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

Equation fitting

  1. Apr 10, 2006 #1


    User Avatar
    Science Advisor
    Homework Helper

    I'm almost too embarrassed to post, but I thought someone might have insight that could help me here. I'm trying to fit an equation to data, but I'm just not sure what sort of equation I should use. I have some requirements on the form of the equation, and then I have points (yet to be determined, actually).

    [tex]f'(x)>0\,\,\forall x>0[/tex] (cost is increasing)
    [tex]f''(x)<0\,\,\forall x>0[/tex] (essentially, average cost is decreasing)
    [tex]\lim_{x\rightarrow\infty} f(x)=k[/tex] (price is bounded above in this fashion)

    It would make the most sense if the origin was included, but this can be one of the data points.

    So before I even think about least-squares vs. minimal points on the curve, I wanted to consider different forms that could make sense. The first that comes to mind was the elementary


    but this has some instabilities and oddities. I guess I'm just looking for thoughts on widely-used models that fit my criteria or could be modified to fit them. In the perfect case I'd have a smooth equation that was nearly linear for small x and essentially hyperbolic for large x.

    Any suggestions would be welcomed.
  2. jcsd
  3. Apr 11, 2006 #2


    User Avatar

    One equation that fits your requirements is:
    [tex]f(x) = k(1 - e^{-ax}) [/tex] k > 0 a > 0
    [tex]f´(x) = ak e^{-ax} > 0[/tex]
    [tex]f"(x) = -a^2k e^{-ax} < 0[/tex]
  4. Apr 11, 2006 #3


    User Avatar
    Science Advisor
    Homework Helper

    Thanks. You know, I was just so out of it I didn't think to use this. :blushing:

    I'm going to play with this for a while and make it work. You gave me just what I needed to get my brain working again.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook