OMC Number Theory Lecture
Uploaded by
Deeba ShahOMC Number Theory Lecture
Uploaded by
Deeba ShahOMC Number Theory Lecture
Shresth Suraiya
June 2023
1 Introduction
Hey folks! This is Shresth Suraiya. In this class, we will be discussing a few
concepts and problems in Number Theory. Enjoy Solving!
2 Introductory Problems
Problem 2.1 (APMO). Determine all positive integers n for which
n2 + 1
√ 2
[ n] + 2
is an integer. Here [r] denotes the greatest integer less than or equal to r.
Problem 2.2 (Balkan MO). Find all ordered pairs of positive integers (x, y)
such that:
x3 + y 3 = x2 + 42xy + y 2
3 Intermediate Problems
Problem 3.1 (All-Russia MO). Given are positive integers a, b satisfying a ≥
2b. Does there exist a polynomial P (x) of degree at least 1 with coefficients
from the set {0, 1, 2, . . . , b − 1} such that P (b) | P (a)?
Problem 3.2 (China TST). Let a1 , a2 , · · · , an be a permutation of 1, 2, · · · , n.
Among all possible permutations, find the minimum of
n
X
min{ai , 2i − 1}.
i=1
4 Advanced Problems
Problem 4.1 (China TST). For positive integer k > 1, let f (k) be the number
of ways of factoring k into product of positive integers greater than 1. The order
1
of factors are not considered. For example f (12) = 4, as 12 can be factored in
these 4 ways: 12, 2 · 6, 3 · 4, 2 · 2 · 3.
Prove that if n > 1, and p is a prime dividing n, then f (n) ≤ np
Problem 4.2. Let N denote the set of all positive integers.
Let ⟨an ⟩n≥0 , ⟨bn ⟩n≥0 be two sequences of positive integers such that
a1 = 1, bn = an + n,
{⟨an ⟩n≥0 } ∩ {⟨bn ⟩n≥0 } = ϕ, {⟨an ⟩n≥0 } ∪ {⟨bn ⟩n≥0 } = N
Show that there do exist such sequences.
Problem 4.3. Let a and b be co-prime positive integers with different parity.
Define the sequence ⟨xn ⟩n≥0 as
xn+1 xn−1 + 3xn
x0 = a, x1 = b, and =
xn xn + 3xn−1
for all positive integers n. Find all ordered pairs (a, b) for which each term of
⟨xn ⟩n≥0 is an integer.
5 Solutions
5.1 Introductory Problems
APMO 2013 P2 We show that there are no such n.
√
Let [ n] = k. Then, n = k 2 + l where 0 ≤ l ≤ 2k =⇒ n2 = k 4 + 2k 2 l + l2
Then, (k 2 + 2) | (k 4 + 2k 2 l + l2 + 1) =⇒ (k 2 + 2) | (l2 − 4l + 5)
=⇒ 1 + (l − 2)2 = m(k 2 + 2)
Note that m ≥ 1 and 1 + (l − 2)2 ≤ 1 + (2k − 2)2 = m(k 2 + 2) =
4k 2 − 8k + 5 < 4(k 2 + 2) So, m ∈ {1, 2, 3}. For m = 1, we get 1 + (l − 2)2 =
k 2 + 2 which yields no solutions. For m = 2, we get 1 + (l − 2)2 =
2k 2 + 4 =⇒ (l − 2)2 − 2k 2 = 3. But 2k 2 ≡ 0, 2 (mod 8) =⇒ (l − 2)2 ≡ 3, 5
(mod 8), a contradiction as (l − 2)2 ≡ 0, 1, 4 (mod 8). For m = 3, we get
1 + (l − 2)2 = 3k 2 + 6 =⇒ (l − 2)2 ≡ 2 (mod 3) which is impossible as
(l − 2)2 ≡ 0, 1 (mod 3). Therefore, there are no solutions.
Balkan MO 2017 P1 Let d = gcd(x, y) =⇒ x = ad , y = bd =⇒
(d(a + b) − 1)(a2 + b2 − ab) = 43ab
We deduce that a, b are coprime with a2 + b2 − ab. Thus we have that
a2 − ab + b2 = 43 =⇒ (a, b) = (x, y) = (1, 7) = (7, 1) or we have that
a2 − ab + b2 = 1 =⇒ (a, b, d) = (1, 1, 22) =⇒ (x, y) = (22, 22)
2
5.2 Intermediate Problems
ARMO 2023 Grade 10, P3 The answer is YES for b ≥ 2. For b = 1, the polynomial is constant.
Write a − b in base-b representation. We have
(a − b)b = kn kn−1 . . . k2 k1 k0 b =⇒ a−b = kn bn +kn−1 bn−1 +· · ·+k1 b+k0
Keep in mind that n ≥ 1 because a − b ≥ b.
Consider P (x) = kn xn +kn−1 xn−1 +· · ·+k1 x+k0 . Note that its coefficients
are indeed from the set {0, 1, 2, . . . , b − 1}. We have
P (b) = a − b | P (a) − P (b) =⇒ P (b) | P (a).
Lastly, the degree of P is n ≥ 1, so P is not constant, as desired.
China TST 2020 P5 Claim: If a < b, c < d then min{a, c} + min{b, d} ≥ min{a, d} + min{b, c}
Proof: It suffices to show min{a, c} − min{b, c} ≥ min{a, d} − min{b, d}.
Let f (x) = min{x, c} − min{x, d}. Note for x ≤ c, f (x) = 0, and for
c < x ≤ d, f (x) = c − x and if x > d, f (x) = c − d. Therefore, f never
increases.
Therefore, the minimum can be obtained with ai are strictly decreasing,
for if ai < aj , i < j then we swap ai , aj .
5.3 Advanced Problems
China TST 3 2014 P6 Clearly we can just assume that p is the largest prime dividing n. We’ll
call the ways to split it up partitions f (n).
Lemma: ϕ(n) ≥ np for the largest prime p|n.
Proof: We have ϕ(n) = n q|n q−1 i−1 n
Q Q
q ≥n i≤p i =p.
Now we use induction. By the hypothesis, f (x) ≤ xp ≤ ϕ(x). Take the
largest prime factor p of a number np. Now for this p look at which factor
of the partition that its in. Say that it is part of pd. Then the remaining
n/d can be partitioned in f (n/d) ways. Note that if p is not the unique
largest, we are
P overcounting P by a lot, so summing over all divisors d of
n, f (np) ≤ d|n f (d) ≤ d|n ϕ(d) = n which is well known. So our
induction is complete.
Submit your solutions to Problems 4.2 and 4.3 at my Email ID,
[email protected] ! (These are my original problems, so you may share
your feedback as well!)
