You are on page 1of 3

18.

781 Homework 4

Due: 4th March 2014

Q 1 (2.2(5)(e-g)). Find all solutions to the equations (e): 64x ≡ 83(mod 105), (f ): 589x ≡ 209(mod 817),
(g): 49x ≡ 5000(mod 999).

Proof. (e)
By running the Euclidean algorithm, we get 25 ∗ 105 − 41 ∗ 64 = 1, so the unique solution to this equation
is given by x ≡ −41 ∗ 83(mod 105) ≡ 33 ∗ 105 − 41 ∗ 83(mod 105) = 62(mod 105).

(f )
The Euclidean algorithm tells us that (589, 817) = 19. Now observe 31 ∗ 19 = 589, 11 ∗ 19 = 209 and
43 ∗ 19 = 817. So we now look for solutions to the equation 31y ≡ 11(mod 43). Running the Euclidean
algorithm, we get 13 ∗ 43 − 18 ∗ 31 = 1, so y ≡ −18 ∗ 11(mod 43) ≡ 43 ∗ 5 − 18 ∗ 11(mod 43) ≡ 17(mod 43).
So the solutions to the original equation are 17 + 19k(mod 817) for 0 ≤ k ≤ 18.

(g)
Euclidean algorithm gives us 367 ∗ 49 − 18 ∗ 999 = 1. 5000(mod 999) = 5(mod 999). So the unique solution
is given by x ≡ 5 ∗ 367(mod 999) ≡ 5 ∗ 367 − 999(mod 999) ≡ 836(mod 999).

Q 2 (2.2(8)). Show that if p is an odd prime, then the congruence x2 ≡ 1(mod pα ) has only the two solutions
x ≡ 1(mod pα ), x ≡ −1(mod pα ).
Proof. α ≥ 1. So, x2 − 1 ≡ 0(mod pα ) implies x2 − 1 ≡ 0(mod p). x2 − 1 = (x − 1)(x + 1). So p | x2 − 1
implies p | x − 1 or p | x + 1.
If p | x − 1, then x + 1 ≡ 2(mod p), so p - x + 1 as p is odd. This tells us that x + 1 and pα are coprime,
and so there is a y(mod pα ) such that (x + 1)y ≡ 1(mod pα ). In this case, if x2 − 1 ≡ 0(mod pα )) implies
(x2 − 1)y ≡ 0(mod pα ) which simplifies to x − 1 ≡ 0(mod pα ).
If p mod x + 1, then x − 1 ≡ −2(mod p) ≡ p − 2(mod p), so p - x + 1 as p is odd. So just as before, we
get that x − 1 and pα are coprime, so we can find a y such that (x − 1)y ≡ 1(mod pα ). x2 − 1 ≡ 0(mod pα ))
then tells us that (x2 − 1)y ≡ 0(mod pα ) which simplifies to x + 1 ≡ 0(mod pα ).

Q 3 (2.2(17)). Let the numbers ci be defined by the power series identity

(1 + x + x2 + . . . + xp−1 )/(1 − x)p−1 = 1 + c1 x + c2 x2 + · · ·

Show that ci ≡ 0(mod p) for all i ≥ 1.

Proof.
p−1   ∞ 
! !
(1 + x + x2 + . . . + xp−1 ) 1 − xp (1 − xp ) − (1 − x)p

X p X −p
= = 1 + =1+ (−x)k x k
(1 − x)p−1 (1 − x)p (1 − x)p k k
k=1 k=0

1
p
 Pp−1 p

Recall we showed in the previous assignment that p | k for 1 ≤ k ≤ p − 1, so k=1 k (−x)k = pxf (x) for
some polynomial f (x) ∈ Z[x], so we have

(1 + x + x2 + . . . + xp−1 )
− 1 = pxg(x)
(1 − x)p−1
P∞
where g(x) is the power series f (x) k=0 −p −p
 k 
k x which has integer coefficients as k ∈ Z for all integers
k ≥ 0.

Q 4 (2.3(3)). Solve the set of congruences: x ≡ 1(mod 4), x ≡ 0(mod 3) and x ≡ 5(mod 7).
Proof. Following the proof of the Chinese Remainder theorem in the book, we have that the product of the
three integers is 84, and we have 84 84 84
4 .1 ≡ 1(mod 4), 3 .1 ≡ 1(mod 3) and 7 .3 ≡ 1(mod 7). So a solution to
these congruences is given by 21 + 0 + 180 = 201. An integer satisfies these congruences if and only if it is
of the form 201 + 84k for some integer k.

Q 5 (2.3(7)). Determine whether the congruences 5x ≡ 1(mod 6), 4x ≡ 13(mod 15) have a common solution,
and find them if they exist.
Proof. 6 | (5x − 1) if and only if 6 | 5(5x − 1) (as (5, 6) = 1), i.e., if and only if 6 | (x − 5). 15 | (4x − 13) if
and only if 15 | 4(4x − 13) (as (4, 15) = 1), i.e., if and only if 15 | (x − 7). Now, 6 | x − 5 implies 3 | (x − 5),
i.e., x ≡ 5 ≡ 2(mod 3). 15 | (x − 7) implies 3 | x − 7, i.e. x ≡ 7 ≡ 1(mod 3). So these two equations cannot
have a common solution.
Q 6 (2.3(10)). Find the number of positive integers ≤ 3600 that are prime to 3600.
Proof. This is just ϕ(3600). 3600 = 24 ∗ 32 ∗ 52 . So ϕ(3600) = (24 − 23 )(32 − 3)(52 − 5) = 8 ∗ 6 ∗ 20 = 960.
Q 7 (2.3(15)). Solve the congruence x3 + 4x + 8 ≡ 0(mod 15).

Proof. A solution to x3 + 4x + 8 ≡ 0(mod 15) would give a solution to x3 + 4x + 8 ≡ 0(mod 5), but there
are no solutions mod 5 as one can check by hand that r(mod 5) for 0 ≤ r ≤ 4 do not satisfy the equation,
so there are no solutions to the original equation.
Q 8 (2.3(24)). Suppose that m1 , . . . , mr are pairwise relatively prime positive integers. For each j, let
C (mj ) denote a complete system of residues modulo mj . Show that the numbers c1 + c2 m1 + c3 m1 m2 + · · · +
cr m1 m2 . . . mr−1 , cj ∈ C (mj ), form a complete system of residues modulo m = m1 m2 . . . mr .
Proof. As there is a bijection between C (m1 ) × C (m2 ) × · · · C (mr ) and the complete system of residues
modulo m, it is enough to show that given any tuple (a1 , a2 , . . . , ar ) ∈ C (m1 ) × C (m2 ) × · · · C (mr ), we can
find cj ∈ C (mj ) such that c1 + c2 m1 + c3 m1 m2 + · · · + cr m1 m2 . . . mr−1 ≡ aj (mod mj ). We will define the
cj by induction. Notice

c1 + c2 m1 + c3 m1 m2 + · · · + cr m1 m2 . . . mr−1 ≡ c1 + c2 m1 + c3 m1 m2 + · · · + cj m1 m2 . . . mj−1 (mod mj )

So set c1 = a1 . Having defined c1 , c2 , . . . cj such that c1 + c2 m1 + c3 m1 m2 + · · · + ck m1 m2 . . . mk−1 ≡


ak (mod mk ) for 1 ≤ k ≤ j, let cj+1 be the unique element of C (mj+1 ) such that

cj+1 m1 m2 . . . mj ≡ aj+1 − (c1 + c2 m1 + c3 m1 m2 + · · · + cj m1 m2 . . . mj−1 )(mod mj+1 )

There is a unique solution to this equation as (m1 m2 . . . mj , mj+1 ) = 1 (this was because (mi , mj+1 ) = 1 for
i 6= j + 1 by Theorem 2.3(3)).
Q 9 (2.3(33)). Find the smallest positive integer n so that ϕ(x) = n has no solution; exactly two solutions;
exactly three solutions; exactly four solutions.

2
Proof. ϕ(1) = 1 = ϕ(2). ϕ(3) = 2. For n = 3, there are no solutions to ϕ(x) = 3. For if 3 =
αp
− pαp −1 ) for S some subset of primes and αp ≥ 1, then 3 = pα−1 (p − 1) for some prime p
Q
p∈S (p
(as 3 is prime), so this tells us by unique factorization that either pα−1 = 3 and p − 1 = 1 or pα−1 = 1
and p−1 = 3. The latter is not possible as 4 is not a prime, and the former is not possible as 2α 6= 3 for any α.

For n = 1, ϕ(x) = 1 has exactly two solutions 1 and 2. For if p > 2, then ϕ(pα ) ≥ p − 1 Q ≥ 2.
So if x Q
is a solution to ϕ(x) = 1, then x cannot have any prime factor other than 2 for if x = pα ,
ϕ(x) = (ϕ(pα )) ≥ ϕ(pα ). So if at all ϕ(x) = 1 has to have some other solution other than x = 1, then
x = 2α for some α. Again, if α > 1, then ϕ(2α ) = 2α−1 ≥ 2.

ϕ(x)
Q = 2 has exactly three solutions 3, 4, 6. This can be argued again using the fact that 2 is prime. If
x = pαp , then no prime strictly bigger than 3 can appear in this product as otherwise, ϕ(x) ≥ p − 1 ≥ 4.
So we are looking for pairs (α1 , α2 ) so that x = 2α1 3α2 is a solution to ϕ(x) = 2. As 3 | ϕ(3α 2 ) if α2 > 1
and 3 - 2, we see that α2 ∈ {0, 1}. If α2 = 1, then α1 is either 0 or 1 as then we are looking for solutions to
ϕ(2α1 ) = 1. This gives x = 3 and x = 6. If α2 = 0, then ϕ(2α1 ) = 2α1 −1 = 2, so α1 = 2. This gives x = 4.

ϕ(x) = 4 has exactly four solutions 8, 12, 5, 10. Firstly observe that x cannot have any prime factor p
bigger than 5 as then ϕ(x) ≥ ϕ(pα ) ≥ p − 1 > 4 for p > 5 and α > 1. So we want to find all triples
(α1 , α2 , α3 ) such that each αi ≥ 0 and x = 2α1 3α2 5α3 is a solution to ϕ(x) = 4. ϕ(x) = ϕ(2α1 )ϕ(3α2 )ϕ(5α3 ).
As 3 | ϕ(3α2 ) if α2 > 1 and 3 - 4, α2 ∈ {0, 1}. Similarly α3 < 2 as 5 - 4.
If α3 6= 0, then α2 = 0. ϕ(x) = ϕ(2α1 )ϕ(3α2 )ϕ(5α3 ) = 4.ϕ(2α1 )ϕ(3α2 ), so ϕ(x) ≥ 8 if α2 = 1. So the
only possible triples are (0, 0, 1) and (1, 0, 1), as the only solutions to ϕ(y) = 1 are 1 and 2. This gives the
integers 5 and 10.
If α3 = 0, then either α2 = 1 or α2 = 0. If α2 = 1, then we are looking for solutions to ϕ(2α1 ) = 2α1 −1 = 2,
so this gives α1 = 2. This gives 12. If α2 = 0, then we are looking for solutions to ϕ(2α1 ) = 2α1 −1 = 4, so
this gives α1 = 3. This gives 8.
Q 10 (2.3(38)). Let (a, b) = 1 and c > 0. Prove that there is an integer x such that (a + bx, c) = 1.
Proof. Let {p1 , . . . , pr } be all the prime factors of b and let {q1 , q2 , . . . , qs } be any additional primes that
appear in the prime factorization of c. (r + s 6= 0 as c 6= 0). To show that (a + bx, c) = 1 for some x, it is
enough to show (a + xb, pi ) = 1 for all i and (a + xb, qj ) = 1 for all j for that x.
Now for any integer x, as pi | b,
(a + bx, pi ) = (a, pi ) = 1
This is because (a, b) = 1, so a and b cannot have any common prime factor. If s = 0, take x = 0.
Otherwise, we have to find an x such that

(a + xb, qj ) = 1

for all j.
Now as (b, qj ) = 1, for each j, we can find a cj such that cj b ≡ 1(mod qj ). Using the Chinese Remainder
theorem, find an x such that x ≡ cj (1 − a)(mod qj ). Now, this would tell us that bx ≡ bcj (1 − a) ≡
1 − a(mod qj ) or in other words a + bx ≡ 1(mod qj ). So (a + bx, qj ) = 1.
Q 11 (2.3(40)). Prove that for n ≥ 2, the sum of all positive integers less than n and prime to n is nϕ(n)/2.
Proof. Let S be the set of positive integers less than n and prime to n. Then ](S) = ϕ(n). If 1 ≤ a ≤ n − 1,
then a ∈ S if and only if n − a ∈ S as (a, n) = (n − a, n). This tells us that f : S → S defined by f (a) = n − a
is a bijection. X X X X X
2 a= a+ f (a) = (a + f (a)) = n = nϕ(n)
a∈S a∈S a∈S a∈S a∈S
nϕ(n)
This tells us that the sum of all positive integers less than n and prime to n is equal to 2 .

You might also like