Posted on

Table of Contents

Q1

Prove there are infinitely many primes such that p1(6)p \equiv -1 (6).

Proof:

We need to prove there are infinitely many prime such that 6(p+1)    p=6k16|(p+1) \implies p = 6k -1

Assume there are finite primes such that p=6k1p = 6k-1. There will be a new prime q=6(p1p2...pn)1q=6(p_1p_2...p_n)-1. This shows q is yet another prime. So there are infinitely many primes.

Q3

For a series aia_i, we can represent any decimal in i=0ai10i\sum_{i=0}a_i10^i. Prove that if 3i=0ai3|\sum_{i=0}a_i or 9i=0ai9|\sum_{i=0}a_i ,the decimal will be divisible by 3 or 9. Also true for 11i=1(1)iai11|\sum_{i=1}(-1)^ia_i.

Proof:

It is obviously that 10i1(3)10^i \equiv 1(3) and 10i1(9)10^i \equiv 1(9). And ai10iai(3)a_i10^i \equiv a_i (3). So if ai10iai(3)\sum a_i 10^i \equiv \sum a_i (3).

Also 1010(11)10 \equiv 10 (11) and 1001(11)100 \equiv 1 (11) that is 10i(1)i(11)10^i \equiv (-1)^i (11).

Q4,5

Prove 3x2+2=y23x^2+2=y^2 and 7x3+2=y37x^3+2=y^3 have no solutions.

Proof

Consider 3x3+2y2(3) 3x^3 +2 \equiv y^2 (3) We already have 3x20(3)3x^2 \equiv 0 (3), so 2y2(3)2 \equiv y^2 (3).

For mod 3, we have 020(3),121(3),221(3)0^2 \equiv 0 (3), 1^2 \equiv 1 (3), 2^2 \equiv 1(3). So y22(3)y^2 \equiv 2 (3) is impossible.

Consider 7x3+2y3(7) 7x^3 + 2 \equiv y^3 (7) We need to show 2y3(7)2 \equiv y^3(7).

It is impossible cause y3(7)y^3 (7) is 0,1,60,1,6.

Q6

If a group integers a1,a2,a3,...,aϕ(n)a_1,a_2,a_3,...,a_{\phi(n)} such that (ai,n)=1(a_i,n)=1. We will have (a,n)=1(a,n)=1 and aa1,aa2,...,aaϕ(n)aa_1,aa_2,...,aa_{\phi(n)} such that (aai,n)=1(aa_i,n)=1.

Proof:

It is obvious that (a,n)=1,(b,n)=1    (ab,n)=1(a,n)=1, (b,n)=1 \implies (ab,n)=1.

On the other hand, assume a U(n)={a1,a2,...,aϕ(n)}U(n)=\{a_1,a_2,...,a_{\phi(n)}\}. {aai}\{aa_i\} is just a permutation of U(n)U(n).

Q7

Prove Euler's Theorem.

Proof:

Assume A=aia2a3...aϕ(n)A = a_ia_2a_3...a_{\phi(n)} and aϕ(n)A=aa1aa2...aaϕ(n)a^{\phi(n)}A = aa_1aa_2...aa_{\phi(n)}. That is $$

By Q6: aϕ(n)AA(n)a^{\phi(n)}A \equiv A (n). Also since (A,n)=1(A,n)=1, so it's safe to cancel A on both side. aϕ(n)1(n) a^{\phi(n)} \equiv 1 (n)

Q8

Let pp is an odd prime. If k{1,2,...,p1}k\in \{1,2,...,p-1\}, there will be a unique bkb_k such that kbk1(p)kb_k\equiv1(p).

Proof Since p is prime, we have (k,p)=1(k,p)=1.

So by Fermat's little theorem: kp11(p) k^{p-1}\equiv 1 (p) So we have a solution bk=kp2b_k = k^{p-2}

Assume there are another solution kbk1(p)kb_k'\equiv 1(p). k(bkbk)0(p) k(b_k - b'_k) \equiv 0 (p) That is p(bkbk)p|(b_k - b_k') which is unique in mod p system, thus bkbk=0b_k-b'_k=0​.

In this way, when k1,kp1k\neq 1, k\neq p-1, kbkk\neq b_k.

Q9

Prove (p1)!1(p)(p-1)!\equiv -1(p).

Proof:

From Q8, we know that there are pairs in 1,2...p11,2...p-1. For each pair, we have kbk1(p)kb_k\equiv 1(p). Except k=1,k=p1k=1,k=p-1. Since p11(p)p-1 \equiv -1 (p).

So (p1)!1(p)(p-1)!\equiv -1(p).

Q10

Prove (n1)!0(n)(n-1)!\equiv0(n) if n is not prime, except when n=4n=4.

Proof:

In general, n=abn=ab where a<n,b<na<n,b<n. So ab(n1)!ab|(n-1)!

When n=4n=4, (n1)!=123=6(n-1)! = 1*2*3=6. So 62(4)6 \equiv 2 (4). The reason 4 is an exception is that the only composition 4=224=2*2 and 2 is only shown once in 3!3!.

Q11

For a reduced residue system modulo nn: {a1,a2,...,aϕ(n)}\{a_1,a_2,...,a_{\phi(n)}\}. Assume NN is the number of solutions of x21(n)x^2 \equiv 1(n). Show that a1a2...aϕ(n)(1)N/2(n)a_1a_2...a_{\phi(n)}\equiv (-1)^{N/2}(n).

Proof:

By Wilson's Theorem in Q9, we can notice that in a reduced residue system, for each k{ai}k\in\{a_i\}, we will have a bkb_k as it's inverse so kbk1(n)kb_k\equiv 1(n), except when k21(n)k^2\equiv1(n)​.

Consider n=piain=\prod p_i^{a_i}.

Given that each piaip_i^{a_i} will provide a residue system modulo that ai1(n)\prod a_i \equiv -1 (n) since k21(n)k^2 \equiv 1(n) will provide 2 solutions (1,n1)(1,n-1). When n=2n=2, the product will be 1, that will not affect anything.

Then there will be (1)N/2(-1)^{N/2} .

Q12

Prove that p(pk)p|\binom{p}{k}, 0<k<p0<k<p

Proof:

By definition: (pk)=p!k!(pk)! \binom{p}{k} = \frac{p!}{k!(p-k)!} Since p!=(p1)!pp!=(p-1)!*p. And p is prime. So no factors in k!(pk)!k!(p-k)! since k<p,pk<pk<p,p-k<p. So pp in upper fraction must remain. That is p(pk)p|\binom{p}{k}

Q13

Prove Fermat's Theorem

Proof:

By Q12, consider : (a+1)pap1=k=1p1(pk)apk (a+1)^p - a^p -1= \sum_{k=1}^{p-1}\binom{p}{k}a^{p-k} Since we know that p(pk)p|\binom{p}{k} so each part in the sum can be divided by p.

That is: (a+1)pap+1(p) (a+1)^p \equiv a^p+1 (p) When a=0a=0, obviously 11(p)1\equiv 1 (p) .

When a>0a > 0, assume apa(p)a^p \equiv a (p) (Fermat). We need to prove for a+1a+1 it still hold. (a+1)pa+1(p) (a+1)^p \equiv a + 1(p) Since we already have apa(p)a^p \equiv a(p). So: (a+1)pap+1(p) (a+1)^p \equiv a^p +1 (p) That is just we proved before.

So Fermat's Theorem holds.

Q14

For p,qp,q are distinct odd primes such that p1q1p-1|q-1. Also if (n,pq)=1(n,pq)=1, prove that nq11(pq)n^{q-1}\equiv 1(pq).

Proof:

By Fermat's Theorem: np11(p) n^{p-1} \equiv 1 (p)

nq11(q) n^{q-1} \equiv 1(q)

Where nq1=nk(p1)n^{q-1}=n^{k(p-1)}

So we have nq11(p) n^{q-1} \equiv 1 (p) That gives: nq11(pq) n^{q-1} \equiv 1 (pq)

Q15

Prove the numerator of 1+1/2+1/3...1/p11+1/2+1/3...1/p-1 is divisible by pp where p is prime.

Proof: numerator=i=2pp!i \text{numerator} = \sum_{i=2}^{p}\frac{p!}{i} Obviously, each term is still integer and all divisible by p.

Q17

Prove for f(x)Z[x]f(x) \in \Z[x] and n=piain=\prod p_i^{a_i}, prove that f(x)0(n)f(x) \equiv 0 (n) has a solution iff f(x)0(piai)f(x) \equiv 0 (p_i^{a_i}) has solution for all ii

Proof:

If f(x)0(n)f(x)\equiv 0 (n), we have nf(x)n|f(x). So for all piaif(x)p_i^{a_i}|f(x) that is f(x)0(piai)f(x) \equiv 0 (p_i^{a_i})

If f(x)0(piai)f(x) \equiv 0 (p_i^{a_i}), by Chinese Reminder Theorem, (piai,pjaj)=1(p_i^{a_i},p_j^{a_j})=1, then f(x)0(piai)f(x) \equiv 0 (\prod p_i^{a_i}) have a unique solution.

Q18

If there are NiN_i solutions for f(x)0(piai)f(x) \equiv 0 (p_i^{a_i}). Then there are Ni\prod N_i solutions for f(x)0(n)f(x) \equiv 0(n).

Proof:

By Chinese Reminder Theorem, for each group of equations: {xxi(piai)} \{x \equiv x_i (p_i^{a_i})\} There are one unique solution. So there are Ni\prod N_i different equations set. So there are Ni\prod N_i unique solutions for f(x)0(n)f(x) \equiv 0 (n)

Q19

Show that x21(pa)x^2\equiv 1(p^a) has only solutions {1,1}\{1,-1\}.

Proof:

We need to show the solution of: (x+1)(x1)0(pa) (x+1)(x-1) \equiv 0 (p^a) The difference of (x+1),(x1)(x+1),(x-1) is 2, which is smaller than any pap^a.

That is, (x+1)(x1)(x+1)(x-1) cannot contain more than 1 p as factor.

That is, there are only 2 solution x=1,x=1x=1,x=-1.

Q20

Show that x21(2b)x^2\equiv 1(2^b) has one solution when b=1b=1 and 2 solution when b=2b=2 and four solution when b3b\geq3.

Proof: (x+1)(x1)0(2b) (x+1)(x-1)\equiv 0(2^b) When b=1b=1, {1,1}\{1,-1\} are same solutions in mod 2.

When b=2b=2, {1,1}\{1,-1\} two solutions.

When b3b\geq3, (x1),(x+1)(x-1),(x+1) can be two even number in sequence. So when x=2b11x=2^{b-1}-1 or x=2p1+1x = 2^{p-1}+1 will also be valid solutions.

Q21

Find the number of solutions x21(n)x^2 \equiv 1(n)

Proof:

For the factorization of n n=piain = \prod p_i^{a_i}.

Assume ν(piai)\nu(p_i^{a_i}) is the solution number of x21(piai)x^2 \equiv 1 (p_i^{a_i}).

For pi=2p_i=2 ν(2)=1;ν(22)=2;ν(2a)=3,a3 \nu(2) =1; \nu(2^2)=2; \nu(2^a)=3, a \geq 3 For other pip_i, ν(piai)=2\nu(p_i^{a_i})=2.

So the solution will be: ν(n)=ν(piai) \nu(n) = \prod \nu(p_i^{a_i})

Q23

Prove a+bi0(1+i)a+bi \equiv 0 (1+i) or a+bi1(1+i)a+bi \equiv 1 (1+i)

Proof:

Assume p+qip+qi such that: (p+qi)(1+i)=a+bi (p+qi)(1+i) = a+bi

(a+b2+ba2)(1+i)=a+bi (\frac{a+b}{2}+\frac{b-a}{2})(1+i)=a+bi

So when a+ba+b and bab-a are both even, there are a+bi0(1+i)a+bi \equiv 0 (1+i)

Assume p+qip+qi such that: (p+qi)(1+i)=a1+bi (p+qi)(1+i)=a-1+bi

(a+b12+ba+12i)(1+i)=a1+bi (\frac{a+b-1}{2}+\frac{b-a+1}{2}i)(1+i)=a-1+bi

So when a+ba+b and bab-a are odd, we have a+bi1(1+i)a+bi\equiv 1(1+i)

If a,ba,b are both even, a+ba+b and bab-a shall both be even.

If a,ba,b are both odd, a+ba+b and bab-a shall both be even.

In other situation, a+ba+b and bab-a shall both be odd.

Q24

Prove a+bω{1,1,0}(1ω)a+b\omega \equiv \{-1,1,0\} (1-\omega)

Proof:

By definition ω2+ω+1=0\omega^2+\omega+1=0. ω1(1ω) \omega \equiv 1 (1-\omega)

a+bωa+b(1ω) a+b\omega \equiv a+b (1-\omega)

So this can be represent to any congruences.

The order of 1ω1-\omega is (1ω)(1+ωˉ)=(1ω)(1ω2)=2ωω2=3(1-\omega)(1+\bar\omega) = (1-\omega)(1-\omega^2)=2-\omega-\omega^2=3

So this is equal to mod 3\text{mod}\ 3: {0,1,2}\{0,1,2\}

Where 21(3)2 \equiv -1 (3)

Q25

For λ=1ω\lambda = 1-\omega, α1(λ)\alpha \equiv 1(\lambda), prove that α31(9)\alpha^3 \equiv 1(9)

Proof:

First we shall see: ω2(1ω)2=3ω(1+ω)=3 -\omega^2(1-\omega)^2 = -3\omega(1+\omega)=3 Then we assume: α=1+λ \alpha = 1+\lambda

α3=1+3λ+3λ2+λ3 \alpha^3 = 1 + 3\lambda + 3\lambda^2 + \lambda^3

α3=1ω2λ3ω2λ4+λ3 \alpha^3 = 1 - \omega^2\lambda^3 -\omega^2\lambda^4 + \lambda^3

We need to show that the last three part is divisible by 9. λ4=9ω2;λ3=3(2ω+1) \lambda^4 = 9\omega^2; \lambda^3 = -3(2\omega+1) So we need to prove that 3(1ω2)(2ω+1) 3|-(1 - \omega^2)(2\omega+1) Obviously (ω+2)(2ω+1)=3ω-(\omega+2)(2\omega+1)=3\omega

Q26

For ζ,η,ξZ[ω]\zeta,\eta,\xi \in \Z[\omega] such that η3+ζ3+ξ3=0\eta^3 + \zeta^3 + \xi^3 = 0 , λ\lambda must divides at least one of them.

Proof:

It is obvious there are only 3 reminder for α{1,1,0}(λ)\alpha \equiv \{1,-1,0\}(\lambda).

By Q25, α3=9k+{1,1,0}\alpha^3 = 9k + \{1,-1,0\}

So η3+ζ3+ξ3=0\eta^3 + \zeta^3 + \xi^3 = 0 is equivalent to η+ζ+ξ=0\eta + \zeta + \xi = 0

That is at least one of them is η0(λ)\eta \equiv 0 (\lambda)