutations of G/H, the set of left cosets of H in G. Define a homomorphism
ϕ : G → S by setting ϕ(a)(bH) = abH for a, b ∈ G. Show that the kernel of ϕ is
the largest normal subgroup of G contained in H.
Solution: We have a ∈ ker ϕ if and only if bH = abH forTall b ∈ G, or in
other words, b−1 ab ∈ H for all b ∈ G. Equivalently, ker ϕ = b∈G bHb−1 . This
intersection is contained in H (take b = 1), and it is a subgroup of G because
it is an intersection of subgroups of G. Furthermore it is normal in G because
conjugation by a fixed element of G merely permutes the terms in the intersection.
Finally, if N is a normal subgroup of G and N ⊂ H then for any T b ∈ G we have
N = bN b−1 ⊂ bHb−1 . Therefore N is contained in the intersection b∈G bHb−1 .
Problem 14. Let A be a square matrix with coefficients in C, and suppose that
A is not diagonalizable but that An is diagonalizable for some n > 2. Show that
det(A) = 0.
Solution: A square matrix over C is diagonalizable if and only if its minimal
polynomial can be factored into distinct monic linear factors. Thus f (An ) = 0
for some polynomial f (x) = (x − λ1 )(x − λ2 ) · · · (x − λmQ) such that λ1 , λ2 , . . . , λm
m Qn
are all distinct. Equivalently, g(A) = 0, where g(x) = j=1 k=1 (x − µj e2πik/n )
where µnj = λj . If every λj is nonzero then g(x) is a product of distinct monic
linear factors and therefore the same assertion holds for the minimal polynomial
of A, which divides g(x) since g(A) = 0. Then A is diagonalizable, contrary to
hypothesis. So one of the λ0j s is zero and therefore det(An ) = 0, whence also
det(A) = 0.

Alg Sol 2016

Uploaded by

Copyright:

Available Formats

Alg Sol 2016

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Alg Sol 2016

Uploaded by

Copyright:

Available Formats

Preliminary Exam 2016

Solutions to Afternoon Exam

Problem 4. Let ϕ : Q → Z be a group homomorphism. Show that ϕ(x) = 0

(Z/8Z) × (Z/2Z) × (Z/2Z),

(Z/4Z) × (Z/4Z) × (Z/2Z),

(Z/4Z) × (Z/2Z) × (Z/2Z) × (Z/2Z),

(Z/2Z) × (Z/2Z) × (Z/2Z) × (Z/2Z) × (Z/2Z).

Problem 11. Let A be an n × n matrix, let Aij be the (n − 1) × (n − 1) matrix

some α ∈ F5 . In fact 33 = 27 = 2 in F5 , so x − 3 is a factor of x3 − 2 in F5 [x]. Using

You might also like