Results 1 to 7 of 7

Math Help - Limit Theorems, Sequences

  1. #1
    Member
    Joined
    Oct 2008
    Posts
    124

    Limit Theorems, Sequences

    Suppose that (sn) and (tn) are convergent sequences with lim sn = s and lim tn = t. Then lim (sntn) = st.

    Write a proof to this theorem that does not use the theorem that states that "every convergent sequence is bounded." I think that this can be done by using the identity sntn - st = (sn - s)(tn - t) + s(tn - t) + t(sn - s). I am lost from here.
    Follow Math Help Forum on Facebook and Google+

  2. #2
    MHF Contributor

    Joined
    Aug 2006
    Posts
    18,381
    Thanks
    1474
    Awards
    1
    If \varepsilon  > 0 then choose \delta  = \min \left\{ {\sqrt {\frac{\varepsilon }{3}} ,\frac{\varepsilon }<br />
{{3\left( {\left| s \right| + 1} \right)}},\frac{\varepsilon }<br />
{{3\left( {\left| t \right| + 1} \right)}}} \right\}.

    Make \left| {s_n  - s} \right| < \delta \;\& \,\left| {t_n  - t} \right| < \delta .

    Then use the identity.
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Member
    Joined
    Oct 2008
    Posts
    124
    I'm still lost...
    Follow Math Help Forum on Facebook and Google+

  4. #4
    Banned
    Joined
    Mar 2009
    Posts
    256
    Thanks
    1
    Quote Originally Posted by bearej50 View Post
    I'm still lost...

    The identities that you must use are:


    1)  s_{n}t_{n}-st = s_{n}t_{n}-st_{n}+st_{n}-st = t_{n}(s_{n}-s)+s(t_{n}-t)

    OR

    2)  s_{n}t_{n}-ts_{n}+ts_{n}-st = s_{n}(t_{n}-t)+t(s_{n}-s).

    In using for example the (1) identity we can work as follows:

    we want to find out if there exists a No ,k, belonging to the natural Nos,such that for any given positive No ε>0

     |s_{n}t_{n}-st| = |s_{n}t_{n}-st_{n}+st_{n}-st| = |t_{n}(s_{n}-s)+s(t_{n}-t)|<\epsilon,whenever ,  n\geq k.

    WHERE n is any natural No.

    But since |t_{n}(s_{n}-s)+s(t_{n}-t)|\leq |t_{n}||s_{n}-s| + |s||t_{n}-t|\leq |t_{n}||s_{n}-s| + (|s|+1)|t_{n}-t|,if we can find a ,k, such that for :

     n\geq k ,then  |t_{n}||s_{n}-s|<\frac{\epsilon}{2} ,and

     (|s|+1)|t_{n}-t|<\frac{\epsilon}{2},we will have proved the desired result.

    Note instead of using |s| we use ||s|+1| in the above inequalities is to avoid the case where s=0.

    HINT: In trying to find this ,k, it must be the max{  n_{1},n_{2},n_{3}} i,e the maximum of three other natural Nos
    Follow Math Help Forum on Facebook and Google+

  5. #5
    Banned
    Joined
    Mar 2009
    Posts
    256
    Thanks
    1
    bearej50.

    Since there is no respond from you ,i will assume that you still do not know how to do the limit,so i will prove it.

    So we want to prove \lim_{n\rightarrow\infty}{s_{n}t_{n}} = st

    Let ε>0.

    Thus \frac{\epsilon}{2(|t|+\epsilon)}>0.................................................. ................................1


    and \frac{\epsilon}{2(|s| +1)}>0.................................................. ........................................2

    Since now \lim_{n\rightarrow\infty}{s_{n}} = s and \lim_{n\rightarrow\infty}{t_{n}} = t,given any +ve No and thus even the +ve Nos (1),(2),ε,
    we can find Natural Nos  n_{1},n_{2},n_{3} and such that:


    If  n\geq n_{1},then |t_{n}|<|t|+\epsilon.................................................. ...........................................3


    If  n\geq n_{2},then |s_{n}-s|<\frac{\epsilon}{2(|t|+\epsilon)}.................................................. .........................................4


    If  n\geq n_{3},then |t_{n}-t|<\frac{\epsilon}{2(|s|+1)}.................................................. ...........................................5

    Choose k = max{  n_{1},n_{2},n_{3}}

    Hence ,

    If  n\geq k,then  n\geq n_{1},n\geq n_{2},n\geq n_{3},

    HENCE

    |t_{n}|<|t| +\epsilon.................................................. ........................................6


    |s_{n}-s|<\frac{\epsilon}{2(|t|+\epsilon)}.................................................. ........................................7


    |t_{n}-t|<\frac{\epsilon}{2(|s|+1)}.................................................. .........................................8

    Now multiply (6) by (7) and we get:


     |t_{n}||s_{n}-s|< (|t|+\epsilon)\frac{\epsilon}{2(|t|+\epsilon)} = \frac{\epsilon}{2}.................................................. ...........................................9

    And by multiplying (8) by (|s|+1) we get:


     (|s|+1)|t_{n}-t|< (|s|+1)\frac{\epsilon}{2(|s|+1)} =\frac{\epsilon}{2}.................................................. ..........................................10

    Add (9) and (10) we get:


    |t_{n}||s_{n}-s|+(|s|+1)|t_{n}-t|< \frac{\epsilon}{2} +\frac{\epsilon}{2} = \epsilon.

    Hence  |t_{n}s_{n}-ts|<\epsilon.

    Therefor we have proved ,given any +ve No ,ε Τhere exist a natural No k such that :

    If  n\geq k then |t_{n}s_{n}-ts|< \epsilon ,for all n
    Follow Math Help Forum on Facebook and Google+

  6. #6
    MHF Contributor

    Joined
    Aug 2006
    Posts
    18,381
    Thanks
    1474
    Awards
    1
    Re the last post.
    Did you note that the OP said not to use the bounded theorem ?
    Your proof uses it.
    Follow Math Help Forum on Facebook and Google+

  7. #7
    Banned
    Joined
    Mar 2009
    Posts
    256
    Thanks
    1
    Nowhere my post uses the the bounded theorem

    Read my post again
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. [SOLVED] Applying the limit theorems.
    Posted in the Differential Geometry Forum
    Replies: 1
    Last Post: September 21st 2009, 05:10 PM
  2. limit theorems- help needed!
    Posted in the Advanced Statistics Forum
    Replies: 1
    Last Post: April 6th 2009, 09:22 PM
  3. Analysis Limit Theorems
    Posted in the Differential Geometry Forum
    Replies: 2
    Last Post: March 12th 2009, 09:38 AM
  4. question on limit theorems hw!!!
    Posted in the Advanced Statistics Forum
    Replies: 3
    Last Post: October 16th 2008, 07:50 PM
  5. binomial distributions and limit theorems!!!
    Posted in the Advanced Statistics Forum
    Replies: 1
    Last Post: October 15th 2008, 11:19 PM

Search Tags


/mathhelpforum @mathhelpforum