# Solving PDE

1. Dec 13, 2004

### ksoy

[SOLVED] Solving PDE

I am just wondering, is there any gerneral method in solving PDE's or just by guess works??

thanks...

2. Dec 13, 2004

### dextercioby

In mathematics there's no room for "guessing".
As for PDE-s,well,mathematicians found a way to classificate them.So far (and more than surely in the future) there has't been found a general method to solving PDE-s,that is to aapply successfully for every kind of PDE.
For example,for nonlinear PDE-s,there is no general method of solving.Analitically,of course.I assume that was the initial question about.
Try to solve (or imagine a way to tackling) somthing like that
$$\frac{u^{3}(x,y,z)}{xy^{\frac{6}{3}}z}[\frac{\partial^{5} u(x,y,z)}{\partial x^{5}}]^{7}+5 u^{8}(x,y,z)-12x^{7}y^{\frac{3}{4}}z=0$$

3. Dec 13, 2004

### ksoy

how about linear PDE??
Is there a general method of solving them??

4. Dec 13, 2004

### dextercioby

Yes,for the linear case,there is.Try first of all to bring them to the canonical form.From there analyze the type (hyperboli,elliptic,parabolic) for every point in the domain of the unknown function.Then look very carefully at the geberal problem and its conditions (boundary type (Dirichlet/Neumann),or initial). Several methods come up then.Green function methods,variable separation methods,Fourier/Laplace transform methods,and so on.
Some equation,after being put in the canonical form may admit immediate integration,and the famous example is the unidimensional wave equation.

Anyway,all these depend from case to case.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?