Homework 11 Solutions

10.3) Suppose we are given a decimal expansion k.d1 d2 d3 d4 ..., where k is a nonnegative integer and
dj ∈ {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} ∀j ∈ N. Define the corresponding sequence as

d1 d2 dn
sn = k + + + ... + n .
10 102 10
Fix n ∈ N. So, we have
d1 d2 dn 9 9 9 1
sn = k + + 2 + ... + n ≤ k + + 2 + ... + n = k + 1 − n < k + 1,
10 10 10 10 10 10 10
because of the formula given in the Hint. Since n ∈ N was arbitrary, we conclude sn < k + 1 ∀n ∈ N.

For the next problem, we need the following fact, which is used in developing the Geometric Series
formula (See Example 1 in Section 14):

Proposition 1. For n ∈ N and a, r ∈ R with r 6= 1, we have the following formula

1 − rn+1
 
2 n
a(1 + r + r + ... + r ) = a .
1−r

Notice that the Hint in the above problem (10.3) uses this fact.

10.6) (a) Suppose {sn } is a sequence with |sn+1 − sn | < 2−n ∀n ∈ N. Before we start proving the
sequence is Cauchy, we need to get an inequality of only one index n. Notice for m, n ∈ N with m > n, we
have by repeated addition by zero (in a bunch of ’clever’ disguises)

|sm − sn | = |sm − sm−1 + sm−1 − sm−2 + sm−2 − sm−3 + ... + sn+1 − sn |

Then, using the triangle inequality many times, we get

1 1 1 1
|sm − sn | ≤ |sm − sm−1 | + |sm−1 − sm−2 |... + |sn+1 − sn | < + + ... + + .
2m−1 2m−2 2n+1 2n
Through algebraic manipulation and using Proposition 1 results in
 2  m−n−2  m−n−1 !
m−n !
1 1 1 1 1 1 1 − 21 1 1
|sm − sn | < n 1 + + + ... + + = n 1 = n−1 − m−1 .
2 2 2 2 2 2 1− 2 2 2

Finally, we can obtain the following inequality which will be used to prove the sequence is Cauchy
1 1 1
|sm − sn | < − < ,
2n−1 2m−1 2n−1
since m > n.
Since n, m ∈ N with m > n was arbitrary, we have the following
1
|sm − sn | < ∀n, m ∈ N with m > n. (1)
2n−1

1
 1
Now we move on to proving the sequence is Cauchy. Let  > 0 be given. Since lim n−1 = 0, ∃N ∈ R
  n→∞ 2
 1 
such that ∀n > N , then  n−1 − 0 < . We choose this N . Without loss of generality we assume m > n
 
2
1
(otherwise, you just switch the indices). Thus, ∀m, n > N , then |sm − sn | < n−1 <  by equation (1)
2
since m > n.
Therefore, we conclude the sequence is Cauchy and must converge by Theorem 10.11.

1 1


(b) No. Consider the sequence sn = ln n. Clearly, |sn+1 −sn | = ln n + 1−ln n = ln 1 + n < n ∀n ∈ N,
but lim ln n = ∞. Since sequence diverges, it cannot be Cauchy by Theorem 10.11.
n→∞

10.10) Let s1 = 1 and sn+1 = 13 (sn + 1) for n ≥ 1 (i.e. a recursion relation).

(a) s1 = 1, s2 = 23 , s3 = 59 , and s4 = 14
27 .

1
(b) First, we show {sn } is bounded below by 2 with induction:

1
(i) This is true for n = 1, since 2 < s1 = 1.

1
(ii) Let n ∈ N. Suppose that sn > 2 is true. Then, we have
 
1 1 1 1
sn+1 = (sn + 1) > +1 = .
3 3 2 2

So the statement is true for n + 1.

1
Thus, by the Principle of Mathematical Induction, we conclude 2 < sn ∀n ∈ N, and the sequence is
bounded below.

(c) Second, we show sn ≥ sn+1 ∀n ∈ N (i.e. sequence is non-increasing) by induction:

2
(i) This is true for n = 1, since s1 = 1 < 3 = s2 .

(ii) Let n ∈ N. Suppose that sn ≥ sn+1 is true. Then, we have

1 1
sn+1 = (sn + 1) ≥ (sn+1 + 1) = sn+2 .
3 3
So the statement is true for n + 1.
Thus, by the Principle of Mathematical Induction, we conclude sn ≥ sn+1 ∀n ∈ N, and the sequence is
monotonically non-increasing.

(d) By parts (b) and (c) we know {sn } is a bounded monotone sequence, and we conclude it must
converge by Theorem 10.2.
Since we know the sequence converges. Let s = lim sn . Then, from the recursion relation, we have
n→∞

1 1 1
lim sn+1 = lim (sn + 1) ⇒ s = (s + 1) ⇒ 3s = s + 1 ⇒ s = .
n→∞ n→∞ 3 3 2

2
 1
Therefore, we also conclude lim sn = .
n→∞ 2

 
1
10.11) Let t1 = 1 and tn+1 = 1 − 2 t2n for n ≥ 1 (i.e. a recursion relation).
4n

(a) First, notice {tn } is non-increasing since

 
1
tn+1 = 1 − 2 t2n < 1 · tn = tn ∀n ∈ N.
4n

Also, this shows that the sequence is bounded above by 1, since t1 = 1. Second, we show {tn } is bounded
below by 0 with induction:

(i) This is true for n = 1, since 0 < 1 = t1 .

(ii) Let n ∈ N. Suppose that 0 < tn is true. Then, we have

 
2 1 1 1
4n > 1 ⇒ 2 < 1 ⇒ 1 − 2 > 0 ⇒ tn+1 = 1 − 2 t2n > 0.
4n 4n 4n

So the statement is true for n + 1.

Thus, by the Principle of Mathematical Induction, we have 0 < tn ∀n ∈ N, and the sequence is
bounded.
Therefore, {tn } is a bounded monotone sequence, and we conclude it must converge by Theorem 10.2.

3
 Worksheet 5 Solutions

4) Suppose lim sn = s and lim tn = t with sn > 0 ∀n ∈ N and s > 0. Then, we have the following
n→∞ n→∞

lim stnn = lim etn ln sn = elimn→∞ tn ln sn ,

n→∞ n→∞

by Problem Sheet 4.2 since f (x) = ex is continuous. Continuing using Theorem 9.4, we have

elimn→∞ tn ln sn = e(limn→∞ tn )(limn→∞ ln sn ) = et ln s ,

by Problem Sheet 4.2 since f (x) = ln x is continuous. Finally, we get

lim stnn = et ln s = st ,
n→∞

and we conclude if lim sn = s and lim tn = t with sn > 0 ∀n ∈ N and s > 0, then lim stnn = st .
n→∞ n→∞ n→∞

5) Suppose E is nonempty subset of R which is bounded below, and define L = {l ∈ R : l is a lower bound for E}.

(a) If l ∈ L, the l ≤ e ∀e ∈ E. Therefore, any e ∈ E is an upper bound for L, and sup L ∈ R exists by
the Completeness Axiom.

(b) Since ∀e ∈ E is an upper bound for L, sup L ≤ e ∀e ∈ E. Therefore, sup L is a lower bound for E.
Since l ≤ sup L ∀l ∈ L, we have sup L = inf E by definition.

Remark: This is an alternate proof of Corollary 4.5 of the Completeness Axiom.