Scribd
Upload
44 views4 pages

Sequence EC

1. The document discusses two examples of applying the Bounded Convergence Theorem: - Finding the value of 6 + 6 + 6 + ... by defining it as the limit of the sequence a1 = 6, a2 = 6 + 6, etc. This sequence is shown to be bounded below 3 and increasing, so it has a limit of 3. - Using Newton's method to find √2 by the sequence x1 = 2, x2 = (x1^2 + 2)/2x1, etc. This sequence is shown to be bounded above 2 and decreasing, so it has a limit of √2. 2. Several extra credit problems are proposed applying the same techniques to

Uploaded by

M Shahbaz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
44 views4 pages

Sequence EC

1. The document discusses two examples of applying the Bounded Convergence Theorem: - Finding the value of 6 + 6 + 6 + ... by defining it as the limit of the sequence a1 = 6, a2 = 6 + 6, etc. This sequence is shown to be bounded below 3 and increasing, so it has a limit of 3. - Using Newton's method to find √2 by the sequence x1 = 2, x2 = (x1^2 + 2)/2x1, etc. This sequence is shown to be bounded above 2 and decreasing, so it has a limit of √2. 2. Several extra credit problems are proposed applying the same techniques to

Uploaded by

M Shahbaz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

On the Bounded Convergence Theorem

I gave two examples of the Bounded Convergence Theorem in class.


Recall: The Bounded Convergence Theorem says that any bounded monotonic
sequence has a limit. Here are my two examples:

1. Make sense of 6+ 6+ 6 + … and determine its value.

For any such thing, we define it as the limit of a sequence. In this case, the
sequence is: a1 = 6 , a2 = 6+ 6 , a3 = 6+ 6+ 6 , ….

To work with such a sequence, it almost always helps to find a description


that relates the terms of the sequence to previous terms. In this case, the relation
is that an+1 = 6 + an . That is, to get the next term in the sequence, add 6 and
take the square root. With this, we can ask about boundedness and about
monotonicity.

The sequence is bounded: Let’s show that 0 < an < 3 for all n. Certainly,
0 < 6 < 3, which we can check by squaring each term. It is clear that all the
a’s are positive numbers, so we want to show that an < 3 for all n. To do this, we
show that once one term is less than 3, so is the next. If an < 3, then
an+1 = 6 + an < 6 + 3 = 9 = 3.
Since a1 < 3, it follows that a2 < 3 as well. Now that a2 < 3, we must have
a3 < 3, and so on.

Note: I could have shown an < M for lots of different M’s. In class, I showed
that an < 10 for all n.

The sequence is monotonically increasing: One way to show an is an increasing


sequence is to show that an+1 > an for all n. Often, there are several approaches.
In class, I did the following: If an+1 = 6 + an , then we can also say that
an = 6 + an!1 . Now an+1 ! an = 6 + an ! 6 + an!1 . We multiply and
divide this by the sum of the radicals:
(6 + an) ! (6 ! an!1) an ! an!1
an+1 ! an = = .
6 + an + 6 + an!1 6 + an + 6 + an!1
page 2

a 2 ! a1
The point of this: Let n = 2, and we get a3 ! a2 = .
6 + an + 6 + an!1
# &
Knowing that a2 > a1 $$"because clearly, 6 + 6 > 6 ''% , it follows that a3 > a2.
Now that we know a3 > a2, the formula (with n = 3) tells us a4 > a3, and so on.

It turns out we could do something different: Knowing an < 3 for all n allows the
(6 + an) ! (an)2
following calculation instead: an+1 ! an = 6 + an ! an = .
6 + an + an
Now 6 + x ! x2 = (2 + x)(3 ! x). Using this on the numerator gives
(2 + an)(3 ! an)
an+1 ! an = > 0 because 3 ! an > 0.
6 + an + an

Now that we know the sequence is bounded and increasing, the bounded
convergence theorem says there is a limit. Given lim an = L, we can find L:
n()
L = lim an = lim an+1 = lim 6 + an = 6 + lim an = 6 + L.
n() n() n() n()

So L = 6 + L . Squaring, L2 = 6 + L, L2 ! L ! 6 = 0, (L + 2)(L ! 3) = 0,
so L = !2 or L = 3. Since L is positive, L = 3.

Just for kicks, here are the first several terms of the sequence: a1 = 6 * 2.449,
a2 = 6+ 6 * 2.9068, a3 * 2.9844, a4 * 2.9974, etc.

One last aspect of this example: Now we know the limit is 3. We can ask “How
quickly does an converge to 3?” This is called the rate of convergence. Here is
how to answer the question: study +an ! 3+. We have:
6 + an ! 9 an ! 3
an+1 ! 3 = 6 + an ! 3 = =
6 + an + 3 6 + an + 3
When n is large, an is near 3 so 6 + an is also near 3. That means
1
+an+1 ! 3+ * +an ! 3+.
6
What this means: The error decreases by a factor of about 6 with each term in the
sequence.
page 3

My second example was to use Newton’s method for finding the square root of 2.
In general, Newton’s method says that if p(x) is nice enough, and a is close to a
p(a)
zero of p(x), then you can improve the accuracy by replacing a by a ! .
p, (a)
p(xn)
You can form a sequence: x0 = a, xn+1 = xn ! . To use this to find 2 ,
p, (xn)
we let p(x) = x2 ! 2, which has 2 as one of its zeros. Then
p(xn) xn2 ! 2 2xn2 ! (x n2 ! 2) xn2 + 2
xn ! = xn ! = = ,
p, (xn) 2xn 2xn 2xn
xn2 + 2
or xn+1 = .
2xn
22 + 2
To find 2 , make an initial guess, say x0 = 2. Then x1 = = 1.5,
2 .2
1.52 + 2
x2 = * 1.41667, and so on. In class, we showed that xn is bounded, and
2.1.5
monotonic, so it has a limit, and the limit is 2 . The check went like this:

First, xn is bounded below by 2 :


xn2 + 2 xn2 + 2 ! 2 2 xn xn2 ! 2 2 xn + 2
xn+1 ! 2 = ! 2 = =
2xn 2xn 2xn
#$x ! 2 &' 2
$" n '%
= > 0.
2xn
This means xn > 2 for all n.

xn2 + 2 2 ! xn2
Next, xn is monotonically decreasing: xn+1 ! xn = ! xn = .
2xn 2xn
Since xn > 2 , the numerator is always negative. This means xn+1 ! xn < 0, so
the x’s are decreasing.

By the Bounded Convergence Theorem, the sequence has a limit. If we call that
xn2 + 2 L 2 + 2
limit L, L = lim xn = lim xn+1 = lim = . That is,
n() n() n ( ) 2xn 2L
page 4

L2 + 2
L= , so 2L2 = L2 + 2 - L2 = 2, or L = ± 2 . Since L > 0, L = 2.
2L

As before, we can say something about the rate of convergence from


#$x ! 2 &' 2
"$ n %'
xn+1 ! 2 = .
2xn
.2 1 2
If xn differs from 2 by ., then xn+1 will differ by , so xn+1 ! 2 < . .
2xn 2
1
That is, if xn ! 2 < 10!k, then xn+1 ! 2 < 10!2k (doubling of the number
2
of correct digits at each stage).

Finally, here is some fairly easy (?) extra credit you can think about:

1. Find the value of 3+ 3+3 + … , with all needed justifications.


That is, see if you can mimic what I did for 6 + 6 + 6 + … .

2. Suppose we look at n + n + n + … for positive integers n.


When n = 6, the value was an integer. What is the next value of n for
which the limit is an integer?

3. Show that Newton’s method works to find 5 , starting with x0 = 3.


That is, mimic what I did for 2 to use the Bounded convergence
theorem for 5 , instead.

4. Still Newton’s method for 5 . Estimate how close x6 is to 5 when


x0 = 3. That is, how many decimal digits do you think will be accurate?
Hint: If you know how close, say x2 is, then you should be able to use
this to estimate x3, and then x4, and so on.

3
5. Can you use Newton’s method to find 5?

6. Can you use Newton’s method to find the zero of x3 ! 3x ! 3?

You might also like