Sec. 3.4

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

3.

4 Hermite Interpolation
Given a differentiable function 𝒇(𝒙) on [𝒂, 𝒃] and a set of
𝒏 + 𝟏 distinct values {𝒙𝟎 , 𝒙𝟏 , 𝒙𝟐 , … , 𝒙𝒏 } ∈ [𝒂, 𝒃]. Then there
exist a polynomial 𝑯(𝒙) of degree at most 𝟐𝒏 + 𝟏 such that
𝒇(𝒙𝒊 ) = 𝑯(𝒙𝒊 ) & 𝑓′(𝒙𝒊 ) = 𝑯′(𝒙𝒊 ) for 𝒊 = 𝟎, 𝟏, 𝟐, … , 𝒏.
𝟐𝒏+𝟏

𝑯𝟐𝒏+𝟏 (𝒙) = 𝒇𝟎 + � 𝒇𝟎𝟏𝟐…𝒌 (𝒙 − 𝒛𝟎 ) (𝒙 − 𝒛𝟏 ) … (𝒙 − 𝒛𝒌−𝟏 )


𝒌=𝟏

where 𝒛𝟐𝒌 = 𝒛𝟐𝒌+𝟏 = 𝒙𝒌 & 𝒇𝟐𝒌,𝟐𝒌+𝟏 = 𝒇′ (𝒙𝒌 ).


𝒇(𝟐𝒏+𝟐) (𝝃(𝒙))
The error is given by 𝑬𝟐𝒏+𝟏 (𝒙) ≅ ∏𝟐𝒏+𝟏
𝒊=𝟎 |𝒙 − 𝒙𝒊 | ,
(𝟐𝒏+𝟐)!
where 𝝃(𝒙) lies in [𝒂, 𝒃].
Divided Differences Table for Hermite

𝒛𝒊 𝒇𝒊 1st order div


𝒛𝟎 = 𝒙𝟎 𝒇𝟎
𝒛𝟏 = 𝒙𝟎 𝒇𝟏 𝒇𝟎𝟏 = 𝒇′ (𝒙𝟎 )
𝒛𝟐 = 𝒙𝟏 𝒇𝟐 𝒇𝟐 − 𝒇𝟏 𝒇𝟎𝟏𝟐
𝒇𝟏𝟐 =
𝒙𝟐 − 𝒙𝟏
𝒛𝟑 = 𝒙𝟏 𝒇𝟑 𝒇𝟐𝟑 = 𝒇′ (𝒙𝟏 ) 𝒇𝟏𝟐𝟑 𝒇𝟎𝟏𝟐𝟑
𝒛𝟒 = 𝒙𝟐 𝒇𝟒 𝒇𝟒 − 𝒇𝟑 𝒇𝟎𝟏𝟐 𝒇𝟏𝟐𝟑𝟒 𝒇𝟎𝟏𝟐𝟑𝟒
𝒇𝟑𝟒 =
𝒙𝟒 − 𝒙𝟑
𝒛𝟓 = 𝒙𝟐 𝒇𝟓 𝒇𝟒𝟓 = 𝒇′ (𝒙𝟐 ) 𝒇𝟏𝟐𝟑 𝒇𝟐𝟑𝟒𝟓 𝒇𝟏𝟐𝟑𝟒𝟓 𝒇𝟎𝟏𝟐𝟑𝟒𝟓

𝑯𝟓 (𝒙) = 𝒇𝟎 + 𝒇𝟎𝟏 (𝒙 − 𝒙𝟎 ) + 𝒇𝟎𝟏𝟐 (𝒙 − 𝒙𝟎 )𝟐 + …

+𝒇𝟎𝟏𝟐𝟑𝟒𝟓 (𝒙 − 𝒙𝟎 )𝟐 (𝒙 − 𝒙𝟏 )𝟐 (𝒙 − 𝒙𝟐 )

1
Example 1 Find the Hermite polynomial that interpolate
𝟏
𝒇(𝒙) = at {𝒙𝟎 = 𝟏, 𝒙𝟏 = 𝟐, 𝒙𝟐 = 𝟒}. Compute 𝑯𝟓 (𝟎. 𝟓), 𝑯𝟓 (𝟑)
𝒙
and 𝑯𝟓 (𝟓. 𝟓) . Find the exact error and comment on the results.
−𝟏
𝒇′(𝒙) =
𝒙𝟐

𝒛𝒊 𝒇𝒊 1st order div


𝟏 𝟏
𝟏 𝟏 𝒇′ (𝟏) = −𝟏
𝟐 𝟎. 𝟓 𝒇𝟏𝟐 = −𝟎. 𝟓
𝟐 𝟎. 𝟓 𝒇′ (𝟐) = −𝟎. 𝟐𝟓
𝟒 𝟎. 𝟐𝟓 𝒇𝟑𝟒 = −𝟎. 𝟏𝟐𝟓
𝟒 𝟎. 𝟐𝟓 𝒇′ (𝟒) = −𝟎. 𝟎𝟔𝟐𝟓

2
Example 2 Use 5-digit rounding arithmetic to find the Hermite
polynomial for the following data:

𝒙 𝒇(𝒙) 𝒇′ (𝒙)
0.8 0.22363 2.1692
1.0 0.65809 2.0467

You might also like