1. Limited time only! Sign up for a free 30min personal 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!

Homework Help: Change in variables

  1. Jul 26, 2012 #1
    1. The problem statement, all variables and given/known data
    I am performing a change of variables, s --> t and am wondering can I just write g(s) --> g(t) or do I have to alter the function e.g. g(s) --> a*g(t).

    2. Relevant equations
    s = (a/b)* t
    g(s) is to be found numerically and therefore we do not have it's definition.

    So Option one:
    just rewrite g(s) as g(t)

    Option two:
    g(s) = (a/b)*g(t)

  2. jcsd
  3. Jul 26, 2012 #2
    Um, generally, if s -> t, then g(s) does NOT go to g(t). If you know that g is continuous, then you may assume that. Otherwise, you will have to work off the official definition of limits.
  4. Jul 26, 2012 #3
    I would assume that if s => ta /b, g(s) => g(ta/b) if a,b are arbitrary / real constants. I could be wrong though.
  5. Jul 26, 2012 #4
    That's definitely not true. Consider the function

    g(x) = x + 1 for x > 0 and x - 1 for x <= 0. Then as s -> 0 from the right, g(x) => 1, even though g(0) = -1.
  6. Jul 26, 2012 #5


    User Avatar
    Science Advisor

    You are misunderstanding the notation. The question is not about limits. The OP is asking 'If we change variable s to variable t, can we write f(s) as simply f(t)?'

    MathMonster, there are times when that is acceptable but I would avoid it. Technically if, for example, [itex]f(s)= s^2- 3s+ 2[/itex] then [itex]f(t)= t^2- 3t+ 2[/itex] but I don't believe that is what you are trying to do. If [itex]f(s)= s^2- 3s+ 2[/tex] then replacing s with t= s-2 gives, since s= t+2, [itex]f(s)= f(t+ 2)= (t+ 2)^2- 3(t+ 2)+ 2= t^2+ 4t+ 4- 3t- 6+ 2= t^2+ t[/tex] but notice that is f(t+2), not f(t). If you like you say explicitely that g(t)= f(t+ 2) and then use g(t).
  7. Jul 26, 2012 #6
    I see - my apologies. Nevertheless, you should not be moving constants out of expressions, g(at/b) =/= (a/b) * g(t). So unless you would like to simply g(t) with some change of variable (again, this as the above advised, this is not suggested - there are many scenarios where this is invalid - an introductory book on Real Analysis can help clarify the conditions under which you may do so), there isn't too much of a point of doing so.
  8. Jul 30, 2012 #7

    Thank you for the reply! Yes I am talking about a change in variables not limits - sorry that's my fault I didn't think of the notation I was using.

    HallsofIvy - thank you! I thought that was the case but was confused how to so it. Basically i've got a function g(s+) and g(s-) and a relation linking s+ & s-. I then have a formula with the two 'g's' in which I need to write as s+ in order to solve it.

    So basically, I will just write g(s-) = g((a/b)*s+) and leave it at that? There's a bit of differentiation that will come in which will take into account the change.

  9. Jul 30, 2012 #8
    Could you possibly post the question you are working on?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook