Professional Documents
Culture Documents
Since none of these groups have the same elementary divisors, none of them are isomorphic to each other. The Fundamental Theorem of Finite Abelian Groups, Theorem 8.12 and Lemma 8.8 tells us that these are all groups of order 720 up to isomorphism. That is, any group of order 720 is isomorphic to one of these ten groups.
a) Show that the function f : G G dened by f (x) = x1 is a group homomorphism i G is abelian b) Dene a new group H to have the same elements as G, but the operation x y = yx, where yx is dened by the operation in G. Show that the function f : G H dened by f (x) = x1 is an isomorphism of G onto H .
Proof :
a) ( = ) Suppose f : G G dened by f (x) = x1 is a group homomorphism. That is, f (ab) = f (a) f (b) for all a, b G. Since G is a group, for any a, b G, ba G and so (ba)1 G. Thus, by Corollary 7.6, ba = (ba)1 again using Corollary 7.6, that for all a, b G
ba = a1 b1
1 1
= a1 b1
= f a1 b1 = f a1 f b1 = a1
= ab.
Therefore, G is abelian.
(=) Suppose G is abelian. Consider the function f : G G dened by f (x) = x1 . Since G is abelian, for all a, b G, ab = ba; so b1 a1 = (ab)
1
= (ba)
= a1 b1 .
= b1 a1 = a1 b1 = f (a) f (b)
proving f is a homomorphism of groups. b) Let H = G and consider the group (H, ), where the operation is dened x y = yx for all x, y H and yx is dened by the operation in G. It is clear that (H, ) is a group since G is a group and how is dened. Moreover, eG x = xeG = x = eG x = x eG which show eG = eH . In addition, x1 x = xx1 = eG = eH = x1 x = x x1 , so the inverses of elements in H are the same as they were for G. Notice that since G and H are both groups, if x, y G, H , then x1 , y 1 G, H . Hence, x1 y 1 = 1 y 1 x1 = (xy) . Consider the function f : G H dened by f (x) = x1 . We must show for all x, y H , f (xy) = f (x) f (y) and that f is bijective. First, see that f is a homomorphism as, for all x, y H ,
f (xy) = (xy)
1
= y 1 x1 = x1 y 1 = f (x) f (y).
Next, suppose for any x, y H , f (x) = f (y). Recall that y 1 y = eH . Then, with x1 = y 1 , multiplying, with respect to , both right sides by y we get x1 y = eH . Associativity is true in H , so
x x1 y = x x1 y = eH y = y .
Whence, multiplying, with respect to , both left sides of x1 y = eH by x, we get y = x. Therefore, f (x) = f (y) implies x = y , proving f is one-to-one. Finally, to prove surjectivity, we must show for any h H, g G such that f (g) = g 1 = h. We luck out, since G and H are both groups which share the same elements, i.e., g H and h G, one clearly sees
g g 1 = g h = eH = g h = eH h1 = g h h1 = h1 = g .
Therefore, since h G, h1 G and it is the element we were searching for. Thus, for all h H , there 1 exists h1 G such that f h1 = h1 = h as desired, making f surjective. Having proven f is a homomorphism, injective, and surjective, it is an isomorphism by denition.
Problem 3 :
a) Show that 1 + i is not a unit in Z [i]. b) Show that 2 is not irreducible in Z [i]. c) Show that 3 is irreducible in Z[i].
Proof :
a) Notice that Z [i] = Z 2
1 is dened by
if N (1 + i) = 1 , then 1 + i is not a unit. Indeed this is precisely what we nd as Hence, 1 + i is not a unit. b) To show that 2 is not irreducible in Z [i], it suces to show that it is the product of irreducibles. Consider 1 i and 1 + i. By evaluating these elements in Z [i] with the norm, we see
N (1 i) = 12 + (1) = 2 and N (1 + i) = 12 + 12 = 2. Therefore, by Corollary 9.22 on page 310, both 1 i and 1 + i are irreducible in Z [i]. Now by considering their product, we see
2
N (1 + i) = N 1 + 1 1 = 12 (1) 12 = 1 + 1 = 2.
