Ask Your Question

solve differential equation

asked 2014-10-14 17:10:24 -0500

Martin Maxa gravatar image

updated 2014-10-17 10:30:38 -0500

I need to solve this third-order linear partial differential equation:

d^2/dx^2 d/dy f(x,y) = f(x,y) - x*y Could you please help me to do this?

Thank you very much for your advise!

edit retag flag offensive close merge delete


Is this something that even has a symbolic solution? There is some numeric stuff that should be usable with PDEs, but I'm not as familiar with that.

kcrisman gravatar imagekcrisman ( 2014-10-16 08:56:09 -0500 )edit

True. It really depends on what you mean by solve... it can be a symbolic solution, an explicit (convergent) Taylor expansion at the origin, a numerical approximation, etc.

vdelecroix gravatar imagevdelecroix ( 2014-10-17 10:31:34 -0500 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2015-05-08 09:05:55 -0500

GM3D gravatar image

Not sure if you still need an answer, but here goes;

Define g(x) = f(x) - x*y, this makes the given equation homogeneous;

d^2/dx^2 d/dy g(x, y) = g(x, y)

And now assume g(x, y) = A(x)B(y). That lets you factorize the equation into two parts;

d^2/xx^2 A(x) = a A(x), d/dy B(y) = (1/a) B(y), with some positive constant a.

This can be solved as A(x) = C1 exp(+-sqrt(a) x), B(y) = C2 exp(y / a)

And by multiplying them back,

g(x, y) = (C1 * C2) exp(+-sqrt(a) x + y / a) = C exp(+-sqrt(a) x + y / a)

Therefore, f(x, y) = C exp(+-sqrt(a) x + y / a) + x*y.

General solution is linear combination of these solution with different values of C and a.

But I might have missed something there, because this solution contains only two arbitrary parameters whereas the original equation is a third order differential equation.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower


Asked: 2014-10-14 17:10:24 -0500

Seen: 524 times

Last updated: May 08 '15