Algebra I & II Flashcards
- a) How many monic polynomial of degree 2 with coefficients in Fp are irreducible over Fp?
b) How many monic polynomials of degree 3 with coefficients in Fp are irreducible over Fp?
1.#irr=#all-#red
a) #all=p^2
f red => f=gh where g,h monic linear. p + pC2 linear possibilities (p for equal factors, pC2 for distinct factors) => #red=p+pC2
#irr = p^2-p-pC2=p(p-1)/2
(Couting the factors works because Fp is field so Fp[x] is PID and so also UFD.)
b) #all=p^3
If f factors into irr then f=(x^2+ax+b)(x+c) or f=(x+a)(x+b)(x+c).
By a) there are p(p(p-1)/2) for first possibility
There are p+2pC2+pC3 possibilities for the second (all the same + 2 diff + 3 diff)
#irr=p^3-p(p^2-p-pC2)-p-2pC2-pC3
#irr=p^3-p(p(p-1)/2)-p-2pC2-pC3=p(p^2-1)/3
- Let G=F*17 be the multiplicative group of the field F17. Does there exist a transitive group action of G?
a) on a set with 8 elements?
b) on a set with an odd number of elements?
- a) If H is a subgroup G acts on the quotient group G/H (G mod H) by multiplication. All cosets xH are in the orbit of H.
Let H={1,16}. Then G mod H has |G/H|=|G|/|H|=16/2=8 elements by Lagrange’s theorem.
b) Unless |T|=1, this is not possible. Since action is transitive Orb(t0)=T for any t0 in T. But by the Orbit-Stabilizer Thm we need that |T|=|Orb(t0)| divides |G|=16=2^4.
- Let G be a group of order 12.
a) Must G be abelian?
b) List all possible abelian G of order 12 (up to isomorphism)
- a) No, A4, the symmetry group of the tetrahedron (4-base/seat) is not abelian: (23)(34)=(234)!=(243)=(34)(23)
Dihedral group D12 the symmetry group of the regular hexagon (six corners) is not abelian. (Just compose a rotation and reflection)
b) Z2 x Z6 and Z12
[From Prop B-3.25 (call it Basis Theorem of finitely generated abelian groups / fundamental theorem of finite abelian groups) it follows that:
G = C(d1) x … x C(dr) s.t. d1|…|dr.
So now either r=1 and d1=12
or r=2 and d1=2 and d2=2*3=6.]
- Is the ideal generated by x^7+1 and x^5+1 a principal ideal in F2[x]? If not, why? If so, find a generator.
- (F2 is a field and F2[x] is a Euclidean ring)
Using the Euclidean algorithm:
x^7+1=x^2(x^5+1)+x^2+1
x^5+1=x^3(x^2+1)+x^3+1
x^3+1=x(x^2+1)+x+1
x^2+1=(x+1)(x+1)
So by backsubstituting, we can see that we can write x^7+1 and x^5+1 in terms of x+1. So (x^7+1,x^5+1) s.s.o. (x+1).
By polynomial division (or by using above equations):
(x^7+1)=(x+1)(x^6-x^5+x^4-x^3+x^2-x+1)
(x^5+1)=(x+1)(x^4-x^3+x^2-x+1)
We get (x+1) s.s.o. (x^7+1,x^5+1), and so (x+1)=(x^7+1,x^5+1).
- How many elements in S5 have order exactly 6?
- Let sigma be in S5 of order 6 with complete factorization sigma=alpha_1 ** alpha_n, where alpha_i is a cycle of length ri. Then 6=lcm(r1,…,rn), so sigma must have a cycle structure of (2,3) since there are no cycles of length 6. Hence there are 5C2*(3-1)!=20 elements of order 6.
- Let d be the smallest index of a proper subgroup of S8. What is d? Do there exist two distinct subgroups of S8 of index d?
- The index [G:H] of H is defined as |G/H| (the number of cosets of H), which by Lagrange is |G|/|H|.
So by definition it has to be an integer. We know A8 has index 2, so d=2.
There can’t be any others. Suppose [S8:H]=2. Then H is a normal subgroup because finite groups can only contain normal subgroups of index 2 by question (7) (ie index 2 implies normality).
[Generally speaking, {1},An,Sn are the only normal subgroups of Sn [see pdf] but specifically:]
By the second isomorphism theorem An n H is normal too. Since An is simple, either An n H is {e} or An. If it is An then H=An because |An|=|H|. If it is {e} then besides e, H only contains odd permutations (and more than one, since |An|=|H|). So let sigma and tao be odd permutations, then by uniqueness of their inverses, either sigma*tao != e or sigma^2 !=0. In either case, H would contain an even permutation, giving us a contradiction.
- Let G be a finite group. Let H be a subgroup of G which is not normal. Can the index of H be 2? Can the index of H be 3?
- 2: No, since if g not in H then gHuH=G=HuHg, so gH=G\H=Hg, i.e. H is normal.
3: Is possible, {(1),(12)} is not normal in S3, for (23)(12)(23)!=(13) iff (132)=(12)(23)!=(23)(13)=(123).
- Are the following rings UFDs? Are the rings PIDs?
a) Z[x]
b) (Z/2Z)[[x]]
c) R[x]/(x^2-1)
d) R[x]/(x^2+1)
8.
a) Z is PID (since for a,b in Z there are s,r s.t. gcd(a,b)=sa+rb), so Z UFD. By Gauss Z[x] is now also UFD. But not PID, since e.g. (2,x) is not principle.
b) It is a PID: For any given nonzero power series p, we can write it as a product of X^nq for some integer n and an invertible power series q (recall that a power series is invertible iff its constant term is a unit)
Thus, (p)=(X^nq)=(X^n). Now, for any two non zero power series p,q there are integers m,n s.t. (p,q)=(X^n,X^m)=(X^min(n,m).
[Exercise: A-3.29, (ii)
https://math.stackexchange.com/questions/1264615/inverse-rule-for-formal-power-series]
Every PID is a UFD, so it is also UFD.
c) x+1 and x-1 are zero divisors, for (x+1)(x-1)=x^2-1=0 in R[x]/(x^2-1).
d) As we know R[x]/(x^2+1)~C, so it is a field, hence PID and UFD.
- Which of the following rings are fields? Explain why or why not?
a) Z[x]/(3,x^2+x+1)
b) Z[x]/(4,x^2+x+1)
c) Z[x]/(2,x^2+x+1)
- a) Z[x]/(3,x^2+x+1)=Z3[x]/(x^2+x+1). It is not a domain: (x+2)(x+2)=x^2+4x+4=x^2+x+1=0
Additionally: Z3=F3=k is a field, and by prop (3.83) k[x]/(f) is a field iff its a domain iff f !=0 is irreducible.
b) Not a domain, since 2*2=4=0.
c) Z[x]/(2,x^2+x+1)~Z2[x]/(x^2+x+1) is a field iff x^2+x+1 is irreducible (prop 3.83), which is its since x^2+x+1=gh => g,h in {x,x+1} but no combination of these yields x^2+x+1.
- What is the automorphism group Aut(G) of the following groups G?
a) Z/2Z
b) Z/8Z
c) Z
d) S3
- a) Let sigma be an automorphism, then it is a homomorphism, so sigma(0)=0. Then sigma(1)=1. So Aut(Z2)={id}
b) sigma in Aut(Z8). Then sigma(0) and sigma is uniquely determined by sigma(1). Since sigma is an isomorphism, sigma(1) must be sent to a generator of Z8. These are 1,3,5 and 7. So, the Aut(Z8)={(0,1,2,3,4,5,6,7), (0,3,6,1,4,7,2,5),(0,5,2,7,4,1,6,3),(0,7,6,5,4,3,2,1)}.
[By the classification theorem of abelian groups every finite abelian group G is a direct sum of cyclic groups G=C(d1)+…+C(dr) where d1|…|dr. So Aut(Z8) is isomorphic to Z4 or Z2xZ2. We can see that every element is its own inverse, so Aut(Z8)=Z2xZ2.
[Lem: If a in G, phi hom, then ord(phi(a))|ord(a). So if phi is hom ord(phi(a))=ord(a).]
c) The general homomorphism is phi(n)=n*phi(1) in Z. Since we need surjectivity, phi(1) is 1 or -1. And so Aut(Z)={-id,id}
d) Again phi((1))=(1).
[(12) and (23) generate S3; since (12)(23)(12)=(13) and we can decompose any 3 cycle into these three elements.]
Since we can generate all elements with transpositions, we only need to assign the transpositions to each other to uniquely determine an automorphism. Assigning 3 elements to 3 elements is S3. So Aut(S3)~S3.
- In the ring C[x,y], do the following inclusions of ideals hold? Prove your answer:
a) (x-1,y-2) s.s.o. (y^2-4x)
b) (y^2-4x) s.s.o. (x-1,y-2)
- a) No, since y^2-4x as a poly in y has degree 2 and y-2 has degree 1.
b) Yes, since y^2-4x=(y+2)(y-2)-4(x-1)
- Let a in Z s.t. [a]=[3] mod 15, [a]=[1] mod 17.
Does this information suffice to compute a mod 255?What is [a] mod 255?
- Yes, since gcd(15,17)=1 and 15*17=255 so the existence follows form the chinese remainder theorem.
Def. a=b mod m iff m|(a-b).
=> 15|a-3 and 17|a-1 => 15p=a-3 and 17q=a-1 => there exist p and q s.t. 15p+3=a=17q+1. (p,q)=(1,1) are solutions and then a=18. So 18 solves both equivalences and a=18 mod 255 also by the chinese remainder theorem.
- Do all groups of order 4 occur as subgroups of S4? Do all groups of order n occur as subgroups of Sn?
- n=4: There are only two groups of order 4 [see lem below]. By the classification theorem of finite abelian groups these are Z2xZ2 and Z4. Z4 is isomorphic to span({(1234)}) and Z2xZ2 is isomorphic to span({(12),(34)}) both are subgroups of S4.
n=general: Let |G|=n. Put elements of G into bijective correspondence with {1,…,} by phi:G–>{1,…,n}. Now G acts on {1,…,n} by gi=phi(gphi^-1(i)) and every g corresponds to a bijective function on {1,…,n} i.e. a permutation. Thus, G is isomorphic to a subgroup of Sn.
“Let G act on itself by left-mult, this is a faithful group action and thus induces an injective homomorphism G–>Sym(G)~Sn.”
Lem: There are only 2 groups of order 4.
Pf: Because of uniqueness of inverses, at least one of a,b,c must be its own inverse. W.l.o.g. say c (w.l.o.g. because we can just switch names). Then either ab=e or not. If so then the other elements in the Cayley table follow. If not then a and b are both their own inverses. Each of a,b,c multiplied with another will then yield the third. Thus there are only 2 ways of filling out the Cayley table.
- Let G be the group of rotational symmetries of a regular solid cube.
a) How many elements does G have?
b) Does there exist a surjective homomorphism from G to S3?
- a) 1 vertex can go to 8 corners of a cube. The cube is then uniquely determined by the orientation of the 3 adjoining vertices/edges. Therefore there are 3*8 possibilities to permute the edges without breaking the symmetry.
b) Yes, each rotation takes the faces of the cube and therefore permutes the three axes (crossing through two opposite faces).
- a) Does the alternating group A17 have any subgroup of index 2 or 3?
b) Does the alternating group A17 have any elements of order 12?
- a) 2: Since A17 is simple and any group with index 2 is normal (gH=G\H=Hg) the only subgroups of A17 with index 2 are itself and {(1)}.
3: Suppose there was a subgroup H with index 3. Let A17 act on A17/H, this gives us a homomorphism phi from A17 to S3, 3 since |A17/H|=3. ker(phi) is normal in A17, which is simple and phi is non-trivial so ker(phi) is trivial. So injective. Thus, A17 is isomorphic to im(phi) some subgroup of S3. Now Lagrange’s theorem delivers the contradiction: 17!/2 | 3!.
https: //math.stackexchange.com/questions/3281671/does-a-17-have-a-subgroup-of-index-3
b) An cycle of length r has order r. We can write a cycle of length r larger than 1 as r-1 transpositions by: (1…r)=(23)…(r-1 r)(r 1). So we can write a 12 cycle as 11 transpositions and we can add a disjoint transposition to make the permutation even and in A17 (this does not change the arder since 2|12).
