(1) Express the symmetric polynomial
over Z in terms of the elementary symmetric polynomials.
Solution. We observe that the leading monomial of f(x, y, z) = x3y2 + x2y3 + x3z2 + x2z3 + y3z2 + y2z3 is x3y2 = x3y2z0, so we subtract s13-2 s22-0s30 from f and find
|
The leading monomial of f1 is -2x3yz = -2x3y1z1, so we subtract -2s13-1 s21-1s31 from f1 and find
|
The leading monomial of f2 is -x2y2z = -x2y2z1, so we subtract -s12-2 s22-1s31 from f2 and find
|
Hence
|
(2) Find a polynomial over Z whose roots are the squares of the roots of t3 + 5t2 + 7t + 1 in Z[t].
Solution. Let's denote the roots of p(t) : = t3 + 5t2 + 7t + 1 by a, b, c. The coefficients of p(t) are the elementary symmetric polynomials in a, b, c, with a suitable sign:
|
Let q(t) = t3 + q1t2 + q2t + q3 be the monic polynomial having a2, b2, c2 as roots. The relationship between its coefficients and roots is as follows:
|
so we have to find si(x2, y2, z2) and substitute a, b, c for x, y, z.
(i) Newton's formula for s2 yields s1(x2, y2, z2) = x2 + y2 + z2 = s2(x, y, z) = [s1(x, y, z)]2 - 2s2(x, y, z). So s1(a2, b2, c2) = [s1(a, b, c)]2 - 2s2(a, b, c) = (-5)2 - 2·7 = 11 and q1 = -s1(a2, b2, c2) = - 11.
(ii) The leading monomial of s2(x2, y2, z2) = x2y2 + x2z2 + y2z2 is x2y2z0, so we subtract s12-2s22-0s30 from it and find s2(x2, y2, z2) - s22 = . . . = -2(x2yz + xy2z + xyz2) = -2(xyz)(x + y + z) = -2s1s3. So s2(x2, y2, z2) = s22 - 2s1s3, so s2(a2, b2, c2) = s22(a, b, c) - 2s1(a, b, c)s3(a, b, c) = 72 - 2(-5)(-1) = 39 and q2 = s2(a2, b2, c2) = 39.
(iii) We have s3(x2, y2, z2) = x2y2z2 = (xyz)2 = [s3(x, y, z)]2, so s3(a2, b2, c2) = [s3(a, b, c)]2 = (-1)2 = 1 and q3 = -s2(a2, b2, c2) = -1.
Thus the required polynomial is t3 - 11t2 + 39t - 1.
(3) Determine whether Q × Q is a vector space if the operations are defined by
|
for all a, b, c, d in Q.
Solution. The given system Q × Q is not an abelian group under addition, so it cannot be a vector space over Q. Indeed, as (1, 1) + (e, f) = (1 + e, 0) ≠ (1, 1) whatever (e, f) maybe, there is no (e, f) in Q × Q for which (1, 1) + (e, f) = (1, 1), so Q × Q has no additive identity, so Q × Q is not a group.
(4) If q is a positive integer and K is a field with q elements, how many elements does Kn have?
Solution. The first entry in an n-tuple in Kn can take on one of the q values from K, likewise the second, third, … , nth entries can take on q values, independently of the other entries, so there are q ·q ·q . . . q = qn elements in Kn.
Quiz Question Construct a field having 125 elements in it.
Solution. From MATH 321 we know that Z5 is a field, from question (4) we know that Z53 has 125 elements in it, from the book we know that Z53 is a vector space over Z5, where the operations are defined as follows:
for all a, b, c, d in Z5.
Thus Z53 is a vector space (over Z5) and has 125 elements in it.
(5) Find all Z2-bases of Z23.
Solution. Since Z23 is finite dimensional over Z2, and since dim Z2 Z23 = 3, the Z2-bases of Z23 are exactly the Z2-linearly independent subsets of Z23 having 3 elements. So we must find three-element subsets of Z23 that are linearly independent over Z2.
