Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

PDE with complex argument

  1. Sep 11, 2012 #1
    I'm in truble with a partial differential equation. Actually it is a system of PDE but It would be useful to solve at least one of them.
    The most easy one is this one

    2 \bar{\xi}\left(\bar{s},\bar{t},\bar{u}\right) - 2 \xi\left(s,t,u\right) + \left(s-\bar{s}\right)\left(\bar{\partial}_\bar{s} \bar{\xi} + \partial_s \xi \right) = 0

    This equation can be simplified to

    2 A^*\left(z^*\right) - 2 A\left(z\right) + \left(z-z^*\right)\left(\bar{\partial}_{z^*}A^*+\partial_{z}A\right)= 0

    I further developed my computation using [tex] A(z) = u(x,y) + i v(x,y) [/tex] with [tex] u,v \in \mathbb{R}[/tex]
    finding (I used Cauchy-Riemann equations)
    [tex] v(x,y) = y^2 f(x+y) [/tex]
    Here is where I get stucked since I cannot find a suitable form of "f(x+y)" in order to obtain "u" and satisfy Cauchy-Riemann equations...
    Any ideas?
  2. jcsd
  3. Sep 11, 2012 #2
    Consider the expression:


    I assume that means:

    [tex]\overline{\frac{\partial \bar{f}}{\partial\bar{s}}}[/tex]

    but we know that:

    [tex]\frac{\partial \overline{f}}{\partial \overline{s}}=\overline{\frac{\partial f}{\partial s}}[/tex]

    which means you have:

    [tex]2\overline{A}(\overline{z})-2A(z)+2(z-\overline{z})\frac{\partial A}{\partial z}=0[/tex]
  4. Sep 11, 2012 #3
    I'm afraid I use the wrong notation or maybe I didn't understand at all! =)
    [tex] \bar{\partial}_{\bar{s}} \xi^* [/tex]
    I mean the derivate of xi* wrt the complex conjugate of s (i.e. \bar{s}). I use the bar over the partial derivative to point out that the derivate is made over \bar{s} and not s. Sorry about this misleading notation! :)
  5. Sep 12, 2012 #4
    Ok, that's confussing. Tell you what, how about we just do it my way:

    [tex]2\overline{A}(\overline{z})-2A(z)+2(z-\overline{z})\frac{d A}{d z}=0[/tex]

    Can we even solve that one? The conjugate variables really hit me with a surprise though and I'm not use to working with DEs like that. I mean what do you do with something like that? Is it even well-posed? Suppose nobody could help us and we had to do something with it, a thesis or something? What do we do? Suppose we could first look at:

    [tex] \frac{dy}{dz}+\overline{y}(\overline{z})=0[/tex]

    Can we even do that one? Does it even make sense? Looks like another whole-semester type problem to me.
    Last edited: Sep 12, 2012
  6. Sep 12, 2012 #5
    I have to admit I'm confused too...
    My problem, I mean in its original formulation, require to find the holomorphic Killing vector of a given Kahler manifold. In order to do that I found I have to solve that equation (and many more to be honest...).
    Now I wondering if by [tex] \bar{A}(\bar{z})[/tex]
    they actually mean [tex] \left(A(z)\right)^*[/tex]
    In that case I can set [tex] A = u(x,y)+iv(x,y) \qquad \bar{A} = u(x,y)-iv(x,y)[/tex]
    For which I found this solution
    u(x,y) = \frac{1}{2} C_1 \left(x^2-y^2\right)+C_2 x + C_3 \qquad
    v(x,y) = C_1 xy + C_2 y
    Which is a bit tempting since it satisfy also Cauchy Riemann equations..
  7. Sep 12, 2012 #6
    I think that means the conjugate of A at the conjugate of z. So if:




    Not sure though ok?
  8. Sep 20, 2012 #7
    Bulletin from the front. :)

    As I supposed they intended just the conjugation of the entire function not of both function and variables... So I solved, thank you anyway!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook