MOPSS, 19th July 2025
(Bay Area MO 2000 P1) Prove that any integer greater than or equal to \(7\) can be written as a sum of two relatively prime integers, both greater than \(1\).
Click here for the spoiler!
Consider the case of an odd integer, the case of a multiple of \(4\), and the case of an even integer, which is not a multiple of \(4\).
Prove that the equation
\[\frac{1}{x} + \frac{1}{y} = \frac{1}{p},\]where \(x, y\) are positive integers, has exactly \(3\) solutions if \(p\) is a prime and the number of solutions is greater than three if \(p > 1\) is not a prime. We consider solutions \((a, b)\) and \((b, a)\) for \(a \neq b\) distinct.
Click here for the spoiler!
Is the given equation equivalent to
\( (x - p)(y - p) = p^2? \)
For any positive integer \(n\), show that the number of ordered pairs \((x, y)\) of positive integers for which
\[\frac 1x + \frac 1y = \frac 1n\]is equal to the number of positive divisors of \(n^2\).
Click here for the spoiler!
Is the given equation equivalent to
\( (x - n)(y - n) = n^2? \)
(India INMO 1991 P10, AoPS, cf. Moscow MO 1973 Day 1 Grade 8 P4) For any positive integer \(n\), let \(S(n)\) denote the number of ordered pairs \((x, y)\) of positive integers for which \(\frac 1x + \frac 1y = \frac 1n\) (for instance, \(S(2) = 3\)). Determine the set of positive integers \(n\) for which \(S(n) = 5\).
Click here for the spoiler!
Is the given equation equivalent to \[ (x - n)(y - n) = n^2? \]
Click here for the spoiler!
For positive integers \(x, y\), note that \[ \frac{1}{x} + \frac{1}{y} = \frac{1}{n} \] holds if and only if \[ (x - n)(y - n) = n^2 \] holds. Observe that if \(x, y\) are positive integers satisfying the given equation, then \(x > n\) and \(y > n\) holds. This shows that the solutions of the given equation over the positive integers are in one-to-one correspondence with the pairs of positive integers \((a, b)\) such that \(ab = n^2\), through the map \[ (a + n, b + n) \leftrightarrow (a, b). \] Hence, the set of positive integers \(n\) satisfying \(S(n) = 5\) is equal to the set of positive integers \(n\) such that \(n^2\) has precisely \(5\) positive divisors. Note that any such integer \(n\) is larger than \(1\). Writing \(n\) as a product of powers of distinct primes, it follows that \(n^2\) has precisely \(5\) positive divisors if and only if \(n\) is the square of a prime. Indeed, if \(p_1, \dots, p_r\) are distinct primes, and \(a_1, \dots, a_r\) are positive integers, then the integer \((p_1^{a_1} \dots p_r^{a_r})^2\) has precisely \(5\) positive divisors if and only if \[ (2a_1 + 1)(2a_2 + 1) \dots (2a_r + 1) = 5 \] holds, which is equivalent to \(r = 1, a_1 = 2\). This proves that the positive integers satisfying \(S(n) = 5\) are precisely the squares of the primes.
(UK BMO 2005 Round 2 P1) The integer \(N\) is positive. There are exactly \(2005\) ordered pairs \((x, y)\) of positive integers satisfying
\[\frac 1x + \frac 1y = \frac 1N.\]Show that \(N\) is a perfect square.
Click here for the spoiler!
- Is the given equation equivalent to
\( (x - N)(y - N) = N^2? \)
- Note that it suffices to show that if \(N^2\) has precisely \(2005\) positive divisors for some positive integer \(N\), then \(N\) is a perfect square.
- Note that \(2005 = 5 \cdot 401\).
- Observe that \(401\) is a prime, and hence, all the prime factors of \(2005\) are congruent to \(1\) modulo \(4\).
(India RMO 1994 P5, AoPS) Let \(A\) be a set of \(16\) positive integers with the property that the product of any two distinct numbers of \(A\) will not exceed \(1994\). Show that there are two numbers \(a\) and \(b\) in \(A\) which are not relatively prime.
Click here for the spoiler!
- Suppose \(\mathcal{A}\) is a set of \(16\) positive integers, and assume that \(\mathcal{A}\) does not satisfy the conclusion of the problem, that is, it is false that some two numbers in \(\mathcal{A}\) are not relatively prime. In other words, \(\mathcal{A}\) has the property that any two numbers in \(\mathcal{A}\) are relatively prime.
- Prove that such a set \(\mathcal{A}\) would fail to satisfy the given condition, which states that the product of any two distinct elements of \(\mathcal{A}\) is smaller than \(1994\). In other words, show that the product of some two elements of \(\mathcal{A}\) is greater than or equal to \(1994\).
Observe that the above argument shows that if a set of \(16\) positive integers violates the conclusion of the problem, then it does not satisfy the given condition. Convince yourself that doing so does prove that if a set of \(16\) positive integers satisfy the given condition, then there is no way that it would fail to satisfy the conclusion.
Further, also convince yourself that establishing that the failure of the conclusion forces the failure of the given condition is equivalent to establishing that the given condition implies the stated conclusion.
When there are more than one condition, the given conditions are to be considered together, and the failure of the totality of the given conditions means the failure of at least one of the given conditions.
Click here for the spoiler!
Note that if \(n \geq 1\), and \(a_1, \dots, a_n\) are distinct and pairwise coprime positive integers such that \(a_i \geq 2\) for all \(i\), then one of them admits a prime factor which is at least as large as the \(n\)-th prime. Indeed, for each \(i\), if we fix a prime divisor \(p_i\) of \(a_i\), then using that \(a_1, \dots, a_n\) are pairwise coprime, it follows that \(p_1, \dots, p_n\) are distinct primes, and hence the largest among them is at least as large as the \(n\)-th prime. Consequently, if \(n \geq 2\), and \(a_1, \dots, a_n\) are distinct and pairwise coprime positive integers, then at least \((n - 1)\) of them are greater than \(1\), and hence one of them is divisible by a prime at least as large as the \((n-1)\)-st prime. If possible, let us assume that the elements of \(A\) are pairwise coprime. Hence, \(A\) contains an element \(x\) which is divisible by a prime at least as large as the \(15\)th prime. Similarly, the set \(A\setminus \{x\}\) contains an element \(y\) which is divisible by a prime at least as large as the \(14\)th prime. Since the \(14\)th and \(15\)th primes are \(43, 47\) respectively, it follows that \(x\geq 47, y \geq 43\), and consequently, \[ xy \geq 47 \cdot 43 = 2021 >1994, \] which contradicts the hypothesis. This proves that there are two numbers \(a\) and \(b\) in \(A\) which are not relatively prime.
Note that the above problem is similar to the following problem in spirit.
For a set \(A\) of consisting of positive integers, let \(\ell(A)\) denote the largest integer which can be expressed as the product of two distinct elements of \(A\). What is the smallest element of the set which consists of the integers of the form \(\ell(A)\) as \(A\) runs over the sets of size \(16\) and consisting of positive integers?
Since \(\ell(A)\) is to be minimized as \(A\) runs over the sets \(16\) pairwise coprime integers and \(\ell(A)\) denotes the maximum of the products of the pairs of elements of \(A\), it follows that the minimum value of \(\ell(A)\) is achieved precisely when the elements of \(A\) are as small as possible. This shows that the minimum value of \(\ell(A)\) occurs when \[ A = \left\{ 1, 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47 \right\} . \] Hence, the minimum value of \(\ell(A)\) is \(43 * 47 = 2021\).
What goes wrong with the above?
(India RMO 1994 P3, AoPS) Find all \(6\)-digit natural numbers \(a_1 a_2 a_3 a_4 a_5 a_6\) formed by using the digits \(1, 2, 3, 4, 5, 6\), once each such that the number \(a_1a_2 \dots a_k\) is divisible by \(k\), for \(1 \leq k \leq 6\).
Click here for the spoiler!
- Show that \(a_2, a_4, a_6\) are equal to \(2, 4, 6\) in some order.
- Prove that \(a_5 = 5\), and \(a_1, a_3\) are equal to \(1, 3\) in some order.
- Using that \(3\) divides \(a_1 a_2 a_3\), determine \(a_2\).
- Using the divisibility condition by \(4\), show that \(a_4 = 6\), and conclude that \(a_6 = 4\).
Click here for the spoiler!
Since \(a_1a_2, a_1a_2a_3a_4, a_1a_2a_3a_4a_5a_6\) are divisible by \(2\), it follows that \(a_2, a_4, a_6\) are even, and hence they are equal to \(2, 4, 6\) in some order. Using that \(a_1a_2a_3a_4a_5\) is divisible by \(5\), we get that \(a_5 = 5\). So \(a_1, a_3\) are equal to \(1, 3\) in some order. Using that \(a_1a_2a_3\) is a multiple of \(3\), we obtain \[ a_1+ a_2+ a_3\equiv 0 \bmod 3, \] which yields \[ a_2 \equiv 2 \bmod 3, \] and hence, \(a_2 = 2\) holds. Note that \(1234, 3214\) are not divisible by \(4\). This shows that \(a_4 = 6\), and hence \(a_6 = 4\). It follows that \(a_1a_2a_3a_4a_5a_6\) is equal to \(321654\), or \(123654\). Note that the integers \(321654, 123654\) satisfy the required conditions too. This proves that \(321654, 123654\) are precisely all the \(6\)-digit numbers satisfying the given condition.
(Tournament of Towns, India RMO 1995 P3, AoPS) Prove that among any \(18\) consecutive three digit numbers there is at least one number which is divisible by the sum of its digits.
Click here for the spoiler!
- Show that one among any such consecutive integers is divisible by \(18\). - Prove that its sum of digits, is a multiple of \(9\), and conclude that it is equal to one of \(9, 18, 27\). - Show that the sum of its digits is not \(27\).
Click here for the spoiler!
Note that among \(18\) consecutive three digit numbers, there is an integer divisible by \(18\). Denote it by \(n = 100a + 10b + c\) with \(a, b, c\) denoting integers lying between \(0\) and \(9\). It follows that \(9\) divides \(n\), and hence \(9\) divides \(a + b + c\). This shows that \(a+b+c\) is equal to one of \(9, 18, 27\). Note that \(a + b + c = 27\) holds only if \(n = 999\). Since \(18\) divides \(n\), it follows that \(a + b + c \neq 27\), and hence, \(a + b + c\) is equal to one of \(9, 18\). This proves that \(a + b + c\) divides \(n\).
(Australian MO 1982, India RMO 2004 P6, AoPS) Let \((p_1, p_2, p_3, \dots , p_n, \dots)\) be a sequence of primes, defined by \(p_1 = 2\) and for \(n \geq 1\), \(p_{n+1}\) is the largest prime factor of \(p_1p_2 \cdots p_n + 1\). Prove that \(p_n \neq 5\) for any \(n\).
Click here for the spoiler!
Show that \(p_1 p_2 p_3 \cdots p_n + 1\) is odd for any \(n \geq 1\), and \(p_n\) is odd for any \(n \geq 2\). Deduce that \(p_1p_2p_3 \dots p_n + 1\) is not a multiple of \(3\). (If you are stuck, then does verifying this statement for small values of \(n\) help?) - What can be said about the smallest prime divisor of \(p_1p_2p_3 \dots p_n + 1\)? - If it is a power of \(5\), then \(p_1p_2p_3 \dots p_n\) is divisible by \(4\). Arrive at a contradiction.
Click here for the spoiler!
Note that \(p_1p_2 \dots p_n + 1\) is odd for any \(n\geq 1\), and hence \(p_n\) is odd for any \(n\geq 2\). Since \(p_1 = 2\) and \(p_2 = 3\), it follows that for any \(n\geq 2\), the integer \(p_1p_2 \dots p_n + 1\) is not divisible by any one of \(2\) and \(3\). So the least prime divisor of \(p_1p_2 \dots p_n + 1\) is at least \(5\) for any \(n\geq 2\). If possible, suppose \(5\) is the largest prime divisor of \(p_1p_2 \dots p_n + 1\) for some integer \(n\geq 2\). This yields \[ p_1p_2 \dots p_n + 1 = 5^k \] for some \(k\geq 1\). This implies that \(4\) divides \(p_1p_2 \cdots p_n\), which is impossible since \(p_1 = 2\), and \(p_r\) is odd for any integer \(r \geq 2\). This shows that \(p_{n+1}\) is not equal to \(5\) for any integer \(n\geq 2\). Consequently, it follows that \(p_n\neq 5\) for any integer \(n\geq 1\).
(India RMO 2005 P2, AoPS) If \(x, y\) are integers and \(17\) divides both the expressions \(x^2 - 2xy + y^2 - 5x + 7y\) and \(x^2 - 3xy + 2y^2 + x - y\), then prove that \(17\) divides \(xy - 12x + 15y\).
Click here for the spoiler!
- Factorize \(x^2 - 3xy + 2y^2 + x - y\) to show that \[ x \equiv y \pmod{17}, \quad \text{ or } x \equiv 2y - 1 \pmod{17} \] holds.
- Consider the above cases seperately, and use the divisibility of the other expression by \(17\) to obtain some congruence conditions on \(y\). Using these conditions to read \(xy - 12x + 15y\) modulo \(17\).
Click here for the spoiler!
Let \(x, y\) be integers such that \(17\) divides both the expressions \(x^2 - 2xy + y^2 - 5x + 7y\) and \(x^2 - 3xy + 2y^2 + x - y\). Note that \[ x^2 - 3xy + 2y^2 + x - y = (x-y)( x-2y+1), \] which is divisible by \(17\). It follows that \[ x \equiv y \bmod 17, \quad \text{ or } x \equiv 2y -1 \bmod 17 \] holds. Let us consider the case that \(x \equiv y \bmod 17\). It follows that \[ x^2 - 2xy + y^2 - 5x + 7y \equiv (x-y)^2 - 5x+ 7y \equiv 2y \bmod 17. \] Since \(17\) divides \(x^2 - 2xy + y^2 - 5x + 7y\), we get \(2y \equiv 0 \bmod 17\), which yields \(x\equiv y \equiv 0 \bmod 17\), and hence \(17\) divides \(xy - 12x + 15y\). Let us consider the case that \(x \equiv 2y -1 \bmod 17\). Using \(x^2 - 2xy + y^2 - 5x + 7y\equiv 0 \bmod 17\), we obtain \[ (2y -1 )^2 - 2(2y -1 )y + y^2 - 5(2y-1) + 7y\equiv 0 \bmod 17, \] which yields \(y^2-5y + 6 \equiv 0 \bmod 7\). This implies that \((y-2)(y-3)\equiv 0 \bmod 17\). This shows that either \(x\equiv 3\bmod 17, y \equiv 2 \bmod 17\) holds, or \(x \equiv 5 \bmod 17, y \equiv 3 \bmod 17\) holds. If \(x\equiv 3\bmod 17, y \equiv 2 \bmod 17\) holds, then \[ xy - 12x + 15y \equiv 6 - 36 + 30 \equiv 0 \bmod 17 \] holds. If \(x \equiv 5 \bmod 17, y \equiv 3 \bmod 17\) holds, then we obtain \[ xy - 12x + 15y \equiv 15 - 60 + 45 \equiv 0 \bmod 17. \] This proves that \(17\) divides \(xy - 12x + 15y\).
Lecture notes for Math Olympiad (IOQM, RMO, INMO)
Click on the icons below to download.
Topics Links Algebra Combinatorics Geometry Number Theory INMO Training Camp 2025, MP IMO Training Camp MOPSS