Maths Olympiad Prep

Library / /22 of 34

, 2010

Number theory Difficulty 6.2 National olympiad Prove it Austria

Determine all tripels (x,y,z)(x, y, z) of positive integers x>y>z>0x > y > z > 0, such that x2=y2x+1x^2 = y \cdot 2^x + 1 holds.

Solution

We first note that the right-hand side of the equation is odd. We therefore know that x2x^2 is odd, and therefore xx is odd. One of the neighbors of xx must therefore be divisible by 44, and we therefore have x=2pa±1x = 2^p a \pm 1 with p>1p > 1 and aa odd.

Let us first assume x=2pa+1x = 2^p a + 1. If a>1a > 1, we first note that x21=2p+1a(2p1a+1)x^2 - 1 = 2^{p+1} a (2^{p-1} a + 1) holds, and therefore, since yy must contain all odd factors of x21x^2 - 1,
y3(2p1a+1)>2pa+1=x, y \geq 3(2^{p-1} a + 1) > 2^p a + 1 = x,
which contradicts x>yx > y. We therefore have a=1a = 1, and x=2p+1x = 2^p + 1. The only possible value for yy is then 2p1+12^{p-1} + 1, since b(2p1+1)>2p+1=xb(2^{p-1} + 1) > 2^p + 1 = x holds for any b>1b > 1. Any possible solution in this case must therefore have x=2p+1x = 2^p + 1, y=2p1+1y = 2^{p-1} + 1 and therefore z=p+1z = p + 1. If p=2p = 2, we obtain x=5x = 5 and y=z=3y = z = 3, which contradicts y>zy > z. For any p>2p > 2 however, we obviously have x>yx > y and y>zy > z holds, since 2p1+1>p+12p1>p2^{p-1} + 1 > p + 1 \Longleftrightarrow 2^{p-1} > p is certainly true. All tripels (2p+1,2p1+1,p+1)(2^p + 1, 2^{p-1} + 1, p + 1) are therefore solutions for p>2p > 2.

Now let us assume x=2pa1x = 2^p a - 1. If a>1a > 1, we have x21=2p+1a(2p1a1)x^2 - 1 = 2^{p+1} a (2^{p-1} a - 1), and therefore
y3(2p1a+1)=2pa1+2p1a2>2pa1+2p2>x, y \geq 3(2^{p-1} a + 1) = 2^p a - 1 + 2^{p-1} a - 2 > 2^p a - 1 + 2^p - 2 > x,
which again contradicts x>yx > y, and we once again have a=1a = 1. If x=2p1x = 2^p - 1, possible values for yy are either y=2p11y = 2^{p-1} - 1 or y=2(2p11)y = 2(2^{p-1} - 1), since b(2p11)>2p1b(2^{p-1} - 1) > 2^p - 1 for any b>2b > 2. We therefore have two further groups of solutions. In the first case, we have x=2p1x = 2^p - 1, y=2p11y = 2^{p-1} - 1 and therefore z=p+1z = p + 1. If p=2p = 2, we obtain x=3x = 3, y=1y = 1 and z=3z = 3, which contradicts y>zy > z. If p=3p = 3, we obtain x=7x = 7, y=3y = 3 and z=4z = 4 again contradicting y>zy > z. If p4p \ge 4 however, y>zy > z holds, since 2p11>p+12p1>p+22^{p-1} - 1 > p + 1 \Longleftrightarrow 2^{p-1} > p + 2 is certainly true. All tripels (2p1,2p11,p+1)(2^p - 1, 2^{p-1} - 1, p + 1) are therefore solutions for p>3p > 3.

Finally, if x=2p1x = 2^p - 1 and y=2p2y = 2^p - 2, we have z=pz = p. If p=2p = 2, we obtain x=3x = 3 and y=z=2y = z = 2, which again contradicts y>zy > z. If p>2p > 2 however, y>zy > z holds, since 2p2>p2p>p+22^p - 2 > p \Longleftrightarrow 2^p > p + 2 is again true. All tripels (2p1,2p2,p)(2^p - 1, 2^p - 2, p) are therefore also solutions for p>2p > 2.

Summarizing, the solutions are given by the tripels
(2p+1,2p1+1,p+1)for p>2,(2p1,2p11,p+1)for p>3,and (2p1,2p2,p)for p>2. \begin{array}{l} (2^p + 1, 2^{p-1} + 1, p + 1) \quad \text{for } p > 2, \\ (2^p - 1, 2^{p-1} - 1, p + 1) \quad \text{for } p > 3, \\ \text{and } (2^p - 1, 2^p - 2, p) \quad \text{for } p > 2. \end{array}

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement and solution reproduced as published; topic and difficulty added by this site.