There is a basic identity by Lander,
(a+b+c)^5 = (-a+b+c)^5 + (a-b+c)^5 + (a+b-c)^5 + 80abc(a^2+b^2+c^2)
Problem: Find {a,b,c} such that,
80abc(a^2+b^2+c^2) = d^5+e^5 (eq.1)
I found this has an infinite number of non-trivial solns given by,
{a,b,c} = {u^5, 25v^5, 50v^5}, and {d,e} = {10u^3v^2, 50uv^4}.
for arbitrary {u,v}. Questions:
1) Is there any other parametrizations to eq.1?
2) What other rational {a,b,c;d,e}, excluding {1,25,32; 20,40}, that
do not belong to this family? (Enough solns may lead to a pattern.)
- Titus
Dear Titus,
The equality
(a+b+c)^5 - (-a+b+c)^5 - (a-b+c)^5 - (a+b-c)^5 = 80abc(a^2+b^2+c^2)
is just a particular case of part of form^(2n+1).t^(2p+1).u^(2q+1)
of a polynomial g(s,t,u)=
(g(s,t,u)-g(-s,t,u)-g(s,-t,u)-g(s,t,-u)+g(-s,-t,u)+g(s,-t,-u)+
g(-s,t,-u) - g(-s,-t,-u))/8
On (s+t+u)^3
we've got:
(s+t+u)^3-(-s+t+u)^3-(s-t+u)^3-(s+t-u)^3 = 24stu
Alain
I. Yes, the basic form is called Boutin's Identity which generalizes
the difference of two squares into a sum and difference of 2^(k-1) kth
powers. Thus,
(a+b)^2 - (a-b)^2 = 2ab
(a+b+c)^3 - (a-b+c)^3 - (a+b-c)^3 + (a-b-c)^3 = 24abc
and so on for all kth powers. See Boutin's Identity
http://sites.google.com/site/tpiezas/001.
II. I found other solns to 80abc(a^2+b^2+c^2) = d^5+e^5. The list
{a,b,c}; {d,e} so far is:
{1, 25, 32}; {20,40},
{2, 352, 355}; {-328, 388}
{4, 125, 155}; {-70, 190}
{32, 101, 205}; {-280, 340}
and, other than the known family of solns, there doesn't seem to be
rhyme or reason to these.
- Titus
Hi Titus ,
two more solutions :
{a,b,c; d,e} {1,2,125 ; 10, 50}
{a,b,c; d,e} {3,4,240 ; 80,100}
As for another equation for :
80abc(a^2+b^2+c^2) = d^5+e^5 (eq.1)
i found that :
C(a,b,d,e) = c
with:
P(a,b,d,e)=(9*a^2*b^2*(d^5+e^5)+sqrt(3)*sqrt(a^4*b^4*(25600*a^2*b^2*
(a^2+b^2)^3+27*(d^5+e^5)^2)))^(1/3);
C(a,b,d,e)=(5^(1/3)*P(a,b,d,e)-40*6^(1/3)*a^2*b^2*(a^2+ b^2)/P
(a,b,d,e))/(2*30^(2/3)*a*b);
(i don't know how to simplify these)
Regards
Gerry
Thanks for the solns, Gerry. Yes, it can be solved as a cubic in
either {a,b,c}. Too bad there's no simple criterion involving the
cubic's discriminant D to determine if it has a rational root. (Unlike
the quadratic where D simply is made a square.)
P.S. I wonder if eq.1, with constraints, can be transformed into an
elliptic curve. After all, it has an infinite number of solns.
- Titus