Scribd
Upload
715 views6 pages

Newton's Divided Difference Formula

Newton's divided difference formula provides a method to approximate functions based on tabulated data points. It defines divided differences recursively to represent polynomials that interpolate the data points. The first divided difference approximates with a linear polynomial, the second with quadratic, and the kth divided difference gives a polynomial of degree k-1. Newton's formula expresses the value of an interpolated function at any point x as a polynomial in terms of the tabulated points and their divided differences. This allows approximating the function where no direct data is given.

Uploaded by

mengelhu
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
715 views6 pages

Newton's Divided Difference Formula

Newton's divided difference formula provides a method to approximate functions based on tabulated data points. It defines divided differences recursively to represent polynomials that interpolate the data points. The first divided difference approximates with a linear polynomial, the second with quadratic, and the kth divided difference gives a polynomial of degree k-1. Newton's formula expresses the value of an interpolated function at any point x as a polynomial in terms of the tabulated points and their divided differences. This allows approximating the function where no direct data is given.

Uploaded by

mengelhu
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/ 6

7/23/2015

Newton'sDividedDifferenceFormula

NEWTON'SDIVIDEDDIFFERENCEFORMULA
f(xi)f(xj)
Letusassumethatthefunctionf(x)islinearthenwehave
(xixj)
wherexiandxjareanytwotabularpoints,isindependentofxiandxj.Thisratioiscalledthe
firstdivideddifferenceoff(x)relativetoxiandxjandisdenotedbyf[xi,xj].Thatis
f[xi,xj]=

f(xi)f(xj)
(xixj)

=f[xj,xi]

Sincetheratioisindependentofxiandxjwecanwritef[x0,x]=f[x0,x1]
f(x)f(x0)
(xx0)

= f[x0,x1]

f(x)=f(x0)+(xx0)f[x0,x1]

1
=
xx0

f(x0)

x0x

f(x1)

x1x

f1f0
=

f0x1f1x0
x+

x1x0

x1x0

Soiff(x)isapproximatedwithalinearpolynomialthenthefunctionvalueatanypointxcanbe
calculatedbyusingf(x)P1(x)=f(x0)+(xx1)f[x0,x1]
wheref[x0,x1]isthefirstdivideddifferenceoffrelativetox0andx1.
Similarlyiff(x)isaseconddegreepolynomialthenthesecantslopedefinedaboveisnot
constantbutalinearfunctionofx.Hencewehave
f[x1,x2]f[x0,x1]
x2x0
isindependentofx0,x1andx2.Thisratioisdefinedasseconddivideddifferenceoffrelativeto
x0,x1andx2.Thesecinddivideddifferencearedenotedas
f[x1,x2]f[x0,x1]
f[x0,x1,x2] =

https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

x2x0
1/6

7/23/2015

Newton'sDividedDifferenceFormula

Nowagainsincef[x0,x1,x2]isindependentofx0,x1andx2wehave
f[x1,x0,x]=f[x0,x1,x2]
f[x0,x]f[x1,x0]

= f[x0,x1,x2]
xx1

f[x0,x]=f[x0,x1]+(xx1)f[x0,x1,x2]

f[x]f[x0]
xx0

= f[x0,x1]+(xx1)f[x0,x1,x2]

f(x)=f[x0]+(xx0)f[x0,x1]+(xx0)(xx1)f[x0,x1,x2]
Thisisequivalenttotheseconddegreepolynomialapproximationpassingthroughthreedata
points
x0
f0

x1
f1

x2
f2

Sowheneverf(x)isapproximatedwithaseconddegreepolynomial,thevalueoff(x)atany
pointxcanbecomputedusingtheabovepolynomial.
Inthesamewayifwedefinerecursivelykthdivideddifferencebytherelation
f[x1,x2,...,xk]f[x0,x1,...,xk1]
f[x0,x1,...,xk] =
xkx0

Thekthdegreepolynomialapproximationtof(x)canbewrittenas
f(x)=f[x0]+(xx0)f[x0,x1]+(xx0)(xx1)f[x0,x1,x2]
+...+(xx0)(xx1)...(xxk1)f[x0,x1,...,xk].
ThisformulaiscalledNewton'sDividedDifferenceFormula.Oncewehavethedivided
differencesofthefunctionfrelativetothetabularpointsthenwecanusetheaboveformulato
computef(x)atanynontabularpoint.
Computingdivideddifferencesusingdivideddifferencetable:Letusconsiderthepoints(x1,
f1),(x2,f2),(x3,f3)and(x4,f4)wherex1,x2,x3andx4arenotnecessarilyequidistantpoints
thenthedivideddifferencetablecanbewrittenas
xi fi

f[xi,xj]

f[xi,xj,xk]

f[xi,xj,xk,xl]

x1 f1
f2f1

f[x1,x2]=

https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

2/6

7/23/2015

Newton'sDividedDifferenceFormula

x2x1
f[x2,x3]f[x1,x2]
x2 f2

f[x1,x2,x3] =

f[x2,x3]=

x3x1
f[x2,x3,x4]f[x1,x2,x3]

f3f2

f[x1,x2,x3,x4] =

x3x2

x4x1

f[x3,x4]f[x2,x3]
x3 f3

f[x2,x3,x4] =

f[x3,x4]=

x4x2

f4f3
x4x3

x4 f4
Example:Computef(0.3)forthedata
x
f

0
1

1
3

3
49

4
129

7
813

usingNewton'sdivideddifferenceformula.
Solution:Divideddifferencetable
xi
0

fi
1

49

23

129
813

19

80

228

3
3
37

NowNewton'sdivideddifferenceformulais
f(x)=f[x0]+(xx0)f[x0,x1]+(xx0)(xx1)f[x0,x1,x2]+(xx0)(xx1)(xx2)f[x0,x1,x2,x3]
f(0.3)=1+(0.30)2+(0.3)(0.31)7+(0.3)(0.31)(0.33)3
=1.831

