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: Limits, geometric series, cauchy, proof HELP

  1. Dec 13, 2011 #1
    i guys, I'm stuck on wording of a homework assignment and thought you might be able to help me. There are several questions...

    Consider the geometric series: (Sum from k=0 to infinity) of ar^k
    and consider the repeating decimal .717171717171 for these problems:

    Question 1:
    Find a formula for the n-th partial sum of the series and PROVE your result using the Cauchy Convergence Criterion. This technique requires to find epsilon, n, N, etc…

    Question 2:
    Use your formula from Q1 above to determine which conditions on "a" and/or "r" guarantee that the geometric series converges. And PROVE your result.

    Question 3:
    Write the repeating decimal .7171717171717171... as a geometric series.

    Question 4:
    Find the sum of the geometric series in Q3 above to get a fractional representation of your repeating decimal.

    My attempts at solving these:

    My issues with Q1 are:
    - is the formula for the nth partial sum a/(1-r) or [a-ar^(n+1)] / (1-r)?
    - the problem notes sum from 0 to infinity; does this change the soln?
    - our professor is very picky when it comes to proofs. we have to use the Cauchy criterion to prove this. We can't just use calculus methods to prove things converge such as limit test, ratio test, integral test etc...

    My issues with Q2 are:
    -How do I prove this??

    My issues with Q3 and Q4 are:

    I wrote in the margin on my hw page that I'd have to "prove it has equivalent rational representation..." but I don't know what this applies to?

    so far I have that .717171… = (sum from k=0 to infinity) of .71(.01)^k
    but I'm lost as far as doing the proofs…

    Thanks for any help guys! I have today and tomorrow to do this. Thanks for any guidance!!
  2. jcsd
  3. Dec 13, 2011 #2
    Can you state the Cauchy convergence criterion??
  4. Dec 13, 2011 #3
    All the cauchy proofs I have been given begin with:

    Let epsilon >0 be given. We want to find N element of the naturals such that n>=N implies that abs[...- thing converges to] < elsilon. Then we need to choose an N and yada yada...

    I'm going to be honest when I say I don't get it either...
  5. Dec 13, 2011 #4
    can anybody help me with this? :confused:
  6. Dec 13, 2011 #5


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I assume that the "Cauchy convergence criterion" is simply that when we want to prove that a sequence of real numbers is convergent, it's sufficient to prove that it's a Cauchy sequence (because we know that every Cauchy sequence is convergent).

    You haven't stated the definition of a Cauchy sequence correctly, so I suggest you start with that.

    The nth partial sum of a series [itex]\sum_{k=0}^\infty x_k[/itex] is [itex]\sum_{k=0}^n x_k[/itex]. If you understand that, you shouldn't have a hard time figuring out which one of your formulas for the nth partial sum of [itex]\sum_{k=0}^\infty ar^k[/itex] is the right one.

    The main reason why you haven't received a lot of help is that you didn't do the most obvious thing: Write down the relevant definitions and make an attempt to apply them. The more work you put in, the more help you will get. If you don't do the obvious first step, you are much less likely to get a reply.
    Last edited: Dec 13, 2011
  7. Dec 13, 2011 #6
    OK, here's what i've got for the definition of the cauchy criterion. But i don't know how to apply it to these problems. We've done literally 5 examples using cauchy, and i don't understand any of them...


    Someone else gave me this but it's all greek to me

    Last edited: Dec 13, 2011
  8. Dec 13, 2011 #7


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I have a question about your Q1. It says "prove your result". What result? The only thing you're supposed to before that is to find a formula for the nth partial sum, so that formula is the only result you have at that point. And if you have found it, I assume that you have already proved it. It's not the sort of thing you would just guess. I think you need to post the exact statement of the problem.
  9. Dec 13, 2011 #8
    believe me, i agree with you that the questions don't make sense. Here is the exact form we were given. Each student was assigned a different value, mine was .71717171...

    And i know we need to use the cauchy criterion to prove something converges to a certain value. Can't just say r < 1 implies it converges.

  10. Dec 13, 2011 #9
    That's something completely different from your OP!!!!

    You need to
    1) find a formula for the partial sums. (Cauchy criterion has nothing to do with this)
    2) prove that the partial sums converge using the Cauchy criterion.

    So I suggest you start by finding and proving a formula for the partial sums:

  11. Dec 13, 2011 #10
    Ok, does this help?


    or this?

  12. Dec 13, 2011 #11


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    OK, I think what you're supposed to do in Q1 (book problem I(a)) is to prove that the formula for the nth partial sum that you already seem to be familiar with holds, nothing more. So Q1 has nothing to do with convergence. It's a finite sum. Do you know how to do this? (Please post your solution, not a picture of someone else's).
  13. Dec 13, 2011 #12

    nope. :uhh:

    All i know is that the cauchy criterion needs to be used somewhere in this assignment...
    And i have to apply the .717171717 to all parts of this assignment, not just the last parts.
  14. Dec 13, 2011 #13
    Yepp, totally lost. Do i do it like these examples?

    Last edited: Dec 13, 2011
  15. Dec 13, 2011 #14


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    You're supposed to solve problems I (a)-(b) and II (a)-(b), right? I see no reason to think that .7171... should be involved in problem I, and no reason to think that the Cauchy criterion should be involved in problem I (a). Problem I (a) isn't even about a series. It's a finite sum.

    Problem I (a) is about one thing: You need to rewrite [itex]1+x+x^2+\cdots+x^n[/itex] in a simpler form. I will tell you the trick. For all real x such that x≠1,
    [tex]1+x+\cdots +x^n=\frac{(1+x+\cdots+x^n)(1-x)}{1-x}=\cdots[/tex] Can you take it from here? (The problem you were given was very slightly different, but you should be able to figure out how to deal with that difference).

    Your solution of II (a) should look like those examples.
  16. Dec 13, 2011 #15
    I guess you're right about not using the repeating decimal .717171 and cauchy for part 1. I have no idea...

    I hate to sound like an idiot but i can't follow where you're trying to go... :/

    EDIT: I posted again, we went to a second page now...
    Last edited: Dec 13, 2011
  17. Dec 13, 2011 #16


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Please multiply those two factors in the numerator together.

    Did it not occur to you to try that? You don't always have to see where a calculation is going when you start. Sometimes you just have to try something and see where it takes you.
  18. Dec 13, 2011 #17
    so [1-x^(n+1)] / (1-x) ?
  19. Dec 13, 2011 #18
    hmm... why am i still lost? :cry:

    I really don't mean to sound like a dolt, i just don't understand what the answers are even supposed to "look" like. My teacher is beyond terrible and can't explain anything...

    and when do i use cauchy??
    Last edited: Dec 13, 2011
  20. Dec 13, 2011 #19


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    That's the final result yes. But how did you get that? When you just write down the final result, I have no idea if you actually did the multiplication, and neither will your teacher when you hand it in. (Are you supposed to hand it in?)

    In this case, I think problem II (b) explains the reason why you're doing these things. The goal is to prove that a repeating decimal expansion can always be expressed as a fraction n/m, where n and m are integers. The other problems can be thought of as intermediate results on the way there, that are also supposed to give you some practice with geometric series and Cauchy sequences, both of which are very important concepts on their own.

    I think the reason why you feel completely lost is that you are so unwilling to try anything unless you know that the result will be what you want. This seems to really be holding you back. Sometimes you just have to ask yourself "is there any way I can rewrite this expression", and then just do it to see if the result is easier to work with.

    In problem I (b). That looks like the hardest part. I think someone else will have to help you with that, because I'm going to bed soon. But even if no one else shows up for a while, don't let that stop you from trying out some ideas.

    By the way, you can't be sure that your teacher won't read your posts here.
  21. Dec 13, 2011 #20
    I contacted my teacher, and this is what they said:

    You need to prove that your partial sum is correct (I cannot comment whether it
    is or not) -- that is, you will need to use induction. And then you need to
    prove that the limit of your partial sum converges to your guess. That is, you
    will need an appropriate epsilon-N proof.

    and yes indeed, we need to write up these answers and hand them in. This is actually a project that's a big portion of our grade. After this i'll never see math again!! Just need to get these problems done so i can move on...!

    So my questions now are, what is the partial sum in 1a? I could probably do the induction to prove it, if i was only clear as to what the partial sum was.... And for 1b, how do i use cauchy? I'm being honest when i say i don't know how to even begin.
    Last edited: Dec 13, 2011
  22. Dec 13, 2011 #21


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    The multiplication we did for I (a) can be considered non-rigorous, because even though it looks obvious, we didn't really prove that all those other terms cancel each other out*. That's why you have to use induction to prove the formula we found, or rather to prove the formula you will find when you repeat the above calculation with [itex]ar^n[/itex] instead of [itex]x^n[/itex].

    *) A mathematician doesn't actually consider that step non-rigorous, because he immediately knows that he could prove it by induction if he wanted to. Students with less experience should however do every step for pedagogical reasons, so I think your teacher is right to demand it here.
  23. Dec 13, 2011 #22


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    You don't need to be 100% sure that you have the right formula, because if you have the wrong one, the induction proof can't possibly succeed, and the failure should make you reconsider the formula.

    The very first thing you need to do is to study the definition of a Cauchy sequence, and make sure you understand it.
  24. Dec 13, 2011 #23
    OK, for the induction, can you at least set it up for me? If the questions are "worded right" i'll know what to do...
    For example, if you ask me to prove that 1 + x + x^2 + ... + x^n = [x^(n+1)-1] / (x-1) using induction, i could do it. So what's the "left side" and "right side" of what i need to do for this assignment?

    And as far as using cauchy, i've only done a few problems, and they've all been worded like:
    "for the sequence Sn= 1 + 1/(2^n) find the limit and prove your sequence converges to your limit..." So i have NO idea how this assignment is asking me to use cauchy. Where am i taking limits? What's converging?

    Do you see what i'm trying to say? :confused:

    So is this right for the induction part?


    And i know that my Sn converges to a/(1-r) ....is this were i'm supposed to use cauchy to prove it? If so, i don't know how to do so. Can you set it up for me?
    Last edited: Dec 13, 2011
  25. Dec 14, 2011 #24


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    The induction part of it is essentially correct. It can be better stated though. Remember that the point of an induction proof is that it allows you to prove infinitely many statements in a finite number of steps. Each P(n) is a statement, and you want to prove that they're all true. Induction allows you to do that by proving only two statements:

    1. P(0) is true.
    2. For all non-negative integers n, if P(n) is true, then so is P(n+1).

    You have the right idea about how to do this sort of thing, but the notation is sometimes ugly. For example, you wrote P(0)=a+ar+...+arn. That doesn't make sense, because it says that a statement is equal to a number. Kind of like saying that an elephant is equal to a Tuesday.

    I have to ask, did you actually do the multiplication I asked you to do? [tex](1+x+\cdots+x^n)(1-x)=?[/tex] Even if you did, I suggest that you do it again. :wink:

    We need to make sure that you understand a few other things:

    1. Do you understand what it means to say that a series is convergent? Specifically, if I write [itex]\sum_{k=0}^\infty a_k=s[/itex], do you know what that means?

    2. Do you understand what it means to say that a sequence is convergent? Specifically, if I say that [itex]s_n\to s[/itex], do you know that means?

    3. Do you understand the definition of Cauchy sequence? Specifically, if I say that [itex]\langle s_n\rangle_{n=0}^\infty[/itex] is a Cauchy sequence, do you know what that means? (That's the notation I use for the sequence [itex]s_0,s_1,\dots[/itex]. Your teacher may use something different).
  26. Dec 14, 2011 #25


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    By the way, if you have to post pictures, try to keep them smaller. It's better if you type the mathematics. Click the quote button next to one of my posts, and you'll see how i did the math there. There's a LaTeX guide for this forum. You just need to read a small part of it to get started.

    This is how I would state the problems you're supposed to do:

    (a) Find a formula for the nth partial sum of [itex]\sum_{k=0}^\infty ar^k[/itex], and use induction to prove that it holds for all non-negative integers n.
    (b) Find conditions on a and r such that the sequence of partial sums of [itex]\sum_{k=0}^\infty ar^k[/itex] is a Cauchy sequence if and only if those conditions are satisfied. (Note that this means that the series is convergent if and only if those conditions are satisfied).
    (c) Use the formula from I(a) to guess a simple formula for [itex]\sum_{k=0}^\infty ar^k[/itex] that holds when the conditions on a and r from II(b) are satisfied. Prove that the sum of the series is in fact what you guessed.

    (Yes, I know that there's no I(c) in your book, but it certainly looks like this is a part of what you're supposed to do, and I can't tell if your teacher considers it a part of I(b) or II(b), so I stated it as a separate problem).

    (a) Write 0.7171... as a geometric series.
    (b) Prove that there exist integers p and q such that 0.7171...=p/q, by using the formula for the sum of a geometric series found in I (c) on the series found in II (a).

    Keep in mind that I have to be careful to not tell you too much. I can only give you hints about how to proceed, so the best way to keep this moving along is to always post your work up to the point where you are stuck. If the next step after that is just to apply a definition, you should only expect replies like "use the definition".
    Last edited: Dec 14, 2011
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook