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!

Continuous functions

  1. Aug 12, 2006 #1
    This is a simple question: given a continuous function, f(x), in a closed interval, how do we show that there is value "a" small enough such as for arbitrary x:

    f(x+a) - f(x) < e

    Where e lower bound is 0.


    Thxs in advance.
  2. jcsd
  3. Aug 12, 2006 #2


    User Avatar
    Science Advisor

    If [itex]x_0[/itex] is in the closed interval then the definition of "continuous" is that [itex]\lim_{x\rightarrow x_0} f(x)= f(x_0)[/itex]. That in turn means "given [itex]\epsilon> 0[/itex] there exist [itex]\delta> 0[/itex] such that if [itex]|x-x_0|< \delta[/itex] then [itex]|f(x)-f(x_0)|< \epsilon[/itex].
    To get your result, let [itex]x_0= x[/itex] and [itex]x= x_0+ \delta[/itex]. Then your "a" is "[itex]\delta[/itex]", "e" is "[itex]\epsilon[/itex] and [itex]|f(x)- f(x_0)|<\epsilon[/itex] becomes |f(x+a)-f(x)|< e. Of course, it's still true without the absolute value since if f(x+a)- f(x) is negative it's still less than the positive e.
  4. Aug 12, 2006 #3


    User Avatar
    Homework Helper

    If the function is uniformly continuous, the proof is easy. This would mean that for any e>0, there is a d>0 such that |x-y|<d implies |f(x)-f(y)|<e for all x,y in the domain. So just pick your a as d/2.

    It's a little more complicated to show this from just the assumption of continuity (which, remember, is a local property, where as you're looking for an a with a ceratin global property), but the http://en.wikipedia.org/wiki/Heine-Cantor_theorem" [Broken] states that these definitions are equivalent if the domain is compact, which a closed interval is. If you can't just cite this theorem, there is a link to a proof on the wikipedia page.
    Last edited by a moderator: May 2, 2017
  5. Aug 12, 2006 #4


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    Mmh, yes, the existence of the 'a' for all x is assured by definition of uniform continuity + Heine-Cantor in the case [itex](X,d(x,y))=(\mathbb{R},|x-y |)[/itex].

    P.S. Halls, do you know how to write in italic w/o using Latex?
    Last edited: Aug 12, 2006
  6. Aug 12, 2006 #5


    User Avatar
    Science Advisor
    Homework Helper

    just subdivide your interval into unit length subintervals, then subintervals of length 1/10, then 1/100, and so on....

    assuming no subdivision ever has intervals on all of which the variation is less than epsilon, resulots in the existence of an infinite decimal, i.e. a real number, in our interval where the function is not continuous.
  7. Aug 16, 2006 #6
    My question has been misunderstood... sorry the lack of clarity. What I meant is:

    Let e and c be constants, for all x. Is there such a constant for which, regardless of x;

    If within the interval [a,b]

    (</= stands for inferior or equal)

    f(x+e) </= a


    f(x+e) - f(x) </= c


    f(x+l) - f(x) ; if l < k < e


    f(x+l) - f(x) < f(x+k) - f(k) < c


    If there is, what is the proof?

    This all has to do with a proof that all Riemann sums converge towards the same value when the span of its subdivision tends to 0. Thanks in advance.
    Last edited: Aug 16, 2006
  8. Aug 17, 2006 #7


    User Avatar
    Homework Helper

    I can't make any sense out of your last post, werg. If what you're trying to do is prove all Riemann sums converge to the same limit for continuous functions on a closed interval, you would want to start with the question we thought you asked and then answered.
    Last edited: Aug 17, 2006
  9. Aug 17, 2006 #8
    Okay, so after a little thought I solved it. Sorry for the lack of clarity again, I tried my best to explain. N/P now :smile:
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook