4.2 Properties of Limits

Now that we have established how to prove the existence of limits, we will consider the uniqueness of them, as well as some algebraic properties of convergent sequences.

Theorem 4.2

If a sequence converges, its limit is unique.

Proof.

Let (xn)n be a convergent sequence and suppose that (xn)nx and (xn)ny. Let ε>0. Then there exists N1 such that nN1|xnx|<ε, but there also exists N2 such that nN2|xny|<ε. Now,

|xy| =|xxn+xny| (for nmax{N1,N2})
|xxn|+|xny| (4.14)
<ε+ε=2ε. (4.15)

Therefore, |xy|=0, so x=y (since ε>0, 0a<εa=0a). Thus, the two limits are in fact the same. ∎

The next theorem will serve as a stepping stone that will make some more useful theorems much easier to prove. First, recall that a set is bounded if and only if it is bounded both above and below. We extend this definition to the elements of sequences, saying that a sequence is bounded if there exists M0 such that |xn|Mn. This is a very quick and easy way to establish the range of a sequence that will come in handy later.

Theorem 4.3

Every convergent sequence is bounded11 1 Note that the converse of this theorem is obviously not true, ((1)n)n is bounded, but does not converge..

Proof.

Let (xn)n be a convergent sequence with xnx. Then there exists N such that nN|xnx|<1. Consider the terms in the sequence x1,x2,,xN1 and let M=max{|x1|,|x2|,,|xN1|}. Now, for all nN, |xn|max{M,|x|+1} (since |xnx||xn||x|<1|xn|<1+|x|), which is a constant, hence (xn)n is bounded. ∎

4.2.1 Adding and Multiplying Sequences

Now that we have this tool, we can proceed to some more interesting results about how limits of sequences behave when we add and multiply sequences together.

Theorem 4.4

Let a, (xn)n and (yn)n converging sequences with xnx and yny.

  1. (i)

    xn+ynx+y

  2. (ii)

    xnynxy

  3. (iii)

    If n,yn0 and y0, then 1yn1y

  4. (iv)

    If n,yn0 and y0, then xnynxy

  5. (v)

    axnax

  6. (vi)

    If n,xnyn, then xy.

Proof.

For all parts of this proof, begin by letting ε>0.

  1. (i)

    Since xnx,N1 such that nN1|xnx|<ε2 and since yny,N2 such that nN2|yny|<ε2. Choose N=max{N1,N2}, now for nN we have

    |(xn+yn)(x+y)| =|(xnx)+(yny)| (4.16)
    |xnx|+|yny| (4.17)
    <ε2+ε2=ε. (4.18)
  2. (ii)

    Since (xn)n is bounded, M0 such that |xn|Mn. Since xnx,N1 such that nN1|xnx|<ε2M and since yny,N2 such that nN2|yny|<ε2|y|. Now, choosing N=max{N1,N2}, for nN we have

    |xnynxy| =|xnynxny+xnyxy| (4.19)
    =|xn(yny)+y(xnx)| (4.20)
    |xn||yny|+|y||xnx| (4.21)
    <|xn|ε2M+|y|ε2|y| (4.22)
    Mε2M+|y|ε2|y|=ε2+ε2=ε. (4.23)
  3. (iii)

    Since yny,N1 such that nN1|yny|<|y|2. Then for nN1

    |y|=|yn+(yyn)||yn|+|yny|<|yn|+|y|2, (4.24)

    so |yn|>|y|2. Now, let N2 be such that nN2|yny|<ε|y|22, then for nmax{N1,N2} we have

    |1yn1y| =|yynyny|=|yny||yn||y| (4.25)
    <ε|y|22|yn||y|<2ε|y|22|y|2=ε. (4.26)
  4. (iv)

    Write xnyn as xn1yn, then the result follows from parts (ii) and (iii).

  5. (v)

    Let yn=a (so (yn)n is the constant sequence (a,a,) with limit a), then the result follows from part (ii).

  6. (vi)

    We will make use of a minor result to assist in proving this.

    Lemma

    Suppose (an)n is a convergent sequence with ana. If an>0n then a0.

    Proof.

    Suppose by way of contradiction that a<0 and let ε=a>0. Then since (an)n converges, N such that nN|ana|<a. This means that

    a< ana<a (4.27)
    ana+a<a+a (4.28)
    an<0 (4.29)

    which is a contradiction, hence a must be greater than 0. ∎

    Now, consider ynxn0n. The sequence (ynxn)n converges to yx by part (i) (and part (v) with a=1) and so by our claim yx0, which implies xy as required.

4.2.2 Sandwich Principle

A little observation from the last part of this theorem is the sandwich principle, which is that if we have three sequences — (an)n, (xn)n and (bn)n with limits a, x and b respectively — with anxnbnn, then axb. In the special case that a=b, we can use this principle to determine the limit of an unknown sequence (xn)n as in this case a=x=b. We will finish off this section with a couple of examples about how to use the arithmetic properties of limits.

Example 4.6.

Let (xn)n=(1+1np)n with p. Show that xn1 as n.
First, let an=1 and bn=1np, then note that we already know that the sequence (1n)n converges to 0, so by theorem 4.4 (ii) (applied repeatedly),

1n1n1np times=1np0 as n. (4.30)

Then (xn)n=(an+bn)n which, by theorem 4.4 (i), converges to 1 since an1 and bn0 as n.

Example 4.7.

Show that the sequence (xn)n=(n+5n2+2n3n3+3) converges to 2.

2n3+5n2+nn3+3=2+5(1n)+(1n)21+3(1n)32+5(0)+01+3(0)=2 as n. (4.31)