Number Theory: Problems and Solutions from Math 126 Final Exam May 2006, Exams of Number Theory

Solutions to selected problems from the final exam of math 126 number theory course held in may 2006. The problems cover topics such as rational solutions of polynomial equations, euclidean algorithm, diophantine equations, continued fractions, and chinese remainder theorem.

Typology: Exams

2012/2013

Uploaded on 02/18/2013

sangha
sangha 🇮🇳

4.3

(43)

87 documents

1 / 3

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Math 126, Number Theory
Final Answers
May 2006
Scale. 85–100 A. 70-84 B. 55 69 C. Median 86.
Problem 1. [20] Carefully prove the following statement.
If x=r/s is a rational solution of the equation
anxn+an1xn1+· ·· +a1x+a0= 0
where all the coefficients an, an1, . . . , a1, a0are in-
tegers, and if rand sare relatively prime, then r
divides the constant term a0while sdivides the
leading coefficient an.
Naturally, there are many proofs. Here’s a straightforward
one.
Proof: Let x=r/x be a solution with rand srelatively
prime. Then
an(r/s)n+an1(r/s)n1+· · · +a1(r/s) + a0= 0.
Multiplying each side of the equation by sn, we have
anrn+an1rn1s+· ·· +a1rsn1+a0sn= 0.
All but one of the terms contains a factor of r, so, moving
the remaining term to the other side, we can rewrite this
equation as
(anrn1+an1rn2s+· ·· +a1sn1)r=a0sn.
Since rdivides the left hand side of the equation (by the
definition of divisibility), therefore rdivides the right hand
side a0sn. But ris relatively prime to s, therefore ris
relatively prime to sn. But rdivides a0sn, hence rdivides
a0, and thus rdivides a0.
Likewise, all but one of the terms contains a factor of s,
so we can rewrite the equation as
(an1rn1+· ·· +a1rsn2+a0sn1)s=anrn.
Since sdivides the left hand side, it also divides anrn. But
sbeing relatively prime to r, it is also relatively prime to
rn. Hence sdivides an, and so also divides an.q.e.d.
Problem 2. [24; 6 points each part] The Euclidean algo-
rithm has been essential for us this semester.
a. Choose two of the following numbers, call them mand
n, and show how the Euclidean algorithm is used to find
their greatest common divisor d= (m, n).
2907,3128,4807,9775
Each choice of mand nleads to different greatest common
divisors. Let’s take m= 4807 and n= 3128. Then
4807 = 3128 + 1679
3128 = 1679 + 1449
1679 = 1449 + 230
1449 = 6 ·230 + 69
230 = 3 ·69 + 23
69 = 3 ·23
Therefore, the greatest common divisor dis 23.
b. Use the work you did in part a to show how dis a linear
combination of mand n, that is, solve the linear Diophantine
equation mx +ny =d.
One way to do that is to work the equations in part a
backwards.
23 = 230 3·69
= 230 3·(1449 6·230)
= 19 ·230 3·1449
= 19 ·(1679 1449) 3·1449
= 19 ·1679 22 ·1449
= 19 ·1679 22 ·(3128 1679)
= 41 ·1679 22 ·3128
= 41 ·(4807 3128) 22 ·3128
= 41 ·4807 63 ·3128
c. Continuing with this example, show how you can use
your results in part b to solve the linear congruence mx
6d(mod n).
That linear congruence is equivalent to the linear Dio-
phantine equation
mx +ny = 6d.
1
pf3

Partial preview of the text

Download Number Theory: Problems and Solutions from Math 126 Final Exam May 2006 and more Exams Number Theory in PDF only on Docsity!

Math 126, Number Theory

Final Answers

May 2006

Scale. 85–100 A. 70-84 B. 55 69 C. Median 86.

Problem 1. [20] Carefully prove the following statement.

If x = r/s is a rational solution of the equation

anxn^ + an− 1 xn−^1 + · · · + a 1 x + a 0 = 0

where all the coefficients an, an− 1 ,... , a 1 , a 0 are in- tegers, and if r and s are relatively prime, then r divides the constant term a 0 while s divides the leading coefficient an.

Naturally, there are many proofs. Here’s a straightforward one.

Proof: Let x = r/x be a solution with r and s relatively prime. Then

an(r/s)n^ + an− 1 (r/s)n−^1 + · · · + a 1 (r/s) + a 0 = 0.

Multiplying each side of the equation by sn, we have

anrn^ + an− 1 rn−^1 s + · · · + a 1 rsn−^1 + a 0 sn^ = 0.

All but one of the terms contains a factor of r, so, moving the remaining term to the other side, we can rewrite this equation as

(anrn−^1 + an− 1 rn−^2 s + · · · + a 1 sn−^1 )r = −a 0 sn.

Since r divides the left hand side of the equation (by the definition of divisibility), therefore r divides the right hand side −a 0 sn. But r is relatively prime to s, therefore r is relatively prime to sn. But r divides −a 0 sn, hence r divides −a 0 , and thus r divides a 0. Likewise, all but one of the terms contains a factor of s, so we can rewrite the equation as

(an− 1 rn−^1 + · · · + a 1 rsn−^2 + a 0 sn−^1 )s = −anrn.

Since s divides the left hand side, it also divides −anrn. But s being relatively prime to r, it is also relatively prime to rn. Hence s divides −an, and so also divides an. q.e.d.

Problem 2. [24; 6 points each part] The Euclidean algo- rithm has been essential for us this semester.

a. Choose two of the following numbers, call them m and n, and show how the Euclidean algorithm is used to find their greatest common divisor d = (m, n).

2907 , 3128 , 4807 , 9775

Each choice of m and n leads to different greatest common divisors. Let’s take m = 4807 and n = 3128. Then

4807 = 3128 + 1679 3128 = 1679 + 1449 1679 = 1449 + 230 1449 = 6 · 230 + 69 230 = 3 · 69 + 23 69 = 3 · 23

Therefore, the greatest common divisor d is 23.

b. Use the work you did in part a to show how d is a linear combination of m and n, that is, solve the linear Diophantine equation mx + ny = d. One way to do that is to work the equations in part a backwards.

23 = 230 − 3 · 69 = 230 − 3 · (1449 − 6 · 230) = 19 · 230 − 3 · 1449 = 19 · (1679 − 1449) − 3 · 1449 = 19 · 1679 − 22 · 1449 = 19 · 1679 − 22 · (3128 − 1679) = 41 · 1679 − 22 · 3128 = 41 · (4807 − 3128) − 22 · 3128 = 41 · 4807 − 63 · 3128

c. Continuing with this example, show how you can use your results in part b to solve the linear congruence mx ≡ 6 d (mod n). That linear congruence is equivalent to the linear Dio- phantine equation

mx + ny = 6d.

Since d = 41m − 63 n, therefore 6d = 6 · 41 m − 6 · 63 n = 246 m − 378 n. Thus, x ≡ 246 (mod n) is a solution.

d. The Euclidean algorithm also is used to find continued fraction expansions. Find the continued fraction expansion of m/n, and show your work.

You can find this as a restatement of part a. 4807 3128

Alternatively, you can read the coefficients 1, 1, 6, 3, 3 di- rectly from the equations in part a.

Problem 3. [18] Consider the system of three congruences

x ≡ 4 (mod 11) x ≡ 3 (mod 9) x ≡ 3 (mod 10)

Since the moduli are pairwise relatively prime, the Chinese remainder theorem says that there is a unique solution to this system modulo 990, the product of the moduli. Find that solution and show your work.

There are a couple of ways to do this. One is to take the first two congruences and replace them by a single congru- ence modulo 99, then take that congruence along with the third to get a single congruence modulo 990. Here’s a different method that uses all three together. Step 1. For each modulus, find a reciprocal of the product of the remaining moduli modulo the given modulus. For the first modulus, 11, that means we need the reciprocal of 90 modulo 11, that is, we need to solve

90 y ≡ 1 (mod 11).

That’s the same as 2y ≡ 1 (mod 11), and that can be easily found by searching to be y ≡ 6 (mod 11). Thus, 6 is the reciprocal we’re looking for. For the second modulus, 9, we need the reciprocal of 110 modulo 9. That’s the same as the reciprocal of 2 modulo 9, which is 5.

For the third modulus, 10, we need the reciprocal of 99 modulo 10. That’s the same as the reciprocal of 9 modulo 10, which is 9. Step 2. To get x sum three products abc, one for each congruence, where a is the constant in the congruence, b is the product of the other moduli, and c is the reciprocal found in the previous step. That gives us

4 · 90 · 6 + 3 · 110 · 5 + 3 · 99 · 9 = 2160 + 1650 + 2673 = 6483

and then reduce this number modulo the product 990 of all three moduli. That gives a final answer of x ≡ 543 (mod 990).

Problem 4. [18] (page 107, exercise 3) Suppose that (a, n) = 1. Prove that

ab^ ≡ ac^ (mod n)

if and only if b ≡ c (mod ordn(a)). (You may use the theorems in the text, of course.) Of course, there are many proofs. The key theorem 3. is the one that says

ax^ ≡ a (mod n) iff x ≡ 1 (mod ordn(a)).

Here’s a proof that uses this theorem in both directions. Proof ⇒: Let ab^ ≡ ac^ (mod n). Suppose first that b ≥ c. Then the congruence ab^ − ac^ ≡ 0 (mod n) can be written as ac(ab−c^ − 1) ≡ 0 (mod n). Thus, n|ac(ab−c^ − 1). But (a, n) = 1, so (ac, n) = 1, and therefore n|(ab−c^ − 1). That says ab−c^ ≡ 1 (mod n). Therefore, by theorem 3.26, b ≡ c (mod ordn(a)). Thus, we’ve shown that if b ≥ c, then b ≡ c (mod ordn(a)) as required. But if c ≥ b, the same argument holds with b and c interchanged in the argument. Thus, in all cases b ≡ c (mod ordn(a)). Proof ⇐: Let b ≡ c (mod ordn(a)). Again, first take the case that b ≥ c. Then b = c+k ordn(a) for some nonnegative integer k. Therefore, modulo n,

ab^ ≡ ac+k^ ordn(a) ≡ ac(aordn(a))k ≡ ac 1 k ≡ ac

Likewise, if c ≥ b, we can also show ab^ ≡ ac^ (mod n). q.e.d.

Problem 5. [20; 10 points each part] Consider the Pell equation x^2 − 30 y^2 = 1. a. Find solution to the equation from the continued frac- tion expansion of

30, which is √ 30 = 5 +