There are just two elements of Z2, namely 0 and 1, so two nonzero vectors in Z23 are linearly dependent over Z2 if and only if one of them is the only nonzero scalar 1 times the other one, i.e., if and only if they are equal. Also, if u and v are linearly independent over Z2, then their span is {0, u, v, u + v}, so {u, v, w} is linearly independent over Z2 if and only if w is distinct from 0, u, v, u + v.
So we are to find subsets {u, v, w} of Z23 such that u ≠ 0, v ≠ 0, v ≠ u, w ≠ 0, w ≠ u, w ≠ v, w ≠ u + v. The simplest method of doing this seems writing all the sets {u, v, w} satisfying all these conditions except the last one, and deleting those sets not enjoying the last property. The elements of Z23 being (0, 0, 0), (0, 0, 1), (0, 1, 0), (0, 1, 1), (1, 0, 0), (1, 0, 1), (1, 1, 0), (1, 1, 1) we obtain the following 35 = 7 choose 3 sets, where {(a, b, c), {d, e, f), (g, h, j)} is abbreviated as abc def ghj:
| 001 | 010 | 011 | * |
| 001 | 010 | 100 | |
| 001 | 010 | 101 | |
| 001 | 010 | 110 | |
| 001 | 010 | 111 | |
| 001 | 011 | 100 | |
| 001 | 011 | 101 | |
| 001 | 011 | 110 | |
| 001 | 011 | 111 | |
| 001 | 100 | 101 | * |
| 001 | 100 | 110 | |
| 001 | 100 | 111 | |
| 001 | 101 | 110 | |
| 001 | 101 | 111 | |
| 001 | 110 | 111 | * |
| 010 | 011 | 100 | |
| 010 | 011 | 101 | |
| 010 | 011 | 110 | |
| 010 | 011 | 111 | |
| 010 | 100 | 101 | |
| 010 | 100 | 110 | * |
| 010 | 100 | 111 | |
| 010 | 101 | 110 | |
| 010 | 101 | 111 | * |
| 010 | 110 | 111 | |
| 011 | 100 | 101 | |
| 011 | 100 | 110 | |
| 011 | 100 | 111 | * |
| 011 | 101 | 110 | * |
| 011 | 101 | 111 | |
| 011 | 110 | 111 | |
| 100 | 101 | 110 | |
| 100 | 101 | 111 | |
| 100 | 110 | 111 | |
| 101 | 110 | 111 |
Quiz Question Find the automorphism group Aut(Q(3√5)).
Solution. Put u := 3√5. Let f be any automorphism of K := O(u) = Q(3√5) subset R. Then f|Q is an automorphism of Q, and from the homework solution, we know that the only automorphism of Q is idQ, therefore f|Q is the identity function on Q. Besides, from u3 - 5 = 0, we get (u3)f-5f = 0f, so (uf)3-5 = 0, so uf is a root of the polynomial x3-5. The roots of this polynomial are u, uw, uw2, and uw, uw2 are nonreal complex numbers, so they are not in our field K. Thus uf, being an element of K, is necessarily the unique root u of this polynomial that belongs to K. Then we get also u2f = (uf)2 = u2. Hence (a + bu + cu2)f = af + bf uf + cf (u2)f = a + bu + cu2 for any a, b, c in Q, that is to say, f is the identity function on K.
Thus Aut(K) = {idK}.
Quiz Question Find a field having 73 elements, describe its elements, explain how they are added and multiplied.
Solution. If f(x) is an irreducible cubic polynomial in F7[x], then F7[x]/<f(x)> is a field with exactly 73 elements. Thus all we have to do is find a cubic polynomial that is irreducible over F7. Put f(x) := x3 - 2 in F7[x]. Then f(0) = - 2 ≠ 0, f(1) = - 1 ≠ 0, f(3) = 6 ≠ 0, f(4) = -1 ≠ 0, f(5) = 4 ≠ 0, f(6) = 4 ≠ 0, so f(x) has no root in F7, so f(x) has no linear factor in F7[x], so f(x) is irreducible over F7. It follows that E := F7[x]/<f(x)> is a field and |E| = 73.
For a in F7, we write a in place of a + <f(x)> and u in place of x + <f(x)>. Then the elements of E have the form a + bu + cu2, with a, b, c in F7. Addition and multiplication are carried out in the usual way, and u3 is replaced by 2 wherever it occurs. Thus
(a + bu + cu2) + (d + eu + fu2) = (a + d) + (b + e)u + (c + f)u2,
and
(a + bu + cu2) (d + eu + fu2) = ad + ae u + af u2 + bd u + be u2 + bf u3 + cd u2 + ce u3 + cf u3u = (ad + 2bf + 2ce) + (ae + bd + 2cf)u + (af + be + cd)u2for a, b, c, d, e, f in F7.
(12) Let E/K be a field extension, and write G := AutKE for short. Prove that, if H is a normal subgroup of G, then H'' is also a normal subgroup of G.
Solution. We know that H'' is a subgroup of G. We need only prove that it is normal in G. This follows from the definitions:
H is normal in G,
For all f in G, for all h in H, we have f-1hf in H,
For all b in H', for all f in G, for all h in H, we have b(f-1hf) = b,
For all b in H', for all f in G, for all h in H, we have (bf-1)h = bf-1,
For all b in H', for all f in G, we have bf-1 in H',
For all a in H'', for all b in H', for all f in G, we have bf-1a = bf-1,
For all a in H'', for all b in H', for all f in G, we have bf-1af = b,
For all a in H'', for all f in G, we have f-1af in H'',
H'' is normal in G.
(13) Let E/K be a field extension, G : = AutKE and L : = G¢. Prove that E/L is Galois.
Solution. We put H : = AutLE. For any intermediate field M of E/L, we set M° : = {j Î H : aj = a for all a Î M}; for any subgroup J of H, we set J° : = {a Î E: aj = a for all j Î J}. We are to prove that L°° = L.
Since L Ê K, any automorphism of E that fixes every element of L fixes also every element of K, so H = AutLE Í AutKE = G. On the other hand, by the definition of L = G¢, any automorphism in G fixes every element of L, so any automorphism in G belongs to H and G Í H. Thus H = G.
We get now
|
|
(15) Let p, k be natural numbers, p prime. Let Fp(x) denote the pth cyclotomic polynomial over Q. Prove that, if d divides Fp(k), then d is congruent to 0 or 1 mod p.
Solution. In this solution, let ≈ stand for the congruence sign.
Suppose first that d is a prime divisor of Fp(k). Then d divides (kp-1)/(k-1), so d divides kp-1, so kp ≈ 1 mod d, so k*p = 1 in Fd×. this shows that the order of k*, as an element in the group Fd×, is a divisor of p, so it is 1 or p.
If the order of k* is 1, then k is congruent to 1 mod d, so p ≈ 1 + 1 + … + 1 ≈ kp-1 + kp-2 + … + 1 ≈ Fp(k) ≈ 0 mod d, so d divides p, and since both p and d are prime numbers, d = p. Thus d ≈ 0 mod p.
If the order of k* is p, we use the fact that the element of any finite group is a divisor of the order of that group. We obtain that p divides f(d) = d-1, so d ≈ 1 mod p.
Hence the claim is established in case d is prime. Suppose that d is not prime. Then d = 1 or d is a composite number. If d = 1, then certainly d ≈ 1 mod p. If d is composite and d = d1d2…dm is its prime decomposition, then dj ≈ 0 or 1 mod p for each j = 1, 2, … m, and it follows that d ≈ (0 or 1)(0 or 1)…(0 or 1) ≈ 0 or 1 mod p.