Sincethegivendataisforthepolynomialf(x)=3x35x2+4x+1theanalyticalvalueisf(0.3)=
1.831
Theanalyticalvalueismatchedwiththecomputedvaluebecausethegivendataisforathird
degreepolynomialandtherearefivedatapointsavailableusingwhichonecanapproximateany
dataexactlyuptofourthdegreepolynomial.
https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

3/6

7/23/2015

Newton'sDividedDifferenceFormula

Properties:
1.Iff(x)isapolynomialofdegreeN,thentheNthdivideddifferenceoff(x)isaconstant.
Proof:Considerthedivideddifferenceofxn
(x1+h)nxn
xn =

nhxn1+...
=

x+hx

=apolynomialofdegree(n1)

Alsosincedivideddifferenceoperatorisalinearoperator,ofanyNthdegreepolynomialisan
(N1)thdegreepolynomialandsecondisan(N2)degreepolynomial,soontheNthdivided
differenceofanNthdegreepolynomialisaconstant.
2.Ifx0,x1,x2...xnarethe(n+1)discretepointsthentheNthdivideddifferenceisequalto
f0

fn

f[x0,x1,x2...xn]=

+...+
(x0x1)...(x0xn)

(xnx0)...(xnxn1)

Proof:Ifn=0f(x0)=f(x0)hencetheresultistrueletusassumethattheresultisvalidupto
n=k
f0

fk

f[x0,x1,...xk]=

+...+
(x0x1)...(x0xk)

(xkx0)...(xkxk1)

Considerthecasen=k+1
f[x1,x2,...xk+1]f[x0,x1,...xk]
f[x0,x1,...xk,xk+1]=
(xk+1x0)

1
=
(xk+1x0)

f1

fk+1

+...+
(x1x2)...(x1xk+1)
(xk+1x1)...(xk+1xk)

(xk+1x0)

f0

fk

+...+
(x0x1)...(x0xk)
(xkx0)...(xkxk1)

f0

f1

https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

fk+1
4/6

7/23/2015

Newton'sDividedDifferenceFormula

+
(x0x1)...(x0xk+1)

(x1x2)...(x1xk)(xk+1x0)

f0

=
+
(x0x1)...(x0xk+1)

x1xk+1

x1x0

f1

+...+
(xk+1x0)...(xk+1xk)

fk+1
+...+

(x1x0)(x1x2)...(x1xk+1)

(xk+1x0)...(xk+1xk)

3.SheppardZigzagrule:
Considerthedivideddifferencetableforthedatapoints(x0,f0),(x1,f1),(x2,f2)and(x3,f3)

Inthedifferencetablethedottedlineandthesolidlinegivetwodifferenctpathsstartingfrom
thefunctionvaluestothehigherdivideddifference'sposssibletothefunctionvalues.The
Sheppard'szigzagrulesaysthefunctionvalueatanynontabulatedfromthedottedlineorfrom
thesolidlinearesameprovidedtheorderofxiaretakeninthedirectionofthezigzagline.That
isanyf(x)throughthedottedlinecanbeapproximatedas
f(x)=f0+(xx0)f[x0,x1]+(xx0)(xx1)f[x0,x1,x2]+(xx0)(xx1)(xx2)f[x0,x1,x2,x3].

Similarlyf(x)overthesolidlineiseuivalentto
f(x)=f2+(xx2)f[x1,x2]+(xx2)(xx1)f[x1,x2,x3]+(xx2)(xx1)(xx3)f[x0,x1,x2,x3].
Example:Findf(1.5)fromthedatapoints

x 0
f(x) 1

0.5
1.8987

1
3.7183

2
11.3891

f(1.5)alongthedottedlineis
f(1.5)=1+1.5x1.7974+1.5(1)x1.8418+(1.5)(1)(0.5)x0.4229
https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

5/6

7/23/2015

Newton'sDividedDifferenceFormula

=6.770
Similarlyf(1.5)alongthesolidlineis
f(1.5)=3.7183+(1.51)x3.6392+(1.51)(1.50.5)x2.6877+(1.51)(1.50.5)(1.52)x0.4229
=6.770
Thedataisgivenforf(x)=x2+exandtheanalyticalvalueforf(1.5)=6.7317

SolutionofTranscendentalEquations|SolutionofLinearSystemofAlgebraicEquations|Interpolation&CurveFitting
NumericalDifferentiation&Integration|NumericalSolutionofOrdinaryDifferentialEquations
NumericalSolutionofPartialDifferentialEquations
search

https://mat.iitm.ac.in/home/sryedida/public_html/caimna/interpolation/nddf.html

6/6

You might also like