{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

# set5 - \$ Set 5 Piecewise Polynomial Interpolation Kyle A...

This preview shows pages 1–11. Sign up to view the full content.

a39 a38 a36 a37 Set 5: Piecewise Polynomial Interpolation Kyle A. Gallivan Department of Mathematics Florida State University Foundations of Computational Math 2 Spring 2011 1

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
a39 a38 a36 a37 Summary Approximation of f ( x ) ∈ C (0) with polynomials. Various metrics possible: i | f ( x i ) p ( x i ) | i | f ( x i ) p ( x i ) | + · · · + | f ( k ) ( x i ) p ( k ) ( x i ) | bardbl f p bardbl bardbl f p bardbl L 2 2
a39 a38 a36 a37 Summary Interpolation f ( x i ) = p ( x i ) f ( x i ) = p ( x i ) , ..., f ( k ) ( x i ) = p ( k ) ( x i ) more general combinations of function values and derivatives Various interpolation forms of unique polynomials Lagrange – standard or barycentric Newton Hermite-Birkoff • bardbl f p bardbl 0 : convergent sequence of polynomial family representations Bernstein polynomials for f ∈ C (0) interpolatory strategies for more constrained class of f 3

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
a39 a38 a36 a37 Polynomial Interpolation Problems: Pointwise error too large at important points • bardbl f p bardbl too large on interval of interest erratic variation, i.e., not smooth enough excessive computational complexity ill-conditioning and instability 4
a39 a38 a36 a37 Polynomial Interpolation Solutions – Complications: choose better points – may not be possible increase n – may or may not improve error, may not converge interpolate derivatives – values may not be available 5

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
a39 a38 a36 a37 Piecewise Lagrange Interpolation Use local interpolants of lower order rather than one global polynomial. a = x 0 <x 1 < · · · <x n = b [ a,b ] = s I s : union of disjoint subintervals (intersect only at subset of grid points) g k ( x ) , on I s = [ x i s ,x i s + k ] is in P k g k ( x ) is a piecewise polynomial local interpolant p k,i s ( x j ) = f ( x j ) , i s j i s + k global interpolant g k ( x i ) = f ( x i ) , 0 i n 6
a39 a38 a36 a37 Choices Form of p k,i ( x ) In practice, each interval is independent in construction and evaluation. For analysis the form matters, e.g., basis choice When used to define a set of relationships between unknown f ( x i ) ,...,f ( k ) ( x i ) the form determines the structure of equations to be solved. 7

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
a39 a38 a36 a37 Forms and Bases monomial p k,i s ( x ) = α ( i s ) 0 + α ( i s ) 1 x + · · · + α ( i s ) k - 1 x k - 1 + α ( i s ) k x k Newton p k,i s ( x ) = f i s + f [ x i s ,x i s +1 ]( x - x i s ) + · · · + f [ x i s ,...,x i s + k ] ω ( i s ) k Lagrange p k,i s ( x ) = k X j =0 ( i s ) j ( x ) f i s + j basis form for analysis and implicit equations g k ( x ) = n X i =0 f i φ i ( x ) = n X i =0 γ i ψ i ( x ) where φ i ( x ) and ψ i ( x ) are piecewise polynomials. 8
a39 a38 a36 a37 Error If f ∈ C ( k +1) [ a,b ] a x b, f ( x ) g k ( x ) = f ( x ) p k,i s ( x ) = f ( k +1) ( ξ ) ( k + 1)! ω ( i s ) k +1 ( x ) x [ x i s ,x i s + k ] The local error expressions can be combined to get a global error bardbl f g k bardbl Ch k +1 bardbl f ( k +1) bardbl where h is maximum size of intervals I i 9

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
a39 a38 a36 a37 Error This is easily shown: | f ( k +1) ( ξ ) ( k + 1)!
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}