1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Does Homogeneous = Separable?

  1. Sep 12, 2004 #1

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Hi,

    Please bear with me, I've only had the first sort of "pseudo-lecture" in ordinary d.e.'s this past week, and I was doing some reading ahead. It occurred to me that if linear first-order differential equations are those that can be written in the general form:

    [tex] \frac{dy}{dx} + P(x)y = Q(x) [/tex]

    and if I understood my prof's remark correctly that homogeneous linear first-order d.e.'s are those for which [itex] Q(x) = 0 [/itex], then all homogeneous first-order linear differential equations are actually separable because:

    [tex] \frac{dy}{dx} + P(x)y = 0 [/tex]

    [tex] \frac{dy}{dx} = -P(x)y [/tex]

    Which can be solved as follows:

    [tex] \int{\frac{dy}{y}} = -\int{P(x)dx} [/tex]

    ^^Looks separable to me. I just wanted to make sure I was getting the basics before I moved on. So is the thread title statement true then? Oh yeah, and I'm guessing that the converse is not always true right? A separable first-order d.e. is not necessarily a homogeneous linear d.e. right? After all, [itex] \frac{dy}{dx} = \frac{6x^2}{2y + cosy} [/itex] is not linear in [itex] y [/itex] right? But it looks like it could be homogeneous non-linear (if such a thing exists, I don't know???). In that case, would the statement homogeneous = separable be true in the most general sense? Moving on with the original example (I'm just solving while "LaTeXing" to see what comes out) ...

    [tex] \ln|y| = -\int{P(x)dx} [/tex]

    [tex] |y| = e^{-\int {P(x)dx}} [/tex]

    Now, the most general solution for [itex] y [/itex] must include the most general antiderivative, so we'll have a [itex] C [/itex] stuck in there if and when we solve the integral:

    [tex] y = \pm e^{-\int {P(x)dx} + C} = \pm e^{-\int {P(x)dx}}e^C = \pm Ae^{-\int {P(x)dx}} [/tex]

    Whoah, cool! So in a homogeneous first order d.e. the solution takes the form of some constant [itex] \pm A [/itex] times the reciprocal of the integrating factor [itex] I(x) [/itex]?! Is this always true?

    EDIT: Yeah actually I can see why that is. Take the usual statement about the integrating factor:

    [tex] (I(x)y)^{\prime} = I(x)Q(x) [/tex]

    If [itex] Q(x) = 0 [/itex], then

    [tex] (I(x)y)^{\prime} = 0 [/tex]

    [tex] I(x)y = \pm A [/tex]

    [tex] y = \frac{\pm A}{I(x)} [/tex]

    Awesome! I discovered a lot more writing this post than I expected. I hope someone will correct me if I've made any errors.
     
  2. jcsd
  3. Sep 12, 2004 #2
    Just wondering, but do you really need to state that the constant of integration is +/- A, instead of just A (in the part where you talk about integrating factors)?
     
  4. Sep 12, 2004 #3

    cepheid

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Good point. In fact, I guess I could have gotten rid of the +/- way up here:

    [tex] \pm (e^{-\int {P(x)dx}})(e^C) = Ae^{-\int {P(x)dx}} [/tex]

    just combining everything into "A", which would remain as A for the ensuing discussion.
     
    Last edited: Sep 12, 2004
  5. Sep 12, 2004 #4

    HallsofIvy

    User Avatar
    Staff Emeritus
    Science Advisor

    In answer to the original question, yes, a first order d.e. that is "homogeneous" in this sense: P(x) dy/dx+ Q(x)y= 0 is trivially seperable: dy/y= -(Q(x)/P(x))dx.

    But be careful: in the limited area of FIRST ORDER d. e.s, the term "homogeneous" is often used in quite a different way (the d.e. A(x,y)dx+ B(x,y)dy= 0 is "homogeneous" if and only if B(&lamda;x,&lamda;y)/A(λx, λy)= B(x,y)/A(x,y) (essentially that means that the total exponent of x and y in each term is the same).
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?