2019-Diophantine Equations

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

2019 Winter Camp Diophantine Equations James Rickards

Diophantine Equations
James Rickards
[email protected]

1 Introduction
A classic number theory problem takes the form

Find all (positive) integer/rational solutions to the following equation/inequality

Such a question is called a Diophantine


√ equation. Despite the area being thousands of years
old (it’s roots being the irrationality of 2), it is still one of the top current research topics.

As with most of math, there is no single method that will solve all Diophantine equations. If
you write down a completely random equation, it’s fairly likely to be quite easy or not possible to
solve with elementary techniques.

2 Strategies
Here are some basic strategies and tips to remember:

• Search for small solutions: if there are non-trivial ones, this may be worth a point if you get
them all;

• Look modulo primes/prime powers to deduce information about the variables;

• Try to rearrange and factor terms; show that such factors have bounded gcd;

• Rewrite equations involving squares, look for squares with a small difference between them;

• Bound the sizes of different variables;

• Sometimes casework is needed, get comfortable of when. Sometimes casework devolves into
more casework and will never end, and sometimes it doesn’t: it’s good to have intuition about
if your casework is useful or not.

Some simple examples are:

Example 1. Find all integer solutions to y 2 = x5 + 7.

Solution. Modulo 11, the possible fifth powers are −1, 0, 1. Thus y 2 ≡ 6, 7, 8 (mod 11), and these
are all not quadratic residues modulo 11. Thus there are no solutions.

Example 2. Find all positive integer solutions to y 2 = (x2 + 1)(2x2 + 6).

1 of 8
2019 Winter Camp Diophantine Equations James Rickards

Solution. The gcd of the right hand side divides (2x2 + 6) − 2(x2 + 1) = 4, so it is 1, 2, 4. Thus
we either have x2 + 1, 2x2 + 6 being both squares, or twice squares. In the first case, we let
(x2 + 1, 2x2 + 6) = (a2 , b2 ), so then a2 − x2 = 1. Thus a > x, and write a = x + r with r > 0 an
integer. Thus 1 = r2 + 2rx ≥ 1 + 2 = 3 > 1, contradiction.
Otherwise, (x2 + 1, 2x2 + 6) = (2a2 , 2b2 ), whence b2 − x2 = 3. As above, b = x + r for r > 0, whence
3 = r2 + 2xr ≥ 1 + 2 = 3, with equality iff r = x = 1. This gives rise to (x, y) = (1, 4), which is
thus the only solution.

3 Two Variable Polynomials


Let f (x, y) be a two variable integer polynomial, and consider the equation f (x, y) = 0 to be solved
in the rational numbers. For a generic such equation, we can assign it a non-negative integer g,
called the genus. For example, the curves:

1. x2 + y 2 = 1 has genus 0

2. y 2 = x3 + x + 1 has genus 1

3. y 2 = x5 + x has genus 2

A genus zero curve will have a family of solutions which can be solved easily. For example,

Example 3. Find all rational solutions to x2 + y 2 = 1.

Solution. If x = 1 then y = 0 necessarily. Next starting with the rational solution (1, 0), the general
equation of a line with rational slope through this point is y = tx − t for any rational number t.
Plugging this in to our original equation, we have (t2 + 1)x2 − 2t2 x + t2 − 1 = 0. But we know that
2 −1
x = 1 is a solution, so this quadratic must factor! The solutions multiply to x = tt2 +1 , hence this
is the other root. But all rational solutions can be constructed in this manner, as the line through
the solution and (1, 0) will have finite rational slope. Thus the solution set is
 2 
t − 1 2t
(x, y) = 2 , , (1, 0)
t + 1 t2 + 1

for any rational t.

A genus 1 curve can have either finitely or infinitely many solutions, and such an equation is
called an Elliptic Curve. Studying these equations is one of the hottest topics in current number
theory, and they are still very mysterious. If one of these equations appears on a contest, then it
will most likely have finitely many solutions, as there is no elementary (high school level) way to
describe the infinite families of solutions.

Finally, a curve of genus at least 2 has finitely many rational solutions by Faltings’s theorem.
So you could eventually find all solutions by ordering all possible (x, y) values and checking them,
but you will have no way of knowing if you’ve found them all or need to search more...

2 of 8
2019 Winter Camp Diophantine Equations James Rickards

4 Local Global Principle


If f (x1 , . . . , xn ) is an integer polynomial, then f = 0 having an integer solution will imply that
there is a solution in R and modulo p for all primes p. The opposite implication is very closely
related to something called the “local-global principle.” This principle holds if f is a quadratic
polynomial, but does not hold in generality. For example,

(x2 − 2)(x2 − 3)(x2 − 6) = 0

has solutions in R and modulo all primes p, but no solution in Q.


Even though the principle does not always hold, looking modulo p is still very useful.

5 A Couple Useful Theroems


Theorem 1 (Lifting the exponent lemma). Let p be prime, x, y integers, and n a positive integer
such that p | x − y and p - x, y. Then we have

• If p is odd, then
vp (xn − y n ) = vp (x − y) + vp (n);

• If p = 2, then
v2 (xn − y n ) = v2 (x − y) + v2 (n) + v2 (x + y) − 1.

Lifting the exponent (LTE) can be proved by induction (useful exercise!), and can be quite
useful.

Theorem 2 (Zsigmondy’s Theorem). Let a > b > 0 be coprime positive integers. Given an integer
n ≥ 1, there exists a primitive prime divisor of an − bn , i.e. a prime p such that p | an − bn but
p - ak − bk for all 1 ≤ k ≤ n − 1, except in the following cases:

• n = 1, a − b = 1 (there are no prime factors);

• n = 2, a + b is a power of 2;

• n = 6, a = 2, b = 1.

Exercise 1. Prove Zsigmondy’s Theorem (or just the special case of a = 2, b = 1) using lifting the
exponent lemma and bounding (if all prime factors came from previous terms, we can use LTE to
know exactly how much they contribute, and can show that there is not enough there).

Zsigmondy’s Theorem is very powerful and is often overkill, but it has its uses.

Corollary 1. Given any integer a > 1 and any n 6= 2, 6, there is a prime p for which a has order
n modulo p (and of course n = 2, 6 will work in all but the exceptional cases described above).

Corollary 2. 2n + 1 has a prime divisor that is at least as big as 2n + 1 for all n 6= 3.

3 of 8
2019 Winter Camp Diophantine Equations James Rickards

4 of 8
2019 Winter Camp Diophantine Equations James Rickards

6 Problems
I will adopt the David Arthur style of ordering problems into 3 sections: A,B,C. A-level problems
should be approximately CMO level, B level problems would be easy-medium IMO problems, and
C level would be medium-hard IMO or beyond problems. This ordering is of course somewhat
subjective, so don’t be surprised if you find some problems to be out of place.

A1. Find the number of ordered pairs of positive integer solutions (m, n) to the equation 20m +
12n = 2012.

A2. Find the positive integer n such that n13 = 21982145917308330487013369.

A3. Given that 34! = 295232799cd96041408476186096435ab000000 in base 10, find a, b, c, d.

A4. Find all positive rational solutions (x, y) to x2 + 3y 2 = 1.

A5. Find all integers a such that x3 − x + a has three integer roots.

A6. Find the positive integer n such that n5 = 1335 + 1105 + 845 + 275 .

A7. Find all integer solutions (x, y) to x3 y 2 (2y − x) = x2 y 4 − 36.

A8. Show that |12m − 5n | ≥ 7 for all positive integers m, n.

A9. Find all integer solutions to 6(6a2 + 3b2 + c2 ) = 5n2 .

B1. Determine all pairs of rational numbers (x, y) such that x3 + y 3 = x2 + y 2 .

B2. Find all pairs (x, y) of integers which satisfy y 2 = x3 + 16.

B3. Find the integer solutions to y 3 = 3x2 + 3x + 7.

B4. Find all triples (a, b, c) of positive integers such that a!b! = a! + b! + c!.

B5. Find all integer triples (a, b, c) with a > 0 > b > c and a + b + c = 0 such that 2017 − a3 b −
b3 c − c3 a is a perfect square.

B6. Find all pairs (m, n) of non-negative integers such that m2 + 2 · 3n = m(2n+1 − 1).

B7. Prove there are unique positive integers a, n such that an+1 − (a + 1)n = 2001.

B8. Find all pairs of positive integers (m, n) satisfying 3m − 7n = 2.

B9. Let a, b, c be positive integers such that gcd(a, b) = 1 and c is coprime to at least one of a, b.
Prove that there exist infinitely many triples (x, y, z) of distinct positive integers such that
xa + y b = z c .

C1. Find all positive integers N such that x2 + y 2 + N (xy + 1) = 0 has an integer solution (x, y).

C2. Find all positive integers N such that x2 + y 2 − N xy + 1 = 0 has a positive integer solution
(x, y).
x7 −1
C3. Find all integer solutions to x−1 = y 5 − 1.

5 of 8
2019 Winter Camp Diophantine Equations James Rickards

C4. Find all non-negative integer solutions (x, y, z, w) to the equation 2x 3y − 5z 7w = 1.

C5. Find all pairs of positive integers m, n ≥ 3 such that there are infinitely many positive integers
m +a−1
a such that aan +a2 −1 is an integer.

C6. Integers x > 2,y > 1,z > 0 satisfy xy + 1 = z 2 . If ω(n) denotes the number of distinct prime
divisors of n, prove that ω(x) ≥ ω(y) + 2.

C7. Show that there are infinitely many pairs (x, y) of rational numbers such that x3 + y 3 = 9.

C8. Show that there are no positive integer solutions (x, y) to (x + 1)(x + 2) · · · (x + 2014) =
(y + 1)(y + 2) · · · (y + 4028).

C9. Solve x2 + 7 = 2n in the integers.

6 of 8
2019 Winter Camp Diophantine Equations James Rickards

7 Hints
A1. Use mods.

A2. Look mod 10, approximate the number.

A3. Find max power of 10 dividing 34!, then look at the rest mod 9, 10, 11.

A4. Look at the example from the lecture and emulate it.

A5. Use Vieta’s formulas.

A6. Look modulo small primes and bound n.

A7. Rearrange to 36 = (xy(x − y))2 , and do some casework (many solutions!).

A8. If not show that |12m − 5n | = 1, and rule this out modulo 11.

A9. Look modulo 2, 3 and get an infinite descent argument so show only a = b = c = n = 0 works.

B1. Write x = yk for k rational (and don’t forget about the case of y = 0).

B2. Bring 16 to the left hand side and factor.

B3. Show that y ≡ 1 (mod 3) and then look modulo 9.

B4. Try small cases to find the one solution. Assume c ≥ b ≥ a and divide by a!, look modulo n
for appropriate n.

B5. Sub in for c, simplify and factor to get the equation a2 + ab + b2 = N ; solve this.

B6. Solve for m, get a new equation involving only n. Look modulo various primes or use lifting
the exponent lemma to finish.

B7. Look modulo a, a + 1 to pin down a, and then find the unique n.

B8. Look at the equations modulo good choices of primes when n ≥ 2 or m ≥ 3.

B9. There are in fact solutions with xa = y b , try with this simplification.

C1. N = 5 is the only N , try Vieta jumping/polynomial division.

C2. N = 3 is the only N , try Vieta jumping/polynomial division.


xp −1
C3. Recall the lemma: prime divisors of x−1 (p is a prime) are all 0 or 1 modulo p.

C4. There are four solutions. You can solve pretty much the whole problem with modular arith-
metic.

C5. Think in terms of polynomials, and play around with the algebra. There is one pair (m, n)
that works.
ay
C6. Manipulate to obtain the equation 4 − by = ±1, and recall the lemma that gcd(xm − 1, xn −
1) = xgcd(m,n) − 1.

7 of 8
2019 Winter Camp Diophantine Equations James Rickards

C7. Try substituting x = a+b a−b


2 and y = 2 . After another inversion/rescaling get an equation of
2 3
the form A = B + c for some constant c, and show this has infinitely many solutions by
using two solutions to construct a third (this is an elliptic curve with rank at least 1).

C8. Show that x has to be quite large, and bound x between two quadratics in y.

C9. The possible n are n = 3, 4, 5, 7, 15. Do the n even case, and for n odd factorize in Q( 7)
(this is really beyond the level of olympiads, unless there is an alternate solution)

8 of 8

You might also like