Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If we have a system of PDE's, where each PDE is different i.e. for

$u:U\subset \Bbb R^2\to\Bbb R^3$, $u(x,y)=(a(x,y),b(x,y),c(x,y))$, which needs to satisfy

$ \left\{ \begin{array}{ll} u_x\cdot u_y=0\\ |u_x|^2-|u_y|^2=0 \\ H(u)=c\end{array} \right.$

Where $H(u)$ is the mean curvature of $u$, and $c$ is a constant.

The reason I ask, is that it seems different to a system $u_t=u_{xx}$ (for example) where $u$ is a vector field. Here the PDE is the same in each case.

Is there a specific theoretical approach to take?

Thanks for any help.

share|cite|improve this question
Since it's a first order system, you could try the characteristics method. I won't lie, I never tried myself, and I'm unsure of how it's going to play out with the second set of equation. – jgyou May 27 '14 at 19:14
@JGab, due to the mean curvature term, it is a second order system, but would you have any ideas on this? – ellya May 27 '14 at 19:14
up vote 1 down vote accepted

it seems different to a system $u_t=u_{xx}$ (for example)

It is different. The linear heat equation on $\mathbb R^n$ is one of the PDE for which we can write down a general solution. There are other such PDE... but you can count them on the fingers of one hand.

Are there standard approaches, to solving a system of nonlinear PDE?

Yes. The standard approach is: you don't solve a system of nonlinear PDE. You decide on the function space in which it should be posed. You consider the appropriateness of initial/boundary conditions and forcing term, if applicable. You carefully approach the problems of existence and uniqueness. And of regularity of solutions. Their asymptotic behavior. Comparison principles (like the maximum principle). And so on. But you don't solve them. Because you can't.

The system you wrote down says that $u$ is a conformal CMC immersion. So, the study of that PDE is the study of CMC surfaces in $\mathbb R^3$, a subject with which books have been filled. The special form of the first two equations (conformality) opens a connection with complex analysis, from which the subject benefits (holomorphic Hopf differential; Kenmotsu's representation formula). So, this system is a lot better understood that what you'd get by, say, changing the first equation to $u_x\cdot u_y = 1$ and keeping the rest the same.

Suggestion: if you are interested in the PDE perspective on constant mean curvature surfaces, begin by reading the lecture notes Constant Mean Curvature Surfaces: Harmonic Maps and Integrable Systems by Frederic Hélein. They are more readable than most books concerning nonlinear PDE. There is also a short book Surfaces with Constant Mean Curvature by Kenmotsu, but it's not really organized as a book from which one could learn the subject.

share|cite|improve this answer
thank you! I was also wondering what approach should be taken if one were to numerically approximate solutions? – ellya May 28 '14 at 5:04
@ellya Some references are given here. – ᴡᴏʀᴅs May 28 '14 at 5:11
thank you for your help, but Which of these references are to do with numerical approximations? – ellya May 28 '14 at 5:22
I couldn't see which ones were numerically oriented, there are about 12 on the Wikipedia page you linked – ellya May 28 '14 at 5:38
As a last supplementary question, in the scenario that I only had the mean curvature equation, is it possible to make approximations of solutions, even though there would be three unknowns and only one equation? – ellya May 28 '14 at 7:14

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.