(1 + i) (1 i) = 1 + i i i2 = 1 (1) = 2.
Therefore, a factorization of 2 as a product of irreducibles in Z [i] is seen with 2 = (1 + i) (1 i). Ergo, 2 is not irreducible. c) To show 3 is irreducible, we will make use of Theorem 9.1. That is, we assume 3 = ab for some a, b Z [i], show that either a or b is a unit, and nally invoke Theorem 9.1 on page 286. We have shown in class that Z [i] is a Euclidean domain, so it is an integral domain by Theorem 9.8 on page 295. Now, using Theorem 9.19 on page 309, N (a) N (b) = N (ab) = N (3) = 9. But, N (a) , N (b) Z+ meaning the only possible values any one of them may take is either 1, 3, or 9. And so, N (a) = 1, 3, or 9. If N (a) = 1, then a is a unit by Theorem 9.20. If N (a) = 9, then N (b) = 1 so b is a unit by the same theorem. Clearly, these are the two cases we desire as they show either a or b is a unit. Hopefully, N (a) = N (b) = 3 is not possible. Since a Z [i], a = x + yi where x, y Z. Thus, N (a) = x2 + y 2 = 3. Most certainly, if |x| 1 and |y| 1 then x2 + y 2 = 3. However, this exhausts all possible solutions for x, y Z. Therefore, N (a) = 3 cannot possibly happen. Ergo, N (a) = 1 or N (a) = 9, which again means a or b is a unit respectively. By Theorem 9.1, as 3 = ab, 3 is irreducible.
Problem 4 :
a) Let R+ be the set of positive real numbers. Dene operations on this set by a b = ab and a b = e(ln(a) ln(b)) , where the right hand sides have the usual meaning in the real numbers. Prove that R+ is a eld with these operations by showing that ex : R R+ is an isomorphism. b) Give a direct proof from the denitions that every maximal ideal is prime c) For the ring of integers Z, describe all the prime ideals. Which ones are maximal; which ones are not? d) For the ring Z4 [x], describe all the prime ideals.
Proof :
a) Let f : R R+ be dened f (x) = ex with the operations on R+ dened above. By showing R R+ , = we can assert that R+ is a eld. Thus, it suces to show that f is an isomorphism. First, for all a, b R we see
f (a) f (b) = ea eb = ea eb = ea+b = f (a + b)
and
f (ab) = eab = e(ln(e
a
)ln(eb ))
= ea eb = f (a) f (b).
As a result, eba = 1 which, may only occur in R if b a = 0. Thus, b = a and we conclude f is injective. Finally, we must prove that f is surjective. Consider any b R+ . We must show a R such that f (a) = ea = b. All we must do is solve ea = b for a. Suppose a = ln (b). Then eln(b) = b as desired. b) Let R be a commutative ring with identity. Suppose M is some maximal ideal in R. Let ab M and a M . To show M is prime, it suces to show b M . / Consider now I := (M, a) = {r1 m + r2 a|r1 , r2 R, m M }. By Theorem 6.3, I is an ideal and M I . But M is maximal, so I = R = (1R ). Hence, there exists x, y R such that
xm + ya = 1 = bxm + bya = b.
However, R is commutative, so bxm + bya = bmx + aby = 1. As M was an ideal, since m M we note that bm M . Finally, as ab, bm M and x, y R, bmx, aby M and so
b = bmx + bay M .
Similarly, one can show a M if b M . Therefore, M is prime. / c) First, we claim that if p is prime, then the ideal (p) is prime. Proof: Suppose p is prime. By denition then, p = 1 and suppose for some a, b Z, p|ab. Then p|a or p|b, i.e., a = pr or b = pq for some r, q Z.
ab (p).
Consider (p) := {zp|z Z}, which contains all multiples of p. Since p|ab, ab is a multiple of p; therefore,
Next, as p|ab, if p|a, then a is a multiple of p and so a (p). If not, then p|b and so b (p). Hence, (p) is prime by denition. Since Z is a PID, every ideal is principle so we now know that we have a prime ideal (p) if p is prime. The question now is, have we discovered all the prime ideals. 5
We must check composite numbers. Suppose n Z is a composite number, that is n = qr for some r, q Z. By denition, 1 < |r| < |n| and 1 < |q| < |n| so we note that r and q cannot possibly be multiples of n. As such, although qr (qr) = (n), r, q (n). / Therefore, (n) is not prime if n is a composite number. Lastly, the ideal (0) is prime in Z trivially as with any integral domain, ab = 0 implies a = 0 or b = 0. This concludes a complete characterization of prime ideals in the PID Z: if p = 0 or p is prime, then (p) is a prime ideal. Now which ones are maximal? We proved problem 12 on page 305 for homework, which told us that for a PID every prime ideal is maximal. However, that's not entirely correct (we were even told to correct this) as (0) is not maximal. Clearly, (0) (2) Z, yet (2) = (0) and (2) = Z. Therefore, every one of the prime ideals (p), where 0 = p is prime is maximal. d) Consider the polynomial ring Z4 [x]. By using Theorems 6.14 and 6.15 on pages 155 and 156 respectively, we can characterize every prime ideal in Z4 [x]. In regards to Theorem 6.14 an ideal is prime if it's quotient ring forms an integral domain. But what is the issue with forming an integral domain? The problem is if we have any
f (x) , g (x) Z4 [x] and some ideal I , we need f (x) g (x) + I = 0 + I , that is f (x) g (x) 0 mod I to imply f (x) + I = 0 + I or g (x) + I = 0 + I .
But in general, f (x) g (x) 0 mod I does not imply f (x) 0 mod I or g (x) 0 mod I because of the zero divisors residing in Z4 [x]. For example 2x + 2x2
2 + 2x2 = 4x + 2x2 = 2x2 yet neither 2x nor 2 is 0 mod 2x2 .
Therefore, let Y be the set generated by all the zero divisors of Z4 [x]. Notice that 0 Y but 1 Y . We / have that Z4 [x] is a commutative ring with Y identity by Theorem 6.9. Moreover, for any f (x) , g (x) Z4 [x], if f (x) g (x) + Y = Y , then
f (x) g (x) Y . If either f (x) or g (x) is zero, then either f (x) + Y = 0 + Y or g (x) + Y = 0 + Y and we are done; as such, we consider f (x) g (x) a nonzero polynomial in Z4 [x].
Thus, there exists some h (x) Z4 [x] such that [f (x) g (x)] h (x) = h (x) [f (x) g (x)] = 0. Now if both f (x) , g (x) Y , this leads to absurdity. / Notice in terms of logic, this is a P Q statement whose negation is P Q. By commutativity and associativity,
[f (x) g (x)] h (x) = f (x) [g (x) h (x)] = 0 = h (x) [f (x) g (x)] = [h (x) g (x)] f (x),
therefore f (x) Y . Similarly, g (x) Y . So f (x) and g (x) are both in Y and not in Y . Therefore, by reductio ad absurdum either f (x) Y or g (x) Y if f (x) g (x) Y as desired. Hence, Y is a prime ideal. But we are not done, as any maximal ideal is also prime. Using Theorem 6.15, it suces to nd an ideal 4 [x] M such that ZM is a eld. Before we get too hasty, we just saw that using Y , the set generated by all the zero divisors of Z4 [x], Z4 [x] is an integral domain. Since every eld is an integral domain, Y we must simply add elements to Y until the result forms a quotient ring that is a eld. Life gets simpler now, as we claim the only elements in Z4 [x] that have multiplicative inverses belong to the
xk + 3 or 2
n k=1
are of order 2, so they are their own inverses. Now, let us baptise the set of elements of Z4 [x] with multiplicative inverses as I . Therefore, we add everything else into Y other than these elements in I and generate a new ideal M , that is M := (Y (Z4 [x] I)).
4 [x] Recall for ZM , that f (x) + M = 0 + M f (x) M . And so, we have that for any f (x) Z4 [x] 4 [x] such that f (x) M (nonzero cosets/elements of ZM ) there /
k=1
Of course the function this function that exists is f (x) itself. Therefore, 6.15, M is maximal. But every maximal ideal is prime, so M is prime.
Z4 [x] M
In general, whenever P is a prime ideal of Z4 [x] and I := {a Z4 [x] |a has an multiplicative inverse}, we can generate another prime ideal (a maximal one) M := (P (Z4 [x] I)).
Problem 5 :
a) Let F be a eld and f (x) an irreducible polynomial of degree 3 in F [x] .Show that if K is an extension eld of F of dimension 10, then f (x) is irreducible in K [x]. b) Let F be a eld and f (x) an irreducible polynomial of degree 5 in F [x] .Show that if K is an extension eld of F of dimension 7, then f (x) is irreducible in K [x].
Proof :
a) Let L = F (u), where u is any one of the three roots of f (x) F [x]. Suppose F L K , and as such, u K . However, we'll see that this leads to absurdity. Since K is an extension eld of F and u K , by Theorem 10.6, there exists p (x) F [x] that is a unique minimal polynomial and p (u) = 0. As a result, either f (x) = p (x) or p (x) |f (x). In either case, we still reach our desired result, but importantly we must notice that deg (p (x)) = 3. This is due entirely to the fact that f (x) is irreducible and p (x) , f (x) F [x], so if p (x) |f (x), then f (x) = c p (x) where c F [x] is a unit. Now, as F K , and u K is an algebraic element over F with minimal polynomial p (x) with degree 3, by Theorem 10.7, [L : F ] = [F (u) : F ] = 3. Yet, as [K : F ] = 10, we know already that [L : F ] = 3 < but we also note that [K : L] < . This was proven in class. Thus, by Theorem 10.4,
10 = [K : F ] = [K : L] [L : F ] = [K : L] 3.
This implies that [K : L] = 10 Z, but [K : L] must be a positive integer. And so, by embedding L in 3 / K , we obtain an impossible result. Thus, L = F (u) K for any of the roots of f (x). That means to say K contains no roots of f (x), call them u, v, and w Now suppose f (x) = g (x) h (x) for some g (x) , h (x) K [x]. Notice we must always satisfy the equality (by Theorem 4.2)
deg (f (x)) = deg (g (x)) + deg (h (x)).
so either g (x) or h (x) is a linear polynomial. But every linear polynomial is of the form ax + b which will have it's root in K , namely a1 b. In any event, this is not possible, as it implies that one of the roots of f (x) is in K , something we've already proven not true. Therefore, f (x) has no rst degree factor. Next, the only other scenario is if either g (x) or h (x) is a nonzero constant, this is either deg (g (x)) = 0 or deg (h (x)) = 0. In this instance, f (x) is the product of polynomial of degree 3 and a nonzero constant. By Theorem 4.11 on page 96, f (x) is irreducible. Therefore, having exhausted all the possible scenarios for f (x) K [x] by showing each case is either impossible or results in f (x) be irreducible, we conclude f (x) is irreducible in K [x] b) Using the exact same argument as before, we reach a similar conclusion. So, by letting u be any of the ve roots of f (x) F [x], let L = F (u).
Again, by supposing F L K and proceeding in the same fashion as the previous problem, we arrive at
7 = [K : F ] = [K : L] [L : F ] = [K : L] 5.
7 / This implies that [K : L] = 5 Z, but [K : L], again, must be a positive integer. And so, K contains no roots of f (x), call them u1 , u2 , u3 , u4 , and u5 .
Now suppose f (x) = g (x) h (x) for some g (x) , h (x) K [x]. Again, we use Theorem 4.2 and evaluate the possibilities of the equality
deg (f (x)) = deg (g (x)) + deg (h (x)).
If either g (x) or h (x) is a nonzero constant, then again, f (x) is irreducible by Theorem 4.11. Thus, suppose g (x) and h (x) are nonconstant polynomials in K [x]. Thus, we see that
deg (f (x)) = 5 = 4 + 1 = 1 + 4 = deg (g (x)) + deg (h (x))
or
deg (f (x)) = 5 = 2 + 3 = 3 + 2 = deg (g (x)) + deg (h (x))
And so, are really only two cases to consider. Case 1: Suppose deg (g (x)) = 4 and deg (h (x)) = 1 (the proof is the same if we assume deg (g (x)) = 1 and deg (h (x)) = 4). As such, f (x) = g (x) (x ui ) where ui is one of the ve roots (as we saw in the previous problem). However, this implies one of the ve roots, ui , are in K , which we've already proven to be false. In the end, this case is not possible and we discard it as possible scenario for expressing f (x) K [x]. Case 2: Suppose deg (g (x)) = 2 and deg (h (x)) = 3 where g (x) and h (x) are irreducible in K [x] (the proof is the same if we assume deg (g (x)) = 3 and deg (h (x)) = 2). Notice that the (deg (g (x)) , deg (h (x))) = 1. From problem 11 on page 351, which we proved in class, we deduce that [F (u, v) : F ] = [K (u, v) : K] = 6 where u is one of the two roots of g (x) and v is one of the three roots of h (x). Therefore, [K (u, v) : F ] = [K (u, v) : K] [K : F ] = 42 by Theorem 10.5. But by problem 13 of the same page (another problem proved in class) we can reason that g (x) is irreducible in K (v1 , v2 , v3 ), where vi is a root of h (x). Similarly, h (x) is irreducible in K (u1 , u2 ) where ui is a root of g (x). By Theorem 10.14, we can prove the existence of an automorphism that extends to an isomorphism between K (u1 , u2 , v1 , v2 , v3 ) and F (u1 , u2 , v1 , v2 , v3 ). By Theorem 10.5, [K (u1 , u2 , v1 , v2 , v3 ) : F ] = [F (u1 , u2 , v1 , v2 , v3 ) : F ]. However, since [K (u, v) : F ] = 42 and [F (u, v) : F ] = 6, we can easily see that this the two splitting elds do not have the same dimension over F . Therefore, this case too is unreliable. As both cases 1 and 2 are not possible, we conclude that f (x) K [x] cannot be expressed as the product of irreducible polynomials. Therefore, f (x) can only be a product of associates, meaning it must be irreducible in K [x].
5, which we know to be a homomorphism. a) What is the kernel of ? Explain why with some reference to eld theory. b) The image of is a subring of C. Describe it and its eld of quotients. No proof is needed.
Proof :
a) Note that (f (x)) := f 5 . Therefore, as ker := f (x) Z [x] | (f (x)) = f conclude that the kernel of is the set of all the polynomials in Z [x] which have
5 = 0 , we
5 as a root. By noting Z C and 5 C,we can invoke Theorem 10.6 which tells us there is a unique minimal polynomial p (x) Z [x] with 5 as a root.
As such, p (x) ker and p (x) is the unique minimal polynomial of ker. The same theorem tells us that p (x) divides every g (x) ker. Thus,
ker = {g (x) Z [x] |for some h (x) Z [x], g (x) = h (x) p (x)} = {h (x) p (x) |h (x) Z [x]}.
However, for p (x) Z [x], (p (x)) := {f (x) p (x) |f (x) Z [x]}. Ergo, ker = (p (x)). b) Consider now Im := z C| (f (x)) = f where f (x) = Thus,
(f (x)) = f
n n k=0
5 =
zk
k=0
5 .
5 =
k=0 k n
zk zk
= =
k
zk
k=0 n k=0
2k
+
n k=0
zk
k=0
2k+1
zk (5) +
k=0 n k
5i
n
2k+1
zk (5) +
k
(1) zk (5)
5.
k=0
But Z is a ring, so
a, b Z.
zk (5) Z and
5 = a + b 5 for some
k=0
k=0
We need this fact however to support the claim that Im contains a eld of quotients. This claim is supported by Theorem 9.30 on page 321. This eld, we'll call it R for now, is built up of elements of the form a , where b = 0 and a, b Im . b Consider a + b 5, c + d 5 Im . We see
a+b5 c+d 5
ac+5bd c2 +5d2
bcad c2 +5d2
5 = e + f 5
by multiplying the left hand side by the denominator's conjugate and separating terms. Again though, Z is a ring and so ac + 5bd, bc ad, c2 + 5d2 Z. Therefore, ac+5bd , cbcad2 Q. 2 +5d c2 +5d2 And so, the quotient eld R is e + f 5|e, f Q = Q
5 .
10
Problem 7 : Let F = Fq be a eld with q = pn elements. Show that every element of F can be written as
We will use the outline provided. a) Let F be a eld and |F | = 2n . By Theorem 10.25, F is a splitting eld of x2 x over Z2 . Therefore, Z2 F and F has characteristic 2; So, for all a F and
n
any n N, a2 = a.
n
Consider the function : F F dened by (x) = x2 , which we will show is an isomorphism. To see that it is a homomorphism, we see for all a, b F
(a + b) = (a + b) = a2 + b2 = (a) + (b) by the Freshman's Dream
2
and
(ab) = (ab) = a2 b2 = (a) (b) since F is a commutative ring.
2
For injectivity, let (a) = a2 = b2 = (b). Thus, a = a2 = b2 = b. Lastly, for surjectivity, for any c F , we must nd a F such that (a) = c. Since c F , c2 = c. Therefore, a2 = (a) = c = c2 and so a = c and we obtain the element we were looking for, namely (c) = c2 = c. As is an isomorphism, we now know we can express any element in F as the sum of two squares. That is, for all a F since (a) = a2 = a, we can express a as
a = a2 = a2 + 02 . b) Assume now p is odd. As such, |F | = pn and Zp F that is 0, 1 F . Dene F := F {0} and 2 := a2 |a F . Surely, F and F2 are abelian groups F
so is a group homomorphism. It is easy to see that F2 F by using the denition of F2 and if for some y, x F such that x = y 2 , then x F2 (i.e., there are elements in F that are not in F2 ) / Thus, F2 F = pn 1 as we dened F to have one less element than F by pulling out 0. We turn our focus now to the kernel of . By denition, ker := x F | (x) = x2 = 1 = x F | |x| = 2 . With no thought, we know that 1 ker, but is there anything else?
We know that Zp F , so there must be another element of order 2 in F . 2 (p n) (p n) = (p n) for p Zp and n N. As such, in Zp
(p n) = p2 2np + n2 = n2
2
In general, consider
so
(p n) = 1 n = 1.
2
Therefore, only 1 and p 1 are in ker. The question now is are there more elements in F Zp of order 2. We see that for any x F
x2 = 1 = x2 1 = 0 = (x + 1) (x 1) = 0 = x = 1, 1.
11
But, 1 Zp as it lies in the equivalence class of p 1. The result is the kernel of only contains 1 and p 1, so |ker| = 2.
By the First Isomorphism Theorem, F 2 ker. Since F is abelian, ker is normal in F . Moreover, as = F the order of F is nite, we invoke Theorem 7.36 on page 217 and see |F | |ker|
pn 1 2 . pn 1 2 .
{0} = F2 + |{0}| =
pn +1 2 . pn +1 2
Let x F be xed. From the denition of x F 2 , we notice that at once that we are subtracting n +1 distinct elements a2 F from x. In other words, there are p 2 subtractions occuring in x F 2 . Therefore, we have
pn +1 2
elements in x F 2 , that is x F 2 =
x F2
pn +1 2 .
Whence, z F such that z F 2 and z x F 2 . But this means z can be expressed as z = b2 F 2 and z = x a2 x F 2 . Then,
b2 = x a2 = x = a2 + b2 .
Since x was arbitrary, we conclude that any element x F can be expressed as the sum of two squares.
12