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

A trick on PDE?

  1. Mar 19, 2007 #1
    A trick on PDE??

    Hi all.
    I am reading a text in mathematical wave theory.
    I saw and am confused by a manipulation of a PDE, as shown in the attached figure.

    I don't really undertand how the equation (1.9) is transformed by "introducing the charcteristic variables). (as indicated by the red arrow and the question mark)

    I guess there are some missing steps? Could somone fill in the missing links so that I could know what's going on there.

    Attached Files:

    • de.jpg
      File size:
      36.4 KB
  2. jcsd
  3. Mar 19, 2007 #2


    User Avatar
    Staff Emeritus
    Science Advisor

    New parameters are introduced such that [tex]\xi=x-t ,\hspace{1cm} \zeta=x+t[/tex]. You need to work out utt and uxx in terms of the new parameters. For example, [tex]u_t=\frac{\partial u}{\partial \xi}\frac{\partial \xi}{\partial t}+\frac{\partial u}{\partial \zeta}\frac{\partial \zeta}{\partial t}=-u_{\xi}+u_{\zeta}[/tex], using the chain rule. Use a similar method to find ux, and then to find utt and uxx.

    When you've worked out utt and uxx sub them into the LHS of (1.9) and you should obtain the result.
    Last edited: Mar 19, 2007
  4. Mar 19, 2007 #3
    Thanks! I think I could work it out.
    But what is the motivation of doing this?
    I mean, is there any phyical meaning attached? Or What benefits can be obtained after this transformation?
  5. Mar 19, 2007 #4


    User Avatar
    Staff Emeritus
    Science Advisor

    I'm not sure of the physical meaning, but the advantage of performing the transformation is that you can integrate the equation! Note that f is now a function of only [itex]\xi[/itex], so the primes denote differentiation wrt [itex]\xi[/itex]. Thus you can integrate once wrt [itex]\xi[/itex] and once wrt [itex]\zeta[/itex] and obtain the result given (noting that this is a PDE and so the constant of integration is not just a constant in the normal sense, but is a function of the coordinate not integrated with repect to.)

    IF you don't make the substitution, then the equation is a fair bit harder to solve.
  6. Mar 19, 2007 #5
    many thanks! Nice and detailed explanation!

    BTW, do you mind answering also my another question on asymptotic expansion?
  7. Mar 20, 2007 #6
    I still find it difficult to implement the two initial conditions to determine functions A and B...
    how to relate the conditions in terms of the characteristic variables?
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: A trick on PDE?
  1. Solutions to this PDE? (Replies: 4)

  2. Tough pde (Replies: 21)

  3. Linearity of PDE (Replies: 4)