To bound or not to bound!

Consider the following diophantine equation in natural numbers x + y 2 + z 3 = x y z x+y^2+z^3=xyz where z = gcd ( x , y ) z=\gcd(x,y) . Let ( x 1 , y 1 , z 1 ) , ( x 2 , y 2 , z 2 ) , . . . , ( x n , y n , z n ) (x_1,y_1,z_1), (x_2,y_2,z_2), ..., (x_n,y_n,z_n) be the all triples which satisfy the equation. Find k = 1 n ( x k + y k + z k ) . \sum_{k=1}^{n} (x_k+y_k+z_k).

Details and Assumptions

gcd ( a , b ) \gcd(a,b) is the greatest common divisor of a a and b b .

This is from IMO1995 shortlist and belongs to this set .


The answer is 37.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

Kazem Sepehrinia
Apr 5, 2015

Suppose p p and q q are natural numbers such that x = z p x=zp and y = z q y=zq . It follows that z p + z 2 q 2 + z 3 = z 3 p q p + z q 2 + z 2 = z 2 p q zp+z^2 q^2+z^3=z^3 pq \\ p+z q^2 +z^2=z^2 pq Therefore z p z|p and there is a natural number r r such that p = r z p=rz . Put this in the last equation: r z + z q 2 + z 2 = z 3 r q r + q 2 + z = z 2 r q q 2 q ( z 2 r ) + ( r + z ) = 0 rz+z q^2+z^2=z^3 rq \\ r+q^2+z=z^2 rq \\ q^2-q(z^2 r)+(r+z)=0 In order for getting integer values for q q , discriminant of the last equation (a quadratic in terms of q q ) Δ = ( z 2 r ) 2 4 ( r + z ) \Delta=(z^2 r)^2 -4(r+z) must be a perfect square, but Δ < ( z 2 r ) 2 \Delta<(z^2r)^2 Hence Δ ( z 2 r 1 ) 2 ( z 2 r ) 2 4 ( r + z ) ( z 2 r 1 ) 2 1 + 4 ( r + z ) 2 z 2 r 1 + 4 z r ( 2 z 2 4 ) 2 z 2 4 2 z 2 4 z 5 0 \Delta \le (z^2 r-1)^2 \\ (z^2 r)^2-4(r+z) \le (z^2r-1)^2 \\ 1+4(r+z) \ge 2z^2 r \\ 1+4z \ge r(2z^2-4) \ge 2z^2-4 \\ 2z^2-4z-5 \le 0 As a result z 2 z\le 2 . So z = 1 z=1 or z = 2 z=2 . If z = 1 z=1 then Δ = r 2 4 r 4 = ( r 2 ) 2 8 \Delta=r^2-4r-4=(r-2)^2-8 Since Δ \Delta is a perfect square, difference of two squares are 8 8 , hence squares are 1 1 and 9 9 . So r 2 = 3 r-2=3 and r = 5 r=5 . In this case, 5 + q 2 + 1 = 5 q q 2 5 q + 6 = 0 5+q^2+1=5q \\ q^2-5q+6=0 Thus q = 2 q=2 or q = 3 q=3 which leads to two solutions: ( 5 , 2 , 1 ) , ( 5 , 3 , 1 ) (5,2,1), (5,3,1) .

If z = 2 z=2 then Δ = ( 4 r ) 2 4 ( r + 2 ) \Delta=(4r)^2-4(r+2) must be a perfect square, but 4 r 2 ( r + 2 ) ( 2 r 1 ) 2 = 4 r 2 4 r + 1 4 r 1 r + 2 r 1 4r^2-(r+2) \le (2r-1)^2=4r^2-4r+1 \\ 4r-1 \le r+2 \\ r \le 1 Therefore r = 1 r=1 . Now we can write 1 + q 2 + 2 = 4 q q 2 4 q + 3 = 0 1+q^2+2=4q \\ q^2-4q+3=0 In conclusion q = 1 q=1 or q = 3 q=3 which gives us two solutions: ( 4 , 2 , 2 ) , ( 4 , 6 , 2 ) (4,2,2), (4,6,2) .

Really enjoyed the problem! +1!

User 123 - 6 years, 2 months ago

Log in to reply

Yep, It is a beautiful problem!

Kazem Sepehrinia - 6 years, 2 months ago

Log in to reply

the most beuatifil problem ever seen

Omar El Mokhtar - 6 years, 2 months ago

I agree too ! Keep up the flow of good questions from your side :)

kazem can you suggest me some nice books for undergraduate(math and physics).

Omar El Mokhtar - 6 years, 1 month ago

Don't wanna be a party crasher but this is a IMO 1995 suggested problem right?

Arian Tashakkor - 6 years, 1 month ago

Log in to reply

Thanks for pointing that out Arian, I saw this once in the third stage of Iran mathematical olympiads, I didn't know that it is an IMO suggested problems.

Kazem Sepehrinia - 6 years ago

Thanks for the solution.

Kazem Sepehrinia - 5 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...