(In reply to
Simplifying matters (Spoiler) by Steve Herman)
ok first I will show how to compute the square root of a complex number.
sqrt(x+yi)=a+bi
x+yi=(a^2-b^2)+2abi
a^2-b^2=x
2ab=y
b=y/(2a)
b^2=y^2/(4a^2)
a^2-(y^2/(4a^2))=x
4a^4-y^2=4xa^2
4a^4-4xa^2-y^2=0
let t=a^2
4t^2-4xt-y^2=0
t=(4x+-sqrt(16x^2+16y^2))/8
t=(4x+-4sqrt(x^2+y^2))/8
t=(x+-sqrt(x^2+y^2))/2
we want positive t to give real x so we have
t=(x+sqrt(x^2+y^2))/2
this gives
a=sqrt((x+sqrt(x^2+y^2))/2)
and this gives
b=y/(2a)
so we have already determined the solutions for n>=0 and n<=625/4
if n<0 then the first square root is real and the second is complex and thus we end up with a complex number overall and thus never an integer.
if on the other hand n>625/4 then sqrt(625/4 - n) is imaginary which then results in the sum of the square roots of 2 complex numbers. For the first one we have x=25/2 and y=sqrt(n-(625/4)) and the second has x=25/2 and y=-sqrt(n-(625/4))
doing the substitutions and simplifying we end up with the sum being
(25+2sqrt(n))/2
now this can only be an integer when n is a perfect square. But then 2*sqrt(n) is always even and thus 25+2sqrt(n) is always odd and thus finally (25+2sqrt(n))/2 can never be an integer.
Now it appears that this proves that n=144 is the only solution but having said that I can not find where the flaw is in Steve's solution and neither can I find a flaw in my own. So hopefully somebody can find one in either or maybe even both and set the record straight here :-)
|
Posted by Daniel
on 2009-03-24 14:36:47 |