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

Find a function F(x,y) whose level curves are solutions to the differential equation

  1. Jan 30, 2006 #1
    Hello everyone!! Its me! I'm stuck, these are suppose to be exact equations, and yet its not in exact equation forum. I thought being an exact equation you have to have the following form:

    http://tutorial.math.lamar.edu/AllBrowsers/3401/Exact_files/eq0008M.gif [Broken]

    But my question is the following:
    Find a function F(x,y) whose level curves are solutions to the differential equation
    http://cwcsrv11.cwc.psu.edu/webwork2_files/tmp/equations/a6/62aa7a4c8735852fb4ffa4df53d26d1.png [Broken]
    http://cwcsrv11.cwc.psu.edu/webwork2_files/tmp/equations/8d/8616ca1dd6ce230911485b98a57fa31.png [Broken]
    I rearranged it so its now:
    y - x dx/dy = 0
    but its not dy/dx, its dx/dy! so what happens? T hanks.
    Last edited by a moderator: May 2, 2017
  2. jcsd
  3. Jan 30, 2006 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    It means that you would be looking for some function

    You could just as well express it as

    [tex] x - y \frac {dy} {dx} = 0 [/tex]

    if it makes you more comfortable.
  4. Jan 30, 2006 #3
    ..test some simple solutions first

    For instance:

    Let M(x,y) = x^n A(y) and N = x^m B(y), where y = (sum over all i's) of:

    r0 + r1 x + r2 x^2 + r3 x^3 ....r(i)x^i for i less than I(sub)n.

    See if M=constant works..N=constant works...etc..

    Then try some transcendentals..exp(x), sin(x) etc..

    Non-linear equations want to you to 'play with them'..before they tell
    you what is going on.


    <---getting LaTex soon
  5. Jan 30, 2006 #4
    Thanks, so what ur saying is i just gota start guessing? how do uknow what to even guess by?
  6. Jan 30, 2006 #5
    simple solutions (part II)

    ...your general form:

    y = +/- sqr(x^2) + C0 when multiplied by a second solution ys = W(x)s

    where W = {(sum over all j) v0 + v1 x + v2x^3 ...v(j)x^j} times s(y),

    may give more solutions for your N= (whatever) ; M(whatever) non-polynomial factors. Some simple polynomials however lead directly into very involved transcendentals..for instance the D.E.: (second DE)

    y" = x^2 + x is a form of the elliptic logarithm.

    You might try the Steven Wolfram site on an "unsolvable".

  7. Jan 31, 2006 #6


    User Avatar
    Science Advisor

    Why would you rearrange it like that? You don't have to have any particular form in order for an equation to be exact. "exact" simply means that it can be put in the form M(x,y)dx+ N(x,y)dy = 0 and that there exist a function F(x,y) such that dF(x,y)= M(x,y)dx+ N(x,y)dy so that the equation is of the form dF(x,y)= 0 and has solution F(x,y)= C. One way to check for that is to check the equality of the mixed second derivatives of F: Fxy= My= Nx= Fyx.

    ydy- xdx= 0 certainly is exact: (y)x= 0= (-x)y. You could then say: Fy= y so F(x,y)= (1/2)y2+ f(x). From that Fx= f'(x)= -x so f(x)= -(1/2)x2. The solution is F(x,y)= (1/2)(y2- x2)= C

    Even more simply, the equation is obviously separable: ydy= xdx. Integrating both parts, (1/2)y2= (1/2)x2+ C which gives exactly the same answer.
    Last edited by a moderator: May 2, 2017
  8. Jan 31, 2006 #7

    should read:

    dy^2/dx^2 = y + y^2 (elliptic logarithm)

    ..and,...hmm they use these as basis spaces for Complex Proofs..

    Have you had numerical series solving? There are existance proofs
    using summation analysis, product functionals and Fourier.

    For instance the 2D Laplace form for F(x,y):

    L(f,s) = int f(x,y) exp(-sy) dy int exp(-vx)dx | from 0 to ymax:xmax

    ...enough already*.

    I am getting LaTex before I come back here.


    *(..cancelling my warp drive journal..till then as well..email me for info)
  9. Jan 31, 2006 #8
    Thanks everyone, and thank you IVEY, i don't know why I thought it had to be in that form. That was alot easier then I expected! and yes the answer was:
    http://cwcsrv11.cwc.psu.edu/webwork2_files/tmp/equations/f5/7102c63fed37197e439d0270e34b001.png [Broken]
    Last edited by a moderator: May 2, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook