Not Quite an Ellipse

y 2 = x 3 3 x + 7 y^2 = x^3 - 3x +7

Let x x and y y be integers satisfying the elliptic curve above.

Find the sum of all distinct values of x x .


The answer is 63.

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.

1 solution

Kushal Bose
Feb 16, 2017

I have done this by Python Programming

import math for i in range(1,100): c=math.pow(i,3)-3*i+7 print(math.sqrt(c),i)

I got x = 1 , 1 , 2 , 3 , 59 x=-1,1,2,3,59

I am interested to know is there any other mathematical way to solve this

The equation can be written as ( y 3 ) ( y + 3 ) = ( x + 1 ) 2 ( x 2 ) (y-3)(y+3)=(x+1)^2(x-2) which easily yields solutions ( 1 , 3 ) ( 1 , 3 ) ( 2 , 3 ) ( 2 , 3 ) (-1,3)(-1,-3)(2,3)(2,-3) . When you go to the linked wiki page you will find information about colinear 3 points on a curve. Connecting points (-1,-3) and (2,3) yields point (3,5); connecting (2,-3) and (3,5) yields (59,453).

Maria Kozlowska - 4 years, 3 months ago

Log in to reply

Precisely. ;)

Worranat Pakornrat - 4 years, 3 months ago

OK, we can use the group structure of the elliptic curve to obtain these extra solutions. How do we know there are no more?

Mark Hennings - 4 years, 3 months ago

Log in to reply

Using Kushal's brute force method we can find there are in fact many more: " for i in range(1000000,1010000): c=math.pow(i,3)-3*i+7 if int(math.sqrt(c))==math.sqrt(c): print (math.sqrt(c),i) " yields 1000002, 1001999, 1002003, 1004002 and five other solutions

Hopefully I am wrong about this. Please let me know how if someone figures it out.

Abel McElroy - 4 years, 3 months ago

Log in to reply

@Abel McElroy You have rounding error issues. If x = 1000002 x = 1000002 then x 3 3 x + 7 = 1.0000030000000000045 × 1 0 9 = 1000003000.0000000045 \sqrt{x^3 - 3x + 7} \,=\, 1.0000030000000000045 \times 10^9 \,=\, 1000003000.0000000045 which is nearly, but not exactly, an integer!

Mark Hennings - 4 years, 3 months ago

Log in to reply

@Mark Hennings nice! thank you. Mark has debunked my brute force counter examples. Rounding errors were the root of the problem. Thanks Mark

Abel McElroy - 4 years, 3 months ago

Log in to reply

@Abel McElroy I have tested for 1 x 2000000 1 \le x \le 2000000 , so "yes".

Mark Hennings - 4 years, 3 months ago

The Wikipedia article mentions Siegel's theorem about a finite number of integral points on an elliptic curve. I think that all integral points can be found in the manner of finding a third point on a line passing through two already known points. There are no more integral points on this curve that could be found in this manner. I do not have a proper proof and that is why I posted my incomplete solution as a comment.

Maria Kozlowska - 4 years, 2 months ago

Log in to reply

@Maria Kozlowska There are, of course, only a finite number of integral points. I am less sure that we cannot be sure that there are not some other integral points, which are obtained from other, rational, points by further applications of the group property. Just because we cannot combine the integral points we know about directly to form other integral points does not mean that there aren't others.

I hunted the net quite extensively, and have not been able to find a definitive statement about this elliptic curve...

Mark Hennings - 4 years, 2 months ago

Log in to reply

@Mark Hennings I was hoping that all the integral points form a chain of solutions. Now I do not think it is exactly the case.

Maria Kozlowska - 4 years, 2 months ago

Does 1 satisfy? Because if x=1 then y 2 = 7 y^2 = 7 .

Worranat Pakornrat - 4 years, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...