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!

Looking for a function with specific properties

  1. Aug 29, 2013 #1
    Hi everyone,
    I'm trying to find a function of single variable f(x) with the following properties:

    -It is symmetric around zero
    -It is differentiable everywhere
    -f'(x)≥0 for all x>0
    -f'(x)=0 when x=0
    -f'(x)≤0 for all x<0
    (I think these last two actually follow from the first three?)
    -It has an upper bound of 1
    -It has a lower bound between 0 and 1, which I can set using a parameter c

    For example, the function
    f(x)=(c/10)*abs(x)+1-c when abs(x)≤10
    f(x)=1 when abs(x)>10

    fulfills all the conditions, except that it is not differentiable at x=-10, x=0 and x=10.

    But I'm hoping to find a relatively simple function that would fulfill all those requirements.

    If anyone has any ideas, I'd be very thankful!
  2. jcsd
  3. Aug 29, 2013 #2
    How about [itex] f(x)=1-ce^{-x^2} [/itex] with [itex] 0 \leq c \leq 1 [/itex] and a minimum of [itex] 1-c [/itex]?

    Edit: And yes, the last two do follow from the first three. If [itex] f(-x)=f(x) [/itex], then [tex] f'(-x)=\lim_{h \to 0} \frac{f(-x+h)-f(-x)}{h}=\lim_{h \to 0} \frac{f(x-h)-f(x)}{h}= -\lim_{h \to 0} \frac{f(x)-f(x-h)}{h}=-f'(x) [/tex]

    Or, if you prefer, [itex] f'(x)=\frac{d}{dx}f(x)= \frac{d}{dx}f(-x)=-f'(-x) [/itex]
    Last edited: Aug 29, 2013
  4. Aug 29, 2013 #3
    Thanks HS! Impressively quick response.
    That looks very promising actually. I might modify it with an extra parameter:

    Then, by changing the value of b I can change the rate at which it approaches the upper bound.

    Anyway, looks very good and helpful!

    PS Sorry about using text for the maths. Still trying to get the hang of this...
  5. Aug 29, 2013 #4
    [itex]f(x)=c[/itex] [itex](0 \leq c \leq 1)[/itex] surely fulfils all your requirements - and that's a pretty simple function!
  6. Aug 29, 2013 #5
    That's true oay, thanks!

    Maybe instead of
    "-It has an upper bound of 1"
    I should have written
    -it's limit at +-Infinity = 1

    Not sure if even that makes it watertight. But anyway, HS-Scientist's suggestion is more suitable for what I need ;)
  7. Aug 29, 2013 #6
    I think when you say "upper bound", you mean to say "least upper bound" or "supremum". Similarly, "lower bound" should be "greatest lower bound" or "infimum".
  8. Aug 29, 2013 #7
    Yes, that's sounds like what I was thinking of, and what I should have written!
    It's been a while since I had to use this terminology...
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook