Convergence: nth-Term Test, Comparing Nonegative Series

Download Report

Transcript Convergence: nth-Term Test, Comparing Nonegative Series

Convergence: nth-Term Test,
Comparing Non-negative Series,
Ratio Test
Power Series and Convergence
We have written statements like:
ln 1 + ๐‘ฅ = ๐‘ฅ
๐‘ฅ2
โˆ’
2
๐‘ฅ3
+
3
โˆ’ โ‹ฏ + โˆ’1
๐‘›
๐‘›โˆ’1 ๐‘ฅ
๐‘›
+โ‹ฏ
But we have not talked in depth about what values of ๐‘ฅ
make the identity true.
Example: Investigate whether or not ๐‘ฅ = 0.5 makes the
sentence above true? ๐‘ฅ = 100?
.52 .53
+
โˆ’ โ‹ฏ + โˆ’1
๐‘ฅ = 0.5: ln 1 + .5 = .5 โˆ’
2
3
๐‘›
.5
๐‘›โˆ’1
+โ‹ฏ
๐‘›
1002 1003
+
โˆ’ โ‹ฏ + โˆ’1
๐‘ฅ = 100: ln 1 + 100 = 100 โˆ’
2
3
Appears to
Converge.
๐‘›
100
Appears to
๐‘›โˆ’1
+ โ‹ฏ Diverge.
๐‘›
We need better ways to determine the values of ๐’™ that make the series converge.
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
Investigate the partial sums of the series and compare the
๐Ÿ
results to ๐’š =
on a graph.
๐Ÿโˆ’๐’™
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
1
Consider the Maclaurin Series for 1โˆ’๐‘ฅ
:
1 + ๐‘ฅ + ๐‘ฅ 2 + ๐‘ฅ 3 + โ‹ฏ + ๐‘ฅ ๐‘›โˆ’1 + โ‹ฏ
๐Ÿ
On (โˆ’๐Ÿ, ๐Ÿ), {๐’š๐Ÿ , ๐’š๐Ÿ‘ , ๐’š๐Ÿ’ , ๐’š๐Ÿ“ , ๐’š๐Ÿ” ,โ€ฆ} converges to ๐’š๐Ÿ =
. The
๐Ÿโˆ’๐’™
polynomial series is a good approximation of ๐’š๐Ÿ on (โˆ’๐Ÿ, ๐Ÿ).
๐’š๐Ÿ
๐’š๐Ÿ
๐’š๐Ÿ‘
๐’š๐Ÿ’
๐’š๐Ÿ“
๐’š๐Ÿ”
๐Ÿ
=
๐Ÿโˆ’๐’™
=๐Ÿ
=๐Ÿ+๐’™
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘
= ๐Ÿ + ๐’™ + ๐’™๐Ÿ + ๐’™๐Ÿ‘ + ๐’™๐Ÿ’
Window: โˆ’2 โ‰ค ๐‘ฅ โ‰ค 2 and โˆ’10 โ‰ค ๐‘ฆ โ‰ค 10
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
Investigate the partial sums of the series and compare the
results to ๐’š = sin ๐‘ฅ on a graph.
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
When is a Taylor Series a Good
Approximation?
Consider the Maclaurin Series for sin ๐‘ฅ:
2๐‘›+1
๐‘ฅ
3
5
๐‘ฅ โˆ’ ๐‘ฅ3! + ๐‘ฅ5! + โ‹ฏ + โˆ’1 ๐‘›
2๐‘› + 1 !
On (โˆ’โˆž, โˆž), {๐’š๐Ÿ , ๐’š๐Ÿ‘ , ๐’š๐Ÿ’ , ๐’š๐Ÿ“ , ๐’š๐Ÿ” ,โ€ฆ} converges to ๐’š๐Ÿ = sin ๐’™. The
polynomial series is a good approximation of ๐’š๐Ÿ on โˆ’โˆž, โˆž .
๐’š๐Ÿ = sin ๐’™
๐’š๐Ÿ = ๐’™
๐‘ฅ3
๐’š๐Ÿ‘ = ๐‘ฅ โˆ’ 3!
๐’š๐Ÿ’ = ๐‘ฅ โˆ’
๐‘ฅ3
๐‘ฅ5
3! + 5!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐’š๐Ÿ“ = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7!
๐‘ฅ3
๐‘ฅ5
๐‘ฅ7
๐‘ฅ9
๐’š๐Ÿ” = ๐‘ฅ โˆ’ 3! + 5! โˆ’ 7! + 9!
Window: โˆ’7 โ‰ค ๐‘ฅ โ‰ค 7 and โˆ’5 โ‰ค ๐‘ฆ โ‰ค 5
Proving Sine Converges
Prove the Maclaurin series
โˆž
๐‘˜=0
โˆ’1
Investigate the error function:
sin ๐‘ฅ for all real ๐‘ฅ.
๐‘ ๐‘ฅ โˆ’ 0 ๐‘›+1
๐‘…๐‘› ๐‘ฅ =
๐‘›+1 !
๐‘” ๐‘›+1 ๐‘ ๐‘ฅ โˆ’ 0 ๐‘›+1
๐‘…๐‘› ๐‘ฅ =
๐‘›+1 !
๐‘”
๐‘›+1
๐‘ ๐‘ฅ ๐‘›+1
=
๐‘›+1 !
1 โˆ™ ๐‘ฅ ๐‘›+1
๐‘ฅ ๐‘›+1
โ‰ค
โ‰ค
๐‘›+1 !
๐‘›+1 !
๐‘”
๐‘›+1
2๐‘›+1
๐‘› ๐‘ฅ
2๐‘›+1 !
converges to
Investigate the limit of the
last statement
๐‘ฅ ๐‘›+1
=0
lim
๐‘›โ†’โˆž ๐‘› + 1 !
This means that ๐‘…๐‘› ๐‘ฅ โ†’ 0
for all ๐‘ฅ.
Therefore the Maclaurin
series converges for all ๐‘ฅ.
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
๐‘ฅโ‰ 0:
๐‘ฅ=0
Investigate the partial sums of the series and compare the
results to ๐’š =
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
๐‘ฅ โ‰  0 on a graph.
๐‘ฅ=0
When is a Taylor Series a Good
Approximation?
โˆ’1 ๐‘ฅ 2
Consider the Maclaurin Series for ๐‘’
0
0
On ๐ŸŽ , {๐’š๐Ÿ , ๐’š๐Ÿ‘ , ๐’š๐Ÿ’ , ๐’š๐Ÿ“ , ๐’š๐Ÿ” ,โ€ฆ} converges to ๐’š๐Ÿ =
๐‘ฅโ‰ 0:
๐‘ฅ=0
2
โˆ’1
๐‘ฅ
๐‘’
0
The series is a good approximation of ๐’š๐Ÿ on ๐ŸŽ .
๐’š๐Ÿ =
2
โˆ’1
๐‘ฅ
๐‘’
0
๐’š๐Ÿ = ๐ŸŽ
๐’š๐Ÿ‘ = ๐ŸŽ
๐’š๐Ÿ’ = ๐ŸŽ
๐’š๐Ÿ“ = ๐ŸŽ
๐’š๐Ÿ” = ๐ŸŽ
๐‘ฅโ‰ 0
๐‘ฅ=0
๐‘ฅ โ‰  0.
๐‘ฅ=0
The Convergence Theorem for Power
Series
There are three possibilities for
respect to convergence:
โˆž
๐‘›=0 ๐‘๐‘›
๐‘ฅโˆ’๐‘Ž
๐‘›
with
Radius of Convergence
1. There is a positive number ๐‘… such that the series
diverges for ๐‘ฅ โˆ’ ๐‘Ž > ๐‘… but converges for ๐‘ฅ โˆ’ ๐‘Ž <
๐‘…. The series may or may not converge at either of
the endpoints ๐‘ฅ = ๐‘Ž โˆ’ ๐‘… and ๐‘ฅ = ๐‘Ž + ๐‘….
2. The series converges for every ๐‘ฅ (๐‘… = โˆž).
3. The series converges at ๐‘ฅ = ๐‘Ž and diverges
elsewhere (๐‘… = 0).
The set of all values of ๐’™ for which the series converges is the
interval of convergence.
The n-th Term Test
If lim ๐‘Ž๐‘› โ‰  0, then the infinite series
๐‘›โ†’โˆž
โˆž
๐‘›=1 ๐‘Ž๐‘›
diverges.
OR
When
determining if a
series converges,
always use this
test first!
If the infinite series โˆž
๐‘›=1 ๐‘Ž๐‘› converges,
then lim ๐‘Ž๐‘› = 0.
๐‘›โ†’โˆž
The converse of this statement is NOT true.
If lim ๐‘Ž๐‘› = 0, the infinite series
โˆž
๐‘›=1 ๐‘Ž๐‘›
does not necessarily converge.
Examples
Use the n-th Term Test to investigate the convergence of
the series.
1.
โˆž 2๐‘›+1
๐‘›=1 ๐‘›
Since the limit is not 0,
๐Ÿ๐’ + ๐Ÿ
=๐Ÿ
lim
the series must diverge.
๐’โ†’โˆž
๐’
2.
5
โˆž
๐‘›=1 10๐‘›
๐Ÿ“
=๐ŸŽ
lim
๐’
๐’โ†’โˆž ๐Ÿ๐ŸŽ
Since the limit is 0, the
n-th Term Test is
inconclusive.
Just because the n-th term goes to zero does not mean the
series necessarily converges. We need more tests to
determine if a series converges.
The Finney Procedure for Determining
Convergence
nth-Term Test
Is lim ๐‘Ž๐‘› = 0?
๐‘›โ†’โˆž
Yes or Maybe
?
No
The series
diverges.
Convergent Geometric Series
๐‘›โˆ’1
The geometric series โˆž
๐‘Ž๐‘Ÿ
converges
๐‘›=1
if and only if ๐‘Ÿ < 1. If the series converges,
๐‘Ž
its sum is
.
1โˆ’๐‘Ÿ
Where a is the first term and r
is the constant ratio.
Example: Determine if
๐‘’ ๐‘›
โˆž
(
๐‘›=1 ๐œ‹)
converges.
๐’†
๐…
This is a geometric series with ๐’“ = .
Since
๐’†
๐…
< ๐Ÿ, the series converges.
This explanation would warrant full credit on the AP Test. Statements like โ€œI
know this is a __ series with __, so it diverges/convergesโ€ are acceptable.
The Finney Procedure for Determining
Convergence
nth-Term Test
Is lim ๐‘Ž๐‘› = 0?
The series
diverges.
No
๐‘›โ†’โˆž
Yes or Maybe
Geometric Series Test
Is ๐‘Ž๐‘› = ๐‘Ž + ๐‘Ž๐‘Ÿ + ๐‘Ž๐‘Ÿ 2 + โ‹ฏ ?
No
?
Yes
Converges to
๐‘Ž
if ๐‘Ÿ < 1.
1โˆ’๐‘Ÿ
Diverges if
๐‘Ÿ > 1.
Example
๐‘›! 2
โˆž
๐‘›=0 2๐‘› !
Investigate the series
to see if it
diverges.
Notice the series is NOT geometric.
โˆž
๐‘›=0
๐‘›! 2
1 1 1
1
1
=1+ + +
+
+
+โ‹ฏ
2๐‘› !
2 6 20 70 252
Using partial sums, it appears the series converges:
1; 1.5; 1.667; 1.717; 1.721
Similar to a Geometric Series, we can investigate the ratio between terms.
a1
a0
๏€ฝ
1 2
a3
๏€ฝ 0.5;
1
a2
a1
๏€ฝ
1 6
1 2
a2
๏€ฝ 0.33;
๏€ฝ
1 20
a5
๏€ฝ 0.3;
1 6
a4
a3
๏€ฝ
1 70
a4
๏€ฝ
1 252
๏€ฝ 0.28
1 70
๏€ฝ 0.29;
1 20
Similar to a converging Geometric Series, the limit of the โ€œcommonโ€ ratio appears to be
less than 1. According to the next theorem. This means the series converges.
The Ratio Test
Suppose the limit
๐‘Ž๐‘›+1
lim
๐‘›โ†’โˆž ๐‘Ž๐‘›
= ๐ฟ either exists
or is infinite. Then:
1. If ๐ฟ < 1, the series
โˆž
๐‘›=1 ๐‘Ž๐‘› converges.
Each successive terms are getting smaller.
โˆž
๐‘›=1 ๐‘Ž๐‘› diverges.
Each successive terms are getting larger.
