Calc BC 10.9 Error Estimates

Section 10.
9
Error Estimates
The Problems
• Say
we want to calculate the value of a power
series for a specified x-value
• We have to use a finite sum to approximate
the value, since , where denotes the sum .
• This problem really has two parts
– How close can we get with minimal effort?
– How many terms are needed to get a good
estimate?
Error Estimates
• There are two cases where we can get useful
error estimates:
– Alternating series
– Taylor series
Alternating Series Estimates
• We saw before, but now it matters . . .

If is a convergent alternating series, then
• We can use this to estimate the sum to within

a specified error tolerance
– find a value n for which . Then
– will provide the “good” estimate we want
Example

• This
series starts out as and it converges AST.
Calculate the sum of the series to the nearest .
0001.
• Since , once we find n with , will be close
enough
– We need n for which
– We need , so
– In other words, if we add at least 99 terms we will
be within .0001 of the sum of the (infinite) series
Example: Calculating

• We
will see later that is the Taylor series for ,
centered at 0
• As a result, How many terms do we need to use
this series to estimate to the nearest 0.001?
– By the AST error condition, we need to find n for
which , which is , which simplifies to
• This will take a while!

Turn the last example around
• Again, use the result
• If we only have time to add up 20 terms, what

is the maximum error we could be making in
estimating ?
– We know that , which is
– For , this error is
– Not a very good estimate!
Taylor Series Estimates

When we derived the Taylor and MacLaurin
series for a function, I showed you an
expression for the difference

LaGrange Remainder Theorem
where is some unknowable number

between
Estimating
(1)
• Remember the TS:
• We want to estimate . If we use to estimate,

what is the max error possible?
• The error
Estimating
(2)
• The
error
• Since , we max this factor by choosing .
– This gives - a snake chasing its tail!
• Time for some really gross estimates
– Since and ,
• Our estimate is correct to 4 decimal places!

Use AST when you can!
• We
know that .
• What is the max error if we use , centered at
0, to estimate the value?
–
– The Lagrange error estimate, would be , which is
• But since this is an alternating series we can
use the AST bound instead!
11− 𝑐𝑜𝑠𝑥
Ch 10.9 #20 approximate to 𝑑𝑥
4 decimal places
∫ 𝑥2
𝑟 ( 𝑥 ) < .00005
𝑛
0
𝑥2 𝑥 4 𝑥6
1 1−(1 − + − + ⋯)
2! 4 ! 6 !
∫ 2
𝑑𝑥
0 𝑥
𝑥2 𝑥4 𝑥6 𝑥8
− + − +⋯ 1 2 4 6
1
1 𝑥 𝑥 𝑥
∫
0
2! 4! 6! 8!
𝑥
2
𝑑𝑥 ¿ ∫
0

( − +
2! 4! 6! 8!
− )
+ ⋯ 𝑑𝑥
1 𝑥
3
𝑥
5
𝑥
7 1 1 1 1
( 2!
𝑥− + −
3 ∙ 4 ! 5 ∙6 ! 7 ∙ 8 ! )
+⋯ ¿ 2 − 3∙ 4 ! + 5 ∙ 6 ! − 7 ∙ 8 ! +⋯
0
How many terms do we need?
1
When is <.00005 ?
( 2 𝑛 −1 ) ( 2 𝑛 ) !
1
We want: <.00005
( 2𝑛 −1 ) ( 2 𝑛 ) !
1
𝑛
( 2𝑛 −1 ) ( 2 𝑛 ) !

𝑛=4
is sufficient , therefore we only need 3terms .
1 1 1
S≈ − +
2 3∙ 4 ! 5 ∙ 6 ! ≈ .4863888889 ≈ .486 4

No guaranteed accuracy past the 4th decimal place...

Ch 10.9 #26 Use a power series 𝑒 𝑥 −𝑒 − 𝑥 − 2 𝑥
lim
to evaluate the limit 𝑥→ 0 𝑥 − 𝑎𝑟𝑐𝑡𝑎𝑛𝑥
𝑥2 𝑥3 𝑥4 𝑥2 𝑥3 𝑥4
(1+ 𝑥 + + + +⋯) − (1− 𝑥 + − + + ⋯)− 2 𝑥
2! 3 ! 4 ! 2! 3! 4!
¿ lim 3 5 7
𝑥→ 0 𝑥 𝑥 𝑥
𝑥 −( 𝑥 − + − + ⋯)
3 5 7
2 𝑥3 2 𝑥5 2 𝑥 3 2 𝑥5 2 𝑥 7
(2 𝑥 + + + ⋯) − 2 𝑥 + + +⋯
3! 5! 3! 5 ! 7 !
¿ lim 3 5 7
¿ lim 3 5 7
𝑥→ 0 𝑥 𝑥 𝑥 𝑥→ 0 𝑥 𝑥 𝑥
− + +⋯ − + +⋯
3 5 7 3 5 7
12 1 2 0 𝑥2 4 2 0 𝑥 4 12
+ + +⋯ 3!
3! 5! 7! =¿ 1
¿ lim 2 4
¿
6
𝑥→ 0 6 60 𝑥 210 𝑥
− + +⋯ 3
3 5 7
Ch 10.9 #32 Calculate , centered at , to 6 dec. places
𝜋
∞ 𝑓 (𝑛 )
3 ( ) 𝜋
𝑛 𝜋
58 °= −
2𝜋
𝑓 ( 𝑥 )= ∑
𝑛=0 𝑛! ( 𝑥−
3 ) 3 18 0

When we substitute
How many terms are needed

so that
√ 3 1 𝜋 √ 3 𝜋 2 1 𝜋 3
𝑓 𝑥=
( ) + 𝑥−
2
−
2 (
𝑥− −
3 ) 2∙ 2 ! ( 3 ) (
2 ∙ 3!
𝑥−
3 )+⋯
√ 3 1 𝜋 √ 3 𝜋 2 1 𝜋 3
𝑓 ( 𝑥 )= + 𝑥− −
2
𝑥−
2 (
−
3 ) 2∙ 2 ! ( 3 ) 2 ∙ 3!(𝑥−
3 )+⋯
( 𝑛 +1)
𝑓 ( 𝑧) 𝑛+1 Need 6 decimal place
𝑅𝑛 ( 𝑥 )= ( 𝑥 −𝑎 )
( 𝑛+1 ) ! accuracy
𝑛 +1
𝜋 𝜋 ± 𝑠𝑖𝑛𝑧 𝑜𝑟 𝑐𝑜𝑠𝑧 − 𝜋
𝑅𝑛 ( −
3 90
= ) ( 𝑛+1 ) ! 90 ( ) ¿ .0000005

𝑛+1
1 𝜋
𝑛 ( ) 1 𝜋 𝑛 +1

( 𝑛+ 1 ) ! 9 0 |𝑅 𝑛 ( 58 ° )|≤ (
𝑛+1 ) ! 9 0 ( )
√ 3 1 −𝜋 √ 3 −𝜋 2 1 −𝜋 3
𝑠𝑖𝑛 58 ° ≈
( )
2
+ ( )
−
2 90
−( )
2 ∙ 2! 90 ( )
2∙ 3 ! 90
already good
enough! ≈ .8480480428 ≈ .8 48048

≈ .84 80480962

Calc BC 10.9 Error Estimates

Uploaded by

Copyright:

Available Formats

Calc BC 10.9 Error Estimates

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calc BC 10.9 Error Estimates

Uploaded by

Copyright:

Available Formats

Section 10.

• We can use this to estimate the sum to within

• This will take a while!

• If we only have time to add up 20 terms, what

LaGrange Remainder Theorem

where is some unknowable number

• We want to estimate . If we use to estimate,

• Our estimate is correct to 4 decimal places!

No guaranteed accuracy past the 4th decimal place...

How many terms are needed

You might also like