- Describe all possible homomorphisms:
a) f:S3–>Z6
b) g: S5–>Z2
- a) Since S3 is generated by (12) and (23) [(12)(23)(12)=(13)]. We only need to determine them. [Make a table of ordes of Z6]. Since they have order 2 and the order of their images must divide their orders and only 0,3 have orders that 2, 1 and 2 respectively, they must be sent to 0 and/or 3. Moreover, since (123) has order 3, its image must be 0,2 or 4. So 2 divides f(123)=f(12)+f(23), so f(12)=f(23). Arguing analagously gives f(132) is 0, 2 or 4 and f(12)=f(13)=f(23)= 0 or 3. So there are two possible homomorphisms: f=const=0 or f(12)=f(13)=f(23)=3 and f=0 else.
b) Since the function only takes on two values, we can determine it by its kernel. We know that ker(f) is normal in S5. The only normal subgroups of S5 are {(1)}, A5 and S5. However the function arising from ker(f)={(1)} is not a homomorphism (it isn’t injective but has trivial kernel or 1=f(123)!=f(12)+f(23)=0). ker(f)=A5 gives a type of sign homomorphism and ker(f)=S5 gives us the trivial homomorphism.
- Which of the following statements are true for all groups G of order 56?
a) They contain a subgroup of order 2.
b) They contain a subgroup of order 3.
c) The number of different subgroups of order 8 contained in G is exactly 2.
- a) Every element of a group has a unique inverse. Since the neutral element is its own inverse, there are 55 elements left. This implies that there is at least one element g in the group that is its own inverse. So {1,g} is a subgroup.
b) Suppose H is a subgroup with |H|. By Lagrange, |G|=56=2^3*7 is divisible by |H|=3. Contradiction.
c) No, consider following counterexamples:
Z7xZ8: Since 7 is prime, all elements of Z7 have order 7, except 0, which has order 1 and is the only element whose order divides 8. So every element in Z7xZ8 with order 8 must be of the form (0,x). But the subgroup {0}xZ8 already has order 8. So this is the only subgroup with order 8.
Z56: Z56 is cyclic and so all its eubgroups are cyclic too. If g has order 8 in Z56, then 56|8g iff 7|g. So g occurs is one of 7, 14, 21, 28, 35, 42, 49. But of these, we see that only 7, 21, 35 and 49 are of order 8.
- In the ring R[x], are the following subsets ideals of R? Prove your answer.
a) {f(x) in R[x] : f(pi)=0}
b) {f(x) in R[x] : f’(pi)=0}
c) {f(x) in R[x] : f(i)=0}
d) {f(x) in R[x] : f(0)=1}
- a) yes, check, std. [equal to (x-pi), since all poly with root pi are div by (x-pi)]
b) Not an ideal since for f(x)=(x-pi)^2+1 or g(x)=x^2-2pix, f’(pi)0=g’(pi) but multiplying p(x)=x does not stay in set for (pf)’(pi)=p’(pi)f(pi)+p(pi)f’(pi)=f(pi)=1!=0 and (pg)’(pi)=p’(pi)g(pi)+p(pi)g’(pi)=f(pi)=-pi^2!=0
c) Yes, check, std. Furthermore, x^2+1 is the minimal polynomial, so the ideal is (x^2+1).
d) No, for 1 the set but multiplying by x we get (x1)(0)=0, so x1 is not in set.
- Compute the conjugate of the group element a=(3524) in S5 by the element (23). How many elements of S5 are conjugate to a?
19.(23)(3524)(23)=(2534).
By Theorem A-4.7 [two elements of Sn are conjugates iff they have equal cycle structure (i.e. their complete factorizations (factorization into disjoint cycles) have the same number of r-cycles for each r in N)] all elements that are conjugate to a have the same cycle structe. So it is enough to count how many 4 cycles there are in S5. There are 5C4 choices of 4 elements of 5 and 4! ways of ordering these in a row. Since we are ordering them in a circle, we counted every ordering 4 times to many, so there are 4!*5C4/4=30.
- Let F2 be the field with two elements.
a) Show that, for every degree n g/e 1, there exists an irreducible polynomial of degree n in F2[x].
b) Let bar(F2) be the algebraic closure of F2. Show that bar(F2) is infinite.
20. a) Corollary to Galois: Let E/Fp be an extension field with p^n elements, which exists by Galois' thm (and proof, constructed field was splitting field of x^p^n-x in Fp[x]). Let alpha in E be primitive, then Fp(alpha)=E. Let g(x)=irr(alpha,Fp) in Fp[x]. It is a irr. poly with alpha as a root [and exists by Thm A-3.87i]. If deg(g)=d, then [Fp[x]/(g):Fp]=d [by Prop A-3.84v] and Fp[x]/(g)~Fp(alpha)=E [by thm A-3.87i]. Also [E:Fp]=n [since if {x1,...,xm} basis, there are p^m lincombs]. Hence, n=d, and so g is an irr. poly of degree n. b) A field is algebraically closed if every nonconstant f in K[x] has a root in K. However, if bar(F2)={a1,...,an} were finite, then f(x):=1+(x-a1)*...*(x-an) is a nonconstant polynomial in bar(F2)[x] with no roots in bar(F2). Contradiction.
- Let A be a module over the ring C[x]. Since C s.s.o. C[x], A is a C-vector space.
a) List all C[x]-modules A with dim_C(A)=1
a) List all C[x]-modules A with dim_C(A)=2
a) List all C[x]-modules A with dim_C(A)=3
21.
- What is the degree over Q of the splitting field over Q of the polynomial x^3-2?
- https://math.stackexchange.com/questions/400660/determine-splitting-field-k-over-mathbbq-of-the-polynomial-x3-2
Let K be a splitting field, a1, a2, a3 three roots of p(X)=x^3-2. Then a1/a1, a1/a2, a1/a3 are three different third roots of unity in K. As there can be at most three third roots of unity, this means that we have found all. Moreover, we know that the third roots of unity form a group. Since any finite subgroup of the multiplicative group of any field is cyclic, so is the group of third roots of unity. So let w be a generator. Now, consider the subextension E=Q(w,a1). It contains a1,a1w,a1w^2, which are three different roots of p, so E is already a splitting field of p, so K=E. Now we can determine [K:Q]: w satisfies w^2+w+1=0 (wiki: Cyclotomic poly), a1 satisfies a1^3-2=0, and these are their minimal polynomials, so [K:Q] has to be divisible by 2 and 3, hence is 6. On the other hand, the splitting field has at most dimension 3!, so indeed we have [K:Q]=6.
https://math.stackexchange.com/questions/1326078/how-should-i-find-splitting-field-of-x3-2-over-mathbb-q?rq=1
- Is the field Q[1+i] s.s.o. C the splitting field in C over Q of some polynomial over Q? If so, give an example of such a polynomial f.
- 1) Show: Q[i+1]=Q[i]=Q(i) of which f(x)=x^2+1 is a splitting field.
2) It is the splitting field of the polynomial (x-(1+i))(x-(1-i))=x^2-2x+2, which is the minimal polynomial over 1+i over Q.
- Let R be a comm. ring with unit, and let R* s.s.o. R be the group of units. Consider elements x in R* s.t. x=x^-1.
a) Show that if R=k is a field, then there are at most two such elements. When is there exactly one such element?
b) What if R is a domain? Are there still only at most two elements x in R* with x=x^-1? If so, give a proof, if not a counterexample.
c) What if we let R be a noncommutative ring with a unit? Are there still at most two such elements?
- a) i) x=x^-1 iff x^2=1 iff 0=x^2-1=(x-1)(x+1). So x=+/- 1.
a) ii) follows that exactly two sols iff +1=-1 iff char(k)=2.
b)Suppose there is a 3rd element a: s.t. a=a^-1 but different so a != 1 so a-1 !=0 and a != -1 so a+1!=0.
It then follows that (a+1)(a-1)=a^2-1=1-1=0. Which contradicts something (domain or assumptions on a).
c) Observe the ring M_n(IR). There are at least 4 such elements: ((a,0),(0,b)), where a,b in {-1,1}.
- Let GL_2(F_2) act on F_2^2{(0,0)} in the usual way. Is the induced homomorphism an isomorphism?
25.”injectivity”: iff ker phi = {1}. Let A be in G - {1}. List the elements of GL_2(F_2). For each find a vector a in F_2^2 that s.t. A*a != a.
“surjectivity”: (pigeon hole principle) follows with injectivity and |G|=|S_3|.
- Let G be a group of order p^n (p prime) acting on a finite set X. Let X^G s.s.o. X be the set of fixed points. Is it true that |X|=|X^G| mod p?
- Yes. Pf:
The orbits of G in T form a partition (t1 ~ t2 iff Ex. g s.t. g*t1=t2). So let Y be a set of representatives, then X = U orb(x) (disjoint union over x in Y). Then |X|=Sum(orb(x)) (over x in Y). By the orbit-stabilizer theorem, |orb(x)| = |G|/|Stab(x)|=p^k for some integer k=
- Give an example (with proof) of a group G which is not abelian and not isomorphic to any of the groups Sn (for n>=Sn).
- A general linear group over an infinite field like Q or R is obviously not isomorphic to Sn.
* Quaternion group Q8={+/-1,+/-i,+/-j,+/-k}. Define: i^2=j^2=k^2= -1, ij=k=-ji, jk=i=-kj, ki=j=-ik but then ji=-k
* Dihedral group D4 (symmetries of a square)
* Fundamental group of 2-rose
* A4 (alternating group n=4), has cardinality 12 but 1!=1, 2!=2, 3!=6, 4!=24. So |Sn| != |A4|
- A general linear group over an infinite field like Q or R is obviously not isomorphic to Sn.
- Calculate gcd(k^2+k+1,3k^2+4k+5) for k in Z
- Use Euclidean algorithm (Lem A-218: (i) gcd(a,b)=gcd(a,r) if b=qa+r).
gcd=3 if k~1mod3, gcd=1 else
- Let z be the real third root of 5. Is the field Q(i,sqrt(3),z) the splitting field in C over Q of some polynomial f in Q[x]? If so give an example of such a polynomial.
- The splitting field of x^3-5 is Q(i,z).
The splitting field of x^2-3 is Q(sqrt(3)).
So since Q[x] UFD the splitting field of (x^3-5)(x^2-3) is Q(i,sqrt(3),z)……maybbe this is wrong…fabian says (x^2+1)(x^3-5)
- Is sqrt(2)-sqrt(3) algebraic over Q? If so, find the minimal polynomial over Q.
- Yes, find poly by turning x=sqrt(2)-sqrt(3) into:
f(x):=x^4-10x^2+1=0. Claim that f is minimal poly.
Pf (Example A-3.89): Factor f in R[x]. As all roots and constant terms are in the form +/-sqrt(2)+/-sqrt(3). So if f is reducible in Q[x] then its factors must be quadratic. Let g|f then g(x):=(x-asqrt(2)-bsqrt(3))(x-csqrt(2)-dsqrt(3))
=x^2-((a+c)sqrt(2)+(b+d)sqrt(3))x+2ac+3bd+(ad+bc)sqrt(6). So g is in Q[x] iff linear coefficient is rational iff a+c=0=b+d but this would imply that ad+bc non zero iff constant term not in Q. So g isn’t in Q[x], f is irreducible in Q[x]. f is obviously also monic and so f is the minimal poly.
- Let Fq be the field with q:=p^n elements where p is an odd prime. For how many elements alpha in Fq does there exist a solution in Fq of the quadratic equation x^2-alpha=0?
- Claim: there are (q+1)/2 such equations.
Pf1: alpha=0 has one solution, namely alpha=0.
Define the group homomorphism phi:Fq–>Fq by phi(x):=x^2 (pf hom uses abelian). Since F field there are 2 sols to x^2=1 iff x=x^-1 by Q-24 and so ker(phi)={+/-1}. By the FIT Fq/ker(phi)~im(phi) and so (by “index formula”) |im(phi)|=|Fq|/2=(q-1)/2. Recalling that 0 is a solution we conclude.
Pf2: If beta is a solution to x^2-alpha=0, then so is -beta. Since the poly then factors, there are no more solitions and Fq[x] is UFD for each alpha there are at most two solutions. There is one solution iff beta=-beta iff beta=0. But we can choose any element in Fq as beta, so but choosing beta or -beta yields the same equation, so there are (q+1)/2 solutions.