2. If ๐ฟ > 1, the series
3. If ๐ฟ = 1, the test is inconclusive.
This is the โ€œcousinโ€ test to determining if a Convergent Geometric Series
test. It is identical except: (1) If the โ€œcommonโ€ ratio is 1, then test fails. (2)
The test does not determine the value the series converges to.
Example 1
Investigate the series
๐‘›! 2
โˆž
๐‘›=0 2๐‘› !
to see if it diverges.
Similar to a Geometric Series, we can investigate the ratio between terms.
๏€จ ๏€จ n ๏€ซ 1 ๏€ฉ !๏€ฉ
๏€จ 2 ๏€จ n ๏€ซ1๏€ฉ ๏€ฉ!
2
lim
n๏‚ฎ ๏‚ฅ
a n ๏€ซ1
๏€ฝ lim
an
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
๏€จ n !๏€ฉ
๏€จ 2 n ๏€ฉ!
2
n๏‚ฎ ๏‚ฅ
๏€จ n ๏€ซ1๏€ฉ
2
๏€จ 2 n ๏€ซ 2 ๏€ฉ๏€จ 2 n ๏€ซ 1 ๏€ฉ
๏€จ ๏€จ n ๏€ซ 1 ๏€ฉ !๏€ฉ ๏€จ 2 n ๏€ฉ !
2
๏€ฝ lim
๏€จ n !๏€ฉ ๏€จ 2 n ๏€ซ 2 ๏€ฉ !
2
n๏‚ฎ ๏‚ฅ
๏€ฝ lim
2
n ๏€ซ 2 n ๏€ซ1
2
n๏‚ฎ ๏‚ฅ 4n ๏€ซ6n๏€ซ2
๏€ฝ
1
4
Similar to a converging Geometric Series, this โ€œcommonโ€ ratio is less than 1.
According to the Ratio Test, this means the series converges.
Notice: Unlike the last
example, this series
depends on a value of x.
Example 2
In #3 on Taylor Series Challenges we worked with the series
1
โˆž
2๐‘› . Find the radius of convergence.
๐‘ฅ
โˆ’
2
๐‘›=1 32๐‘› 2๐‘›
If the series
1
lim
n๏‚ฎ ๏‚ฅ
a n ๏€ซ1
an
๏€ฝ
3
lim
2 ๏€จ n ๏€ซ1๏€ฉ
๏€จ 2 ๏€จ n ๏€ซ 1๏€ฉ ๏€ฉ
1
n๏‚ฎ ๏‚ฅ
3
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
2n
๏€จ 2n ๏€ฉ
2n๏€ซ2
๏€จ x ๏€ญ 2๏€ฉ
๏€จ 2n ๏€ฉ ๏€จ x ๏€ญ 2 ๏€ฉ
2n
2n๏€ซ2
3
๏€จ 2 ๏€จ n ๏€ซ 1๏€ฉ ๏€ฉ ๏€จ x ๏€ญ 2 ๏€ฉ
3
2n
๏€จ x ๏€ญ 2๏€ฉ
2n
2 ๏€จ n ๏€ซ1๏€ฉ
converges, the
ratio is less
than 1.
๏€จ x๏€ญ2๏€ฉ
2
9
๏€จ x ๏€ญ 2๏€ฉ
2
๏€ผ1
๏€ผ9
x๏€ญ2 ๏€ผ3
The interval of
convergence is
The limit depends on n. Separate the nโ€™s.
2
2
2 centered at 2 with
n ๏€จ x ๏€ญ 2๏€ฉ
๏€จ x ๏€ญ 2๏€ฉ
n
๏€จ x ๏€ญ 2๏€ฉ
๏ƒ—
๏€ฝ 1๏ƒ—
a radius
lim 2
๏€ฝ lim
2
n๏‚ฎ ๏‚ฅ 3
3
9
๏€จ n ๏€ซ 1๏€ฉ n ๏‚ฎ ๏‚ฅ n ๏€ซ 1
convergence of 3.
Example 3
Find the radius of convergence for the Maclaurin series for
2
๐‘“ ๐‘ฅ = ๐‘’๐‘ฅ .
โˆž
2
3
๐‘›
๐‘ฅ
๐‘ฅ
๐‘ฅ
We know: ๐‘’ ๐‘ฅ = 1 + ๐‘ฅ + + + โ‹ฏ + + โ‹ฏ =
2! 3!
๐‘›!
Thus: ๐‘’
๐‘ฅ2
4
6
2๐‘›
๐‘ฅ
๐‘ฅ
๐‘ฅ
= 1 + ๐‘ฅ2 + + + โ‹ฏ +
+โ‹ฏ=
2! 3!
๐‘›!
Investigate
the ratio:
lim
n๏‚ฎ ๏‚ฅ
a n ๏€ซ1
an
๐‘›=0
โˆž
x
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
๐‘ฅ๐‘›
๐‘›!
๐‘›=0
๐‘ฅ 2๐‘›
๐‘›!
2 ๏€จ n ๏€ซ1๏€ฉ
๏€จ n ๏€ซ 1๏€ฉ !
x
2n
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
x
2n๏€ซ2
n!
๏€จ n ๏€ซ 1๏€ฉ ! x
2n
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
n!
The series has an infinite radius.
x
2
n ๏€ซ1
๏€ฝ 0
Example 4
Find the radius of convergence for the series
๏€จ ๏€ญ 1๏€ฉ
a n ๏€ซ1
Investigate
lim
๏€ฝ lim
the ratio: n ๏‚ฎ ๏‚ฅ a
n๏‚ฎ ๏‚ฅ
n
๏€จ n ๏€ซ 1๏€ฉ ๏€ญ1
5
x
n ๏€ซ1
n ๏€ซ1
๏€จ ๏€ญ 1๏€ฉ
n ๏€ญ1
5
x
โˆ’1 ๐‘›โˆ’1 ๐‘ฅ ๐‘›
โˆž
.
๐‘›=1
๐‘›
5
n
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
5
5
n
๏€จ ๏€ญ 1๏€ฉ x
n ๏€ซ1
n
๏€จ ๏€ญ 1๏€ฉ
n ๏€ซ1
n ๏€ญ1
n
Since the ratio tests uses
n
n ๏€ซ1
5 x
x
x
an absolute value, the
๏€ฝ lim
๏€ฝ
powers on -1 do not ๏€ฝ lim
n ๏€ซ1 n
n๏‚ฎ ๏‚ฅ 5
n๏‚ฎ ๏‚ฅ 5
x
5
affect the limit.
If the series
converges, the
ratio is less
than 1.
x
5
๏€ผ1
x ๏€ผ5
The interval of convergence is
centered at 0 with a radius
convergence of 5.
x
n
White Board Challenge
Use the Ratio Test to determine the radius of
โˆž ๐‘›+1
convergence for ๐‘›=1 ๐‘› ๐‘ฅ.
lim
n๏‚ฎ ๏‚ฅ
a n ๏€ซ1
an
๏€จ n ๏€ซ1๏€ฉ ๏€ซ1
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
n ๏€ซ1
n ๏€ซ1
n
x
x
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
n๏€จn๏€ซ 2๏€ฉ
n ๏€จ n ๏€ซ1๏€ฉ
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
n๏€ซ2
n ๏€ซ1
๏€ฝ 1
The Ratio Test is inconclusive because the
limit is 1. We need to use other tests to
determine if the series converges.
The Direct Comparison Test
Suppose 0 โ‰ค ๐‘Ž๐‘› โ‰ค ๐‘๐‘› for all ๐‘› โ‰ฅ ๐‘.
Both series must have only positive terms.
โ€ข If
โ€ข If
โˆž
๐‘›=1 ๐‘๐‘›
โˆž
๐‘›=1 ๐‘Ž๐‘›
converges, then โˆž
๐‘›=1 ๐‘Ž๐‘› converges.
diverges, then โˆž
๐‘›=1 ๐‘๐‘› diverges.
Example: Prove
For all ๐‘›: 0 โ‰ค
2๐‘›
๐‘ฅ 2๐‘›
๐‘›! 2
๐‘ฅ 2๐‘›
โˆž
๐‘›=0 ๐‘›! 2
โ‰ค
๐‘ฅ 2๐‘›
๐‘›!
converges for all real ๐‘ฅ.
๐‘ฅ 2๐‘›
โˆž
๐‘›=0 ๐‘›! is the Taylor
Series for ๐‘“ ๐‘ฅ =
โˆž ๐‘ฅ
Since ๐‘›=0
converges for all real ๐’™,
๐‘›!
2
๐‘ฅ
๐‘’ .
๐‘ฅ 2๐‘›
โˆž
๐‘›=0 ๐‘›! 2 converges
all real ๐’™ by the Direct Comparison Test.
for
What about Negative Terms?
Does the Direct Comparison Test fail if there are negative terms?
Consider the
sin ๐‘ฅ ๐‘›
โˆž
series ๐‘›=0
๐‘›!
โˆž
๐‘›
sin 3๐œ‹
2
๐‘›=0
๐‘›!
if ๐‘ฅ = 3๐œ‹
:
2
1
1
= 1 โˆ’ 1 + 12 โˆ’ 16 + 24
โˆ’ 120
โ€ฆ
Using partial sums, it appears the series converges:
1; 0; 0.5; 0.333; 0.375; 0.367
But we donโ€™t have a test to prove the series converges.
Similar to the Ratio Test, what happens when we look at the
absolute value of each term:
โˆž
๐‘›
sin 3๐œ‹
2
1
1
1
= 1 + 1 + 12 + 16 + 24
+ 120
โ€ฆ + ๐‘›!
๐‘›!
๐‘›=0
Using partial sums, this series ALSO appears to converges:
1; 2; 2.5; 2.667; 2.708; 2.717
Since the new series is less than the original, if we can prove the new
series converges, then the original series must converge.
Absolute Convergence Test
If
โˆž
๐‘›=1 ๐‘Ž๐‘› converges, then
โˆž
๐‘›=1 ๐‘Ž๐‘› converges.
Unlike the Comparison Test, this test does not require the terms
to be positive.
Example: Prove
Investigate
sin ๐‘ฅ ๐‘›
๐‘›!
sin ๐‘ฅ ๐‘›
โˆž
๐‘›=0
๐‘›!
Notice:
sin ๐‘ฅ ๐‘›
1
โ‰ค
๐‘›!
๐‘›!
converges for all ๐‘ฅ.
โˆž 1
๐‘›=0 ๐‘›!
lim
n๏‚ฎ ๏‚ฅ
a n ๏€ซ1
an
๏€ฝ
converges by the Ratio Test:
1
lim
n๏‚ฎ ๏‚ฅ
๏€จ n ๏€ซ 1๏€ฉ !
1
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
n!
๏€จ n ๏€ซ 1๏€ฉ !
๏€ฝ lim
n๏‚ฎ ๏‚ฅ
1
n ๏€ซ1
n!
sin ๐‘ฅ ๐‘›
Since
converges for
๐‘›!
sin ๐‘ฅ ๐‘›
โˆž
converges for all
๐‘›=0
๐‘›!
all real ๐‘ฅ by the Direct Comparison Test,
real ๐‘ฅ by the Absolute Convergence Test.
๏€ฝ
0
Definition of Absolute Convergence
If โˆž
๐‘›=1 ๐‘Ž๐‘› converges, then
absolutely.
โˆž
๐‘›=1 ๐‘Ž๐‘›
converges
Example: The Alternating Harmonic Series
converges but it does not converge absolutely:
โˆž
(โˆ’1)๐‘›โˆ’1
1
๐‘›
= 1 โˆ’ 12 + 13 โˆ’ 14 + โ‹ฏ = ln 2
๐‘›=1
โˆž
(โˆ’1)๐‘›โˆ’1
๐‘›=1
โˆž
1
๐‘›
1
๐‘›
=
๐‘›=1
= 1 + 12 + 13 + 14 + โ‹ฏ
This is the divergent
Harmonic Series.
Note about Absolute Convergence
If โˆž
๐‘›=1 ๐‘Ž๐‘› converges, then
absolutely.
โˆž
๐‘›=1 ๐‘Ž๐‘›
converges
๐‘Ž๐‘›+1
๐‘Ž๐‘›
The Ratio Test uses an absolute value:
.
Thus, every test that converges by the Ratio
Test also converges absolutely:
โˆž
๐‘›โˆ’1
4(โˆ’0.5)
๐‘›=1
โˆž
๐‘›โˆ’1
๐‘›=1 4(0.5)
converges absolutely because
converges by by the ratio test
The Ratio Test is incredibly strong