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: Proving Cauchy Sequences

  1. May 16, 2010 #1
    The problem statement, all variables and given/known data

    Well, my problem is proving that sequences are in fact Cauchy sequences. I know all the conditions that need to be satisfied yet I cannot seem to apply it to questions. (Well, only the easy ones!)

    My question is, prove that [tex]X_{n}[/tex] is a Cauchy sequence, given that [tex]X_{n+1}= f(X_{n})[/tex] where f(x)=cos(x).

    I have been told that [tex]|X_{n+1}-X_{n}| \leq 2^{(1-n)}[/tex]

    The attempt at a solution

    Well basically, I tried using m,n>N and N>0 and m>n>N to say that

    [tex]|X_m-X_n| = cos(X_{m-1}) - cos(X_{n-1})[/tex]

    because this is the way i was doing it for simpler sequences.

    However, I havent got a clue where I would go from here which leads me to thinking I have gone the wrong way about it!
    Any help would be greatly appreciated.
    Last edited: May 16, 2010
  2. jcsd
  3. May 16, 2010 #2
    OK, how do we prove that something is a Cauchy sequence?

    First, a very mean man who hates undergrad analysis students is going to choose an [itex]\epsilon > 0[/itex]. What we have to do is, for whatever [itex]\epsilon [/itex] he chooses, be able provide a number N such that for any two numbers m and n that are both greater than N,

    [tex] |{X_n - X_m}| < \epsilon [/tex].

    If we can do this, then it's a Cauchy sequence.

    So now it would be nice if we were able to do cool things with [itex] cos X_{m-1} [/itex], but I don't really see how that's going to help us. However, the inequality they told you about looks promising. Try to figure out a way to provide a number N in response to any [itex] \epsilon [/itex] using that inequality.
  4. May 17, 2010 #3
    Hello, thanks for the reply.

    Here is what I have done:

    Take m,n>N
    [tex]\epsilon > 0[/tex]

    Take m = n+1 then

    [tex]|X_m-X_n| \leq 2^{1-n} \leq 2^{-n}[/tex]

    But then i don't know where to fit epsilon into it?

    I actually have the answer in front of me, but it doesnt use epsilon at all, or N, which is what I have been used to using. So I wanted to see if there was a way to do it with them included.
  5. May 17, 2010 #4
    What definition of Cauchy sequences are you using that doesn't contain an epsilon? Maybe there is some other variable used, like a lot of calculus textbooks now use "h" instead of [itex] \DELTA x [/itex] in the definition of a derivative?

    In any event, you're not allowed to choose m and n; the inequality has to be true for ANY m and n which are both greater than N. Otherwise, you'd have the following (FLAWED) proof:

    Let [itex] \{ x_n \} = \ln x [/itex]. Let [itex] m = n+1 [/itex]. Then [itex] |X_m - X_n| = | \ln (n+1) - \ln n | = \ln \frac{n+1}{n} [/itex].

    Now [itex] \lim_{n \rightarrow \infty} \ln \frac{n+1}{n} = 0. [/itex] Hence for any [itex] \epsilon [/itex], there is an N such that for n > N [itex] \ln \frac{n+1}{n} [/itex] < [itex] \epsilon [/itex]. So [itex] \{ x_n \}[/itex] is Cauchy.

    Of course, this isn't true; ln x doesn't converge at all.
  6. May 17, 2010 #5
    Yes I thought that was odd. But it is a follow on question so perhaps they had proved it before, I'm not sure.
    But I want to be able to do it just based on this.

    So I can't take m = n+1, that makes sense. I was just clutching at straws really, trying to use the inequality provided.

    So could we take [tex]\epsilon > 0[/tex] and then we have [tex]|X_m-X_n| \leq 2^{1-n}[/tex]
    but m,n>N, so if we take N > 1

    [tex]|X_m-X_n| \leq 2^{1-n} \leq 2^{1-N} \leq 2^{-N}[/tex]

    Is that wrong to just assume that?
    Sorry, I am not very good at these at all! Thanks for your patience :)
  7. May 17, 2010 #6
    You might try doing this.

    Let [itex] \epsilon = .1 [/itex].

    Now give me N such that that no matter what m and n I choose greater than N, [itex] | X_m - X_n | < \epsilon [/itex].

    Then try it for [itex] \epsilon = .001 [/itex]. Then .00001, etc, until you get the method down.

    Then generalize so that you can give me an N for any [itex] \epsilon [/itex] I choose.

    Then in your proof, just write down the method of generating Ns in response to an [itex] \epsilon [/itex] and show that it makes [itex] | X_m - X_n | < \epsilon [/itex].

    There will be a DIFFERENT N for each [itex] \epsilon [/itex], by the way. As [itex] \epsilon [/itex] gets smaller, N gets bigger.
  8. May 17, 2010 #7
    you don't have to go back to the cauchy definition, perhaps it was proven somewhere that either

    [tex] | a_{n} - a_{n+1} | \leq c^{n} [/tex] where |c| < 1
    [tex] | a_{n} - a_{n+1} | \leq c^{-n} [/tex]
    implies that the sequence is cauchy. In both cases, you can obviously see that as n gets large, c^n (where 0 < c < 1) or c^(-n) can be made arbitrarily small.

    even if it was not proven formally anywhere, you can still use this
  9. May 18, 2010 #8
    Okay that sort of makes sense, i have C defined in the set [-1,1] in a previous question. (if it is the same C !)
    but what about the [tex]2^{n-1}[/tex]
    where does that come in in terms of C ?
  10. May 20, 2010 #9
    [tex] 2^{1-n} [/tex] becomes small as n becomes large
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook