UNIT -I
INTERPOLATION
Introduction:
If *f is a function from x into y in the domain (Xo, Xow) such that y = f(x), for each value
of x in the given interval, the value of *y’ can be determined just by substitution. However, ifthe
function ‘f is not known, the value of “y' can be obtained when a set of values of x and f(x) are
given. The method to find out such is based on principles of finite differences.
Argument and Entry:
Suppose for a function y = f(x), the function ‘f” assumes the values f(xa),
£(K04h) £(K0+Dh)....so++ns KotH) (OF) Yo, Yt, Y2 vee
xot2l «a, Xotnh (Of) Xo, X1,X2,....,Xe of *x’ respectively, Each value of the independent
vatiable *x’ is called “Argument” and corresponding values of the dependent variable *y’ is called
“entry”. The difference between aff twOgongPcutt¥ ary Known as interval of difference.
Which is denoted by ‘h’.
Interpolation: x
‘Suppose we are given the followin les ol fox) Mor a seM¥E values of x:
,, Yn corresponding to Xo, , Xo+th,
xoth | x0+2h xotnh
x | x0 _
x x Xn
| fex0) | focorh) | £(xo#2h) |---| fXornh)
yo Ly fo a
‘Then the process of finding the value of y corresponding to any value of x between xo and xotnh
is called “interpolation”. Interpolation is the technique of estimating the value of a function for any
intermediate value of the independent variable in the given range. While the process of computing
the value of the function outside of the given range is called “extrapolation”.
Finite Differences:
Suppose that the function y = f(x) is tabulated for the equally spaced values x = Xo, Xoth, xo +2h, ~
Xo-#h (Or) XoX1,X2,.....Xn giving y= yo, Yi, Ya ya. To determine the values of f(x) for
some intermediate value of x the following three types of differences are found useful. They are
(i) Forward differences (li) Backward differences (lif) Central differences.
PagetForward Differences ( A):
Suppose yoynye
| xo-tnh (01) Rox, ,...XnOf X respectively. Then yi-Yos Y2-Yin Y3-Y2s
- yadenote a set of values of *y’ corresponding to the values Xo, xo,
x0 +2h,
yar Yor
are called forward differences of y if these differences are denoted by Ayo, Ayt, Aya, Ayn
respectively.
sAyo = yi-Yor Ayn = ya-yie Aya = YsrY2s~ AYo-t = YarYo
nu where A is called the forward difference
In general, Ay+= yeet-Ye where t= 041,23,
‘operator and Ayo, A yi, Aya
Ayo are called first forward differences. The difference of first
forward differences are called second forward differences and are denoted by A?yo, A’y1, A?y2,——
Myo = Ay: - Ayo
Myr = Aya -Ayt
A?y2 = Ays - Ay2 —
In general, A*yp= = Aysst-Ayr, t= 01,2,3,-~ =
Similarly we can define third forward differences, forth forward difference etc.
Forward difference Table:
Value of | yatue of at ard 3" forward
x ¥ | waitterd cay) J) diferen¥Q(ay) | difference (4*y)
xo yo
yiryo = Ayo
xoth yt Ayi-Ayo=A?yo
yryi = Ayr My1-A?yo=A3y0
xot2h ye Ayr-Ayi=Ay1
Yary2 = Ayo
xot3h ys
Page2o 2Construct forward difference table for the data given below.
x[i]2]3]4]5
y | 4] 13 | 34[73] 136
Value of | Vatue of y 1" forward Hon forward a Forward AE focward
x difference (Ay) | difference (Ay) | difference (Ay) | difference (A*y)
1 4
13-4=9
2 13 21-9212
34-13=21
18-12=6
3 34
73-34=39
24-1856
4 B 3-391
136-43=63
3 136
Backward Differences (V):
Suppose yoy,y2r -yn denote a set of values of y corresponding to the values
xo,xoth.xot2hy----+, Xotth (Or) xox, X,
are called backward differences of y if these differences are denoted by Vyi,Vy2, Vy3-—~
Xaf x respectively. Then y1-Yo, Y2-Yi,Y3¥2.
respectively.
Va = YarYort
In general, Vye= yeyr1, where r= 1,2,3---------n where V is called the backward difference
__ Vynare called first backward differences. The difference of the
+ Vyi = yiryo, Vy2 = yo-¥isVy3 = Y3-Y2r
operator and Vy;, Vy2,Vy3——
first backward differences are called second backward differences and are denoted
byV*y2,V2y3,V7y4
2Vy2= Vy- Vy).V2ys = Vys - Vy2, ys = Vye-Vys,
In general, Vy= Vy Vyra, P= 1,2,3,4,-—-----
Similarly, we can find third backward differences, forth backward differences etc.
Page 3 >=Backward Difference Table:
Value of | Vatue of T* Backward 2 Backward | 3"“Backwarddifference
x ¥ | difference (Vy) _| difference (V*y) Wy)
xo yo
yryo=Vyi
xoth yi
Vy2-Vyi= Vy2
yoy = Vy2 Vy3-VyViy3
xot2h ye
Vys-Vy2=V°ys
yorye= Vys
xot3h yp
Construct backward difference table for the data given below:
x]0
1] 2
yf-l
1| 5
‘Value
of x
Value of y
4*Backward
difference (V*y)
“I
itt
19
0-0-0
Page 4Central Differences (5):
Suppose yo,yi,yar~
xa,Xo+h,xo+2h,-
‘ya denote a set of values of ‘y’ corresponding to the values
th (OF) 0X1, X,..... XaO£ X respectively. Then yi-yo,yo-Y1Ys-Yoo—sYor Yar
are called first central differences of y if these differences are denoted by dyi2, SysaSysi
~Syo.in respectively
dyin = yi-yo,Oysn = y2-Yr, Sys2=Y3-Y2, by:
= YarYout
.n and ‘6’ is called the central difference
+ Syria = Yr Yer, where r= 1,2,3+
In general
operator and dyv2, Sys2,5ysn.~
-6ye-in are called first central differences of y. The differences
of the first central differences are called second central differences and are denoted by 5*y0, 5?y2,
oy,
25%yy = Bysn — byt BPy2 = Bysn — Syse , 8y3 = Syn2 — ys2 —
Similarly we can find third central differences, forth central differences etc,
Central Difference Table:
Value | Value of 3° Central
of x y Difference 6°y)
Xo Yo
Xoth Va
Oy, - Oy = Sys
x+2h| ye a
Xo + 3h Ys
Page 5°" 2-Construct central difference table for the data given below.
x]2]4] 6] 8/10
y |15| 19] 27 | 30 | 50
Value ot] 1 Central 2Central | 3%Central 4®Central
value of x | Y@ef| difference | difference | difference | difference
z y) @y) @y) @y
2 15
19-15=4
4 19 B-d=4
27-19=8 -5-4=-9
22-(-9)=31
6 27
30-27=3 17-(-5)=22
8 30
30=.
10 50.
Note: A f(x) = f(x+h) — f(x) (or) ca
Problems based on First Forward differences:
(A) Find the first forward differences of the following functions:
(@ cos x (v) tan'x
(ii) log x (vi) £@) g(x)
(ili) sin (ax+b) oi
1
(iv) e* log bx (vill) 725
(@ A [cos x]
Sol: Let f(x) = cos x
Let ‘h’ be the interval of difference
‘We know that A f(x) = f(x+h) ~ £00)
‘A [eos x] = cos (x+h) — cos x
== asin 2B, sigH8S2 [cos ~ cos # = ~2 in 222) sin (4)]
{ feos x] =-2 sin(x +4) sin (2)
Page(ii) A[ log x]
Sol: Let f(x) = log x
Let ‘h’ be the interval of difference
‘We know that A f(x) = f(x+h) - f(x)
Al log x] = log (x+h) - log x
=log (2%) [log —logB = log @]
=log (1+4)
A [logs] = 10g (=
i) A [sin (ax+b)]
Sol: Let f(x) = sin (axtb)
Let ‘h’ be the interval of difference
‘We know that A f(x) = f(c+h) — (0)
[sin (axb)] = sin [a(x+h) +b] - sin (ax)
= sin (axtah+b) ~ sin (ax+)
A [sin (axtb)] =2 cos (ax +
(iv) A Le* log bx]
Sol: Let f(x)
Let ‘h’ be the interval of difference
We know that A f(x) = f(a+h) ~ f(x)
A [et log bx] = e%*" log b(x+h) ~ e* log bx
=e Jog b (xth) — e€ log bx
= ee" log b (x+h) — e™ log bx
* Jog bx
A [e log bx] = e* [e* log b(x+h)- log bx]
(y) A [tan x]
Sol: Let f (x) = tan "x
Let ‘h’ be the interval of difference
We know that A f(x) = f(x+h) ~ £0)
A (tant x] = tan (xth) ~ tan! x
tan (#=*] te tanctx — tanvty = tan?
= tan 23] [e tan“ x — tan“4y = tan’
Page7+ Atan! x =tan™
ae
(wi) A [£800]
Sol: Let f(x) = f()g(x)
Let ‘h’ be the interval of difference
‘We know that A f(x) = f(x+h) — f(x)
2A [flos)eCa)] = fcth)g(x#h) ~ fo0)g(%)
Subtracting and adding f(x+h)g(x) in R.HLS of the above equation we get
= f(xth)g(x+h) - f(octh) g(x) + flx+h)g(x) - fx) g00)
= [fOcth) g¢xth) - f(x+h) go] + [fGc+h) g(x) ~ £6) g@0)]
= fixth) [g(c+h) — g(x)] + g@0Lf+h)-£00]
= f(xth) Ag(x) + gOOAL@)
A [Eg] = fOr) Ags) +8Q)ALC)
oii [2 cS
Sol: Let f(x)
Let ‘h’ be the interval of diffe
‘We know that A f(x) = f
a] = 10
7 [E een),
jae) gen) 9)
flxrnae)=fOale+h)
‘g@+hace)
‘Subtracting and adding {(x)g(x) in the numerator of R.H.S, of the above equation , we get
flesnals)-eow+sod9GI-fG. +h)
aCehg@)
= Serer h)=F(aII- Flora)
acer)
= 92D) ~ 604 868)
acetha(@)
2 a[LQ] = seater 4900
1). ‘act a@)
(viii) A Gl
Sal: Let £09 =Fe5
Let ‘h’ be the interval of difference
‘We know that A f(x) = fcth) — f(x)
Page 8=2160)
a
* FCOF=+H)
erm) 7G) Fes C+h)
=A f00)
Fos Cen)
(2) If the interval of difference is unity, then prove that
(HA [x(x+1)(542)(43)] = ACH) (42) 43)
Gi) A taw*(2) = ta(S5)
Sol: (i) A [x(x )(x+2)(X+3)]
Let f(x) = x(x+1)(x+2)(x+3) and given that interval of difference is unity (ie) h=1
‘We know that Af(x) = f(x+h) — £00)
f(x) = fot) =f) [+ h=1]
A [x(t 1)(x42)(043)] = (IY OCHL HL (O42) (4143) — xO) OC+2) KES)
sc 1) (x42) (043) [x4 -x] = 44 1)0042)0043)
A [x(chL) (42)(43)] = 4x41) (4+2)(43)
Gi) 4 tan(—)
Let f(n) = tan"! that il is unity (ie) h=1
We know that A f(qph) -
A f(a) = far
«stan (2) =u
>
a n—(n-a}(ne9|
na)
(nt) 4n-0)
Ge
n2—(nt—a)
mn neteni
=tan]
tan'(Za)
Paged >Problems based on Second and higher Forward differences:
(1) Evaluate (i) Ae") (ii) A?
Sol: (i) A? (3e)
Let f(x) =3e*
Also given that h=1
‘We know that A f(x) = flo#h) ~ £0)
And A? f(x) = A [Af()]
=A [fcth) - £()]
= A f(cth) - AF)
= [i(a+h+h) - f(x+h)] - [f(+h) - £@)]
= [f(x+2h) — fGx+h)] ~ [fGe+h) - £00]
A? f(x) = f(x+2h) - 2 f(x+h) + f(x)
A? £(x) = f(oc+2) — 26(%+1) +600)
Now f(x+1) = 3e"#?-~
and f(x+2) = 3et2----
Substituting (1) (3) and (4) in (2SMget
A? Be%) = 30%? — 23e"*!) +3e*
A? Be’) = 3e**? — Ge"! + 3e*
= 3e* [e2e41]
«A? Be) = 3e* (0-1)?
Sati?
Fa Sa+6.
Let f(x) = 22
ata
$382) ge he
$112) taking h = 1
(0)
=1] -
Gi) a?|
-- (1)
Also given that h=1
We know that A f(x) = f(x+h) ~ f(x)
And A? f(x) = A [Af(%)]
A [f(x+h) - £60]
= A f(x+h) - Af)
= [flxeh+h) - fx+h)] ~ [£Gx+h) - £)]
= [fGc#2h) — f(x+h)] = [f+h) - £00]
A? f(x) = f(x+2h) — 2 f(x+h) + f(x)
A? f(x) = f(x#2) -26GcH1) Hf) [eh
Pageto. —stetye12
Gaptscrre
= Sota)
= Gauss
eft 1=
Now f(x+1) =
se+i7
Sa
s(x42) +12
GerDesGraes
sxi22
and f(x42h) =
-- (4)
f(c+2h) =
240x420)
»Substituting(1),(3) and (4) in (2), we get
agri __sei22_ 4 (_Satt7_) , setae
Savsxsé ~ xe9ne20~ \Fae7atia) Tx245xe6
(3) Evaluate A"{e™*"]
Sol: let f(x) = e®*? and ‘h’ be the interval of difference
a AG
co (emt) = Aor _ goed
Aer se fe]
Now A? [c™**] = A [A(e™**)]
=A fe (o® -1)] fr
=™- DAE)
= (e®-1).{(e%1) e*} [from (1)]
wer)=eye”
Proceeding on n times we get, A? (e™**) = (e*-1)Fe™"
(4) prove that A® (e*) = (e=1)"e™ (or) If f(x) = e%, then show that A f(x) = (e*-1)Pe™*
Sol: Let f(x) = e"also let *h’ be interval of difference.
We know that A f(x) = f(th) - £()
a A(e8) =e — eo = oe 1]
#A(e®) = e™(e"-1)- -——- (1)
And A? (e") = A[Ae™]= A [e® ("™-1)] [from (1)]
=D AC)
=(e™1) [(e-1e"] [from (1)]
2A? (e%) = (CM 1Pe™
Proceeding on n times we get, A® (e%) = (e*-1)"e™
Page 14Other Difference Operators:
We have already introduced the operators A, V and 5, Besides these, there are two operators. They
are (1) Shift (or) displacement operator (2) Averaging operator.
Shift Operator: Shift operator (or) displacement operator is denoted by ‘E’, Shift operator E is the
operation of increasing the argument x by h. So that
E f(x) = f(x+h) E ys= Yuen
E? f(x) = f(x+2h) E’ys Yoon
E? f(x) = f(x+3h) BPyx = Yuen
B® f(x) = f(c4nh) Ey = Yasoh
Eis called inverse shift operator and is defined as
EB" f(x) = fh) Ely,
E? f(x) = flx-2h)— E’y= Et
EB" f(x) = f(x-nh)
Where ‘n’ is any real number.
Averaging Operator: “st byw’ ands defined as
at +7,
:
peont[r (er) + 16-2)
Note: In the difference calculus ‘E” is regarded as the fundamental operator and A, V and 6 and p
can be expressed in terms of ‘E*.
Problems:
x,
(1) Prove that e* = (© Set. Fil the interval of difference being b.
Proof: Leth be the interval of difference
Consider = (")e*=
AP E~*)e*= A2(E~*e*) = A?(e*"") = A2(e*e™") =e NA?e*
Hhexths exth-h
Page 12.2
(2) Evaluate (=) x? if the interval of difference being bh.
Proof: Let h be the interval of difference
@
_ (et28+1)
Pees
=(E-2+E4)x?
= Ex? — 2x? 4 B4x?
= (x +h)? 2x74 (x - A)?
=x? + hx +h? — 2x? + x? — Zhe +h?
) x2 on?
Relations between Operators:
(1) Prove that A = E-1 (or) E=
Proof: By the def of forward diffe Ss, We
Af(X) = f(x+h) - £00) AYR= Yeh Bs,
=E M(x) -f@) =EyA
Af(x) = (E-1)f(x), Ayx = (E-1) yx
A=E-1 (or) E= Atl
(2) Prove that V=1-E4 (or) E4=1-V
Proof: By the definition of backward differences, we have
E(x) = f(x) ~F(-h) Vs =¥x-Yeu
= F(x) - Bf) = yx- BAys
VEG = (-E4) £08) Vyx= (1-E4) yx
V=1-E" (or) E'= 1-0
(3) Prove that (1+A) (1-V) =1
Proof: By the def of forward differences, we have
A sE-1 =>At1=E
By the definition of backward differences, we have
V=LEt =>1V=E"
Page 13.0" =(+d) (-V) = BEI =1
(14a) (LV) = 1
(4) Prove that AV = A-V
Proof: By the def of forward differences, wehave A =E-L
By the definition of backward differences, we have V=1-E+
AV = (E+1) (1-E*)
=E-EB)-1+E7
BE =1]
AV=A-V
(6 Prove that VE=EV=4
Proof: By the def of forward differences, we have A =E-I
By the definition of backward differences, we have V = 1-B*
2VE= (1-E) B= EEE =E-1=8
EV =E(1-E2) =E-BE?=E-1=4
VE =EV=4
(Prove that 6= EF -E?
Proof: By the definition of central differences, we have
B= Yeh Yad
2 at [BY = Vy
Eiyx-E® yx eS ie
= (82-87 )yxb= Ei—
4
(8) Prove that A= VE = 6E?
Proof: By the def of forward differences, we have
By the def of backward differences, we have
By the definition of central differences, we have
edit
=5-8-E
eB
= 88 =E-1- )
a
«From (1) (2) and (3) we get A= Jig 5
(9) Prove thaty =3 [2 +5
Proof: By the definition of averaginipperat
(10) Prove that 4+ ¥) = 1d
Proof: By the def of forward differences, we have A = E-1
By the definition of backward differences, we have
By the definition of averaging operator, we have
By the definition of central differences, we have 5 = E? — E7
Naty =He-1+1-8")= XE-EY)
(ei+ 87)(#-F
z
from (I) and) 4+) = nb
(11) Prove that AV=A-V=3"Proof: By the def of forward differences, we have A =E-1
By the definition of backward differences, we have V=1-E1
By the definition of central differences, we have’ 6 = BEF
AV = (B-1) (1-E)
= E-BE1-1484
= E1148 [s BE = 1]
=(€1)-(1E1)
=A
= E+E!-2=E-1-14E"
= (E-1)-(-E") = 4
HAV
from (1) and@Q) AV=a-
(12) Prove that Et =p +2
Proof:
By the definition of averaging operator, we have = 2(E2+ EF)
By the definition of central differences, wehave 8= £2? — E=
Consider +
+
(e+ 8?) +4(- £7)
(13) Prove that £7 =p-
Proof: By the definition of averaging operator, we have w=2(B? + E)
By the definition of central differences, we have 6 = E2—E=
Comider p- $=2(eE+ £F)-2(et- #2) =1(207) 07
a 8
Be ap-d
Page 16 +43(14) Prove that (e+ FF) (+a) =24A
Proof: We know that A= E-1
(é
i+ F)a+ar=(B + EF)Q+E- ve
=(Ei+ EF) OF
=E4+1=A+1+1=A+2
(E+ BF) +aye=2+4
(15) Prove that E = 1+ =e"?
Proof: By the definition of forward differences, we get
AsE-1 >E=1d~ -(1)
By the definition of shift operator, we have
Ef(x) = flxth)
= f0o+h FO) +2 FQ) + +E elas
= f(x) +h.Df(x) +p foOe> 3D
= (x) + (hD) f(9) +22 RY
=[i+no+ &™ or ow!
set? fix)
E f(x) =e? f(x)
1
etaltxt =
=E - (2)
From (1) and (2) we getE = 1+A =e"?
e
(16) Prove that 14 w26?= (1. + 35%)?
Proof: We know that = and 5
L(ee+ BFP
+ (ee ps2 ERE}
wailE+ E+ 2]
8=(B- Ez)
a= (E)P+ (EF) 2 Bee
Page 17 of >S=HE'-2
Now 1+ y= 1+ 3[E+ E+ E+ &*— 2]
sit ie + EP - (2)
=l+ {E+ EOP -4}
st ${[E+ EF }-1
(1)
Le w8? EE + BO
and (1 + 56°)*= 1 +P +E 2)?
=(1+$+E%)- 1?
= GE+E)*
(2)
(a + $6%)?= 3 +E)
From (1) and (2) we get 1+y6?= (1 + 35%)?
(17) Prove that p?=1+35?
Proof: By the def of central differen« wes
= E+E
state + B+ 2]
=i [6 +2) + 2B +B = 8 +2)
_ S44
Page 18:Differences of a polynomial:
‘Suppose f(x) is a polynomial of degree ‘n’ and the values of x are equally spaced, then 4" f(x) is a
constant.
Proof: Let f(x) be a polynomial of degree ‘n’
(ice) f(x) =agx" + ayx"? + ax 2+
anday 40.
Also let ‘h’ be the interval of difference
We know that, A f(x) = f(cth) — f(x).
2.AE(R) = Lag (e+ A)" aye] + [a(x + YOY = a8] + —— HH yn +H) — Anant} tla nl
a agl(x +h)" = x7] + aul Ge + AYE aE] t+ nal +H) — 2]
se ad|nega" + nex" MN Max 2R? 4 = = = = 4 hey th" $e, ht] — 2°}
OnaaX + dy » Where d,81,82 fag are constants
tarf[(n— Dex" + (= Nex ht Vogt” 3.h? + —— +0 Neg gh™
On aah
+
+ Ayah.
1+ [a
= agnhe'
+ [agnh™? + ayeqaayh2] x (@oh™ + ayn™4)
f(x) = ao nbx™! +2 x"? +s x9 + + by ix+ba, where ba, bs,
Here the polynomial Af(x) is of degree (a-1). Thus, the first difference of a polynomial of n
--baare constants.
degree is a polynomial of degree (n-1).
Now A? f(x) = 4 [Af]
= Afagnhx"@? + box"? + bax”3 + ——— + Dyix + bal
= ath [(xc + AYP2 = x4] 4 bal (a + AYP? — 2-2] + + ba lxth-x] + (bn-bn)
= agnh [Ge + AYE = x4] + bal + A)? — x] + et Bes 40
=agnh {[n- 1x2 +n— 1,0" hen 128 E -+n-1,,_,h"
vty
beth.
+ baf[n — 2,277 +m —2,,x" R$ xh? + —— 2n ati 2-2},
nod 4 CHMOD 52 yn’ a
= agnih{(n — thet? + SED peat? + —— 4h )]
+ bof — apne? + SAPD ptens s —— ht 2] 4 nt ba he
= aon(n-1) h? x"? +[a.2 A + by(n — 2h] x”?
20? f(ox)=a0 n(n-1) RPL xB? tor xP 3402 x4 toot Coa X4C02]
Page 19Where ¢1,02,085 ~
Ca Ca2 are constants,
Here the polynomial A?f(x) is of degree (n-2). Thus, the second difference of a polynomial of n™
degree is a polynomial of degree (n-2).
By continuing tis process, we get A? f(x) = aon(n-L)(n-2)(0-3) r= 3:21. aght'n!
2A f(x) = ach nt which is @ constant,
Hence the proof.
Note) IF (x) is a polynomial of degree n, then AP4(x) is constant. It follows thata™ f(x) = 0,
AP? f(x) =0, AX) =O,
(2) The converse of the above result is also true. That is, if A” f(x) is constant and f(x) is tabulated at
equally spaced, then the function f(x) is @ polynomial of degree ‘x’
Problems:
(1) Evaluate A"((1 — ax)(1 — bx2(1 —cx8(1-dx*)] ifh=1
Sok: Let f(x) = (1-ax)(1-bx2)(1-0x2)(1-dx) 50 that £3) is 10 degree polynomial. Also given that b=1
Now Consider A! [(1-ax)(1-bx°)(1-x?)(1-°)]
= A! {(-a)(-b)(-c)(-d) x"? + terms cont Ww
= AXfabed x!9 + terms containing poffers of W leggthan
= abcd 10! 119 If f(x) = aox™+ax™ =
(2) Show that 4? [(1-)(1-2s)0-290-4
Sol: Let f(x) = 1-x)(0-2x\1- 2091-4
Also given that h=2
A! fxet)C1-2320- 32) (LA) = A (CICD)CSICA) x1 + terms containing powers of x less than 10]
= 8(24x:9 + terms containing powers of x less than 10]
2481002" h-= 2 and f(a) = agx® + aye? + ~~ — + anthend"F) = aon! h")
2 A[(1 — x)(1— 2x7(1 - 3x81 — 4x*)] = 247210!
‘Newton’s Interpolation Formull
Let the function y = f(x) takes values yo,ytoY2s , Ye corresponding to the equi-spaced values
Xo,xoth,xo#2h, xotnh (0f) XojX1,X2, _ xaof x respectively. Whenever we find the value of
the function y = £() using the given equi-spaced data within the given range (Koxo¥0) (07) (oka) We Use
‘Newton's forward interpolation formula (or) Newton's backward Interpolation formula,
Newton’s Forward Interpolation Formula:
Let the function y = 9) takes values yo.y1.y2s "Ye corresponding tothe equl-spaced values
xoxoth,xot+2h, +, Xo+mh (Or) Xo,X1,X2s
xa of x respectively. Suppose it is required to estimate
the value of y or fix)using the given equi-spaced data within the given range (oaxrtnh) (0 (%0.%0) for x=
xetph, where pis a real number, Newton's derived a formula which is known as ‘Newton's forward
Page 20 1:33interpolation formula (ox) Newton's forward difference formula (ot) Newton's Grigaryforward interpolation
formula (or) Newton's Grigary forward difference formula.
For any real number P,
y= f(x) = f(xotph) = FP fo)
= EP yo
= (1 +A)? yolv 1+ 4= E]
= [pcjt7a? + pera + Pel? 28? + poy? B+ -—— + Pep??? yo
efit pa sda atten a ————]y0
= yotpbyot2 222 099 +
POD Atyy + BODED ayy + HODEDED aby 4.
e-0-2)
Say
y=fo
Which is called Newton’s forward Interpolation formula
jotpAyo
229 Were h = interval of difference
Where P
xo= Initial Value of x or initial argument in the given data
x= where we nd of
Note: Newton's forward Interpolatio ed to 1e valt ce) .ction near the initial argument (x0)
‘and also for extrapolating values tothe left Ojo ff the ta Mgaui-spaced,
Problems:
(), Find y(S4) given that y(50) = 205, ¥(60) = 225, ¥(70) = 248, and y(80) = 274 using Newton’s
forward difference formula,
Sol; Given that y(S0) = 205, y(60) = 225, ¥(70) = 248,y(80) = 274
x [50 60 | 70 | 80
y [295 225 | 248 | 274
(ie) The difference table is
x y Ay ay ay
50: 205 yo
60
3A" yo
Se 0%
70 eats
3
26
80 247
Page 21 oi 8Here so = 50, yo = 205, Ayo= 20, A*yo=3, Ayo = 0
X =5x-0, h =60-50 = 10
oxy _ S40
pare.
nT 40
2. Newton's forward difference formula is
y=1(e) = yoepayn 2 sty HEED ty, 4 REDEDEND) pty
(64) = 205+40.4)(20)+ enesn 9) + LNSNOAD 0)
= 205+8-0.3640 = 212.64
0.7071, Sin 50° = 0.7660, sin 55° = 0.8192 and sin 60° = 0.8660. Find sin 52° using
Newton’s forward interpolation formula.
(2). Given sin 45°
Sol: The difference table is
py ay wy aty
45x, 0.7071 yo
0.03894
50 0.7660
= tHQOS 7A? yo
0, -0,00074°¥o
55 0.8192
0.0064
0.0468
60 0.8660
Here xo =45, yo = 0.7071, Ayo = 0.0589, A? yo =-0. 0057,A? yo = -0.0007
43
h=50-4
x= 52, P= =14
_
Newton's forward interpolation formula is y = f(x) = yotPAyo ED atyy + BEDERD pSyo +.
yesin52?=
7880
@), Find £(1.6) using Newton's forward difference formula for the data given below.
x] i [14] 1s | 22
Hex) | 3.49 | 4.82 [5.96 | 6.5
Sol: The difference table is
Page 22°35x y ay wy aty
1, | 349%
1.334y0
14 | 482
-0.1907¥o
041A Yo
18 5.96 Lid
0.60
22 65 0.54
Here x0 =1, yo = 3.49, Ayo=1.33, Ayo = 0.19. A°y0
wom 16-115
He 14-1204, x16, p= 8 =4
Newton's forward interpolation formula is
yefeo) = yoaphye AE Datyg + ODED yg + HEMEDED ayy i
“(1.6) = 3.49+4(1.5) (138) + © DEED (0, 19) +2DESDES? (-0.41)
(1.6) = 5.4394
(Apply Newton's forward interpélggipn for Pt 1c of V5.5, given that V5 = 2.236,
YG = 2.449, V7 = 2.646, and VB = 4828.
Sol: Given that v5 =2.236, V6 =2. 2. BB 2.828
8
2.828
The difference table is
x y oy ay ay
By | 223630
0.213Ay0
6 | 2489
=0.0164*¥o
0.197
7 | 2686 0.0018
0015
0.182
3 | 2828
i
5, yo= 2.236, Ay
ete S55
is = S205
Here \D13 Arye = -0.016, A*yo = 0.001
He
x=5.5,
Newton's forward interpolation formula is
Page 23 i>y= fo) = yorpbye BP.
y= V5.5 = 2.35075
{jp + OME p2yy + RODEEDE Alyy
(8). For x= 0,1,2,3,45 la) = 11415,5,6. Find £(3) using forward difference formula.
‘Sol; Given that f(x) = 1,14,
.15,5,6; for x=0, 1,2,3.4
{= 0 1]?
nk
‘The difference table is
x ¥ Ay ay ay ay
Ox, lye
13Qy,
1 4 -12.b7y0
1 1A°Y0
2 15 11 21 D*Y9
l
2
3 5 ul
1
Newton's forward interpolation formula is
y=ftx) =yotpdyo 2220 ODE sy, 4 REEDED ayy
9) = HS ESD HE (y+ SoDaDED
= 1439-3641 4055
(6 From the following data, estimate the mumber of students who obtained marks between 40 and 45
Marks 30-40 | 40-50 | 50-60 | 60-70 | 70-80
No. ofstudents | 31 a a | 38 3
Page 24.133the Rearranged data is
Marks Less No. of students
40 31
50 31442273
60 734515124
70 124435159
80 159+31=190
Marks Tess than (x) | 40 50 60 70 80 |
No. of students (y) | 3! B wa | 159 190)
‘The difference table is
Newton’s forward interpolation formula is
fhm = 1) (p- -1)(p-00-:
yefoy =yorntyn PE D g2yy + BEDE 2 p3y, + PODEDOD hy, +-~
aye = 314(08)(42) (08105-9 (9 (osy0s-D1052) 95) (osyos-1105-B05-D an
2eyss = 47.86 = 48
“The no. of students who obtained marks less than 45 = 48
‘The no. of students who obtained marks less than 40 = 32
«The no, of students who obtained marks between 40 and 45 = 8-31=17
Page 25al which takes the following values Yoo) = YQ) =O¥@ = 1 and
(2) Find the cubic polynomi
formula and hence find ¥(4)-
‘ycq) = 10 using Newton's forward difference
Sal: Given that yo) = 14 Yea) = 0172) = 1G) = 10
Ge) Poeeee
yi 0 | 1 \10
‘The difference table is
‘Newton's forward difference formula is
4 BODE 12y, 4 HODEDE aly, +.
y= 10s) =yotpye ASP Ay
ay= fo) =H CDHP@ BED)
are @ HEE
SLEEK? BR OK
S241
f(x) = x- 2x41
ay(d) = F(A) = 42-24)841 = 64-32-1533
Page 26Let the function y = f(x) takes values YoY1sY2
. yacorresponding to the equi-spaced values
xo,xoth Xo#2h, ==» Roth (OF) KOKI xe of x respectively. Suppose it is required to estimate
the value of of f(x)using the given equi-spaced data within the given range (xoxxo#nh) (or) (Xos%n) For x=
xoaph, where pis areal number, Newton’s derived a formula which is known as Newton's backward
interpolation formula (or) Newton's backward difference formula (or) Newton's Grigarybackward
interpolation formula (or) Newton's Griganybackward difference formula,
For any real number P,
y= 00) = frm + A)
= EP Exp)
EP Yn
E'y"yn
=0-)%— (EASE)
ofr eyv Opt BEE Ya
{using Binomial theo!
(1-2) = Meg Beak + Re
Goa-otei-m ie -&--
if we replace x = fect
a-vyPeat + pr Ot De -RELDOtP +oocc oT
yor pyar LOD Vy, 4 HENOMD pay, 2ODEHDOD yy,
Which is known as Newton's backward interpolation formula.
Here h=
= interval of difference
iq = last value of x or last argument in the given data
x = where we have to find the value of y
Note: Newton's backward interpolation formula is used for interpolating the values of y near the end of a
set of tabulated values and also for extrapolating values to the right of Xp if the given data is equispaced.
Page 27 *(@) Using Newton’s backward difference formula, find the viue of y at 9 from the following data
x) 2])5 /8 [2
y | 948 | 879 | 813 | 751
C ay
&
bs,
= 75.:1460.67)(-62)4+ 22S e149) (9.4) 42 snicaste067+ (0)
15.1+4.154-0.04422-0.004901
9.2049
@ Using Newton’s backward interpolation formula, find the polynomial satisfied by (3,6)(4,24)(5,60)
and (6,120)
‘Sol: Given that (3,6)(4,24)(5,60) and (6,120)
xT2])4)]51)¢
y] 6 | 2 | 6 | 120
‘The difference table is
Page 28ay
607Yn
Ox 120yy
Here x= 6, ye=120, Vy:
pager xx peta
Newton's backward interpolation yer)
= 1204 (x-6)(60)+ 2-2" (24)
= 120+60(x-6)+12(%-5) (¢-6) +
R ~
2)
f(9)= 99-3242
For the folowing data extrapolate the value ofy at 4.2 using Newton's backward difference formula.
xy 02 | 04 | 06 | 08 | 10
Uap | Bazza | 3.2499 | 3.3825 | 3.5205 | 3.6042
‘Sol: The difference table is
(ee y 4) x |
02 3.1204
0.1275
04 3.2499 0.0051
0.1326
08 3.3825 0.0054
0.1380
06 3.5205 00879
| ousszese |
1.0, _|_ 3.66429.
Page 29O,ysH3.6642, Vye=0.1437, Veyn=0.0057
¥3=0.0003, V4ys=0, n= 04-0.2=0.2,x= 12
ete _ bao 7
Pe apa
“=: Newton's backward interpolation formula is
y=¥o py 2 Dany ryt OO et
= 3.66424100.183) HD (0.0057) 2222 (@.0002)+ aHaHIE (0)
= 3,6642+0.1437+0.0057+0.0002+0
= 3.8138
(4) The population of a town is as follows.
year [eae ea Tn Tost | 1991
ylation i ro
populationin | 49 | 24 | 29 | 36
lacks
Estimate the population increase 4)
Sol: The difference table is
¥
x | (dis difference,
1941 20
1951 oa
1961 29 *
tom | 36
1981 | 46
i991 | SI
Poy the year 1946; Here we use NF.LF.
Here xo 1941, yo= 20, Ayo=4, Ayo = 1, A'yo= . Aty0=0, Afyo.
H= 1951-1941 = 10, x= 1946, p= 222
yn.
=05
Newton's forward interpolation formula is
y=yorpAyeeZd Atyer BORED ay 2O-DE=DEND) nye
Page 30eyhpng = 20K0 SHH CNED OE ME $04 4 Q5108-DOSDOED
(os 5-05-0081
= 20#2-0,125+0.06254+0-0.24609
= 21.09
2yygqe = 21.69 lacks
Population in the vear 1986: Here we use N.BLLF.
Way, = -5, V2yn = -8, VF
1906-1991
sen = 05
: Newton's backward interpolation formula is
y= yt yn DT PONE, pA 2ODEDED I, PERDODONNEED GE
jngae = 51+ (-05)5+= cH 5 costcnsencns castcasencost2 9)
posse
2 51-2:540.625405+0.351564 424609
Yasee = 50.223 lacks
‘: Increase population during 1946 to 1986 = Ys86~ Yass
= 50.223-21.69
= 28.533 lacks
Here xp = 1991, yn= SL Yn’
H = 1951-1941, x=1986,P
acs 543)
’
game oseDcoseDIEDs
Exercise:
apestimate the values of (22) and f(42) from the following data
x, a] 3 | 2 35] «40 s |
[9 354 | 332 | 291 | 260 | 231 204 |
(2)From the following data estimate the no. of students who obtained marks between 65 and 70
Marks: [200 40-50 | 50-60 | 60-70 70-80 |
No. of Students | 31 42 SL 35 31 ‘|
Pare 32Central Difference Interpolation formulas:
Tn the preceding section, we derive Newton's forward interpolation formula and Newton’s backward
interpolation formulas which are applicable fr interpolating near the beginning and ending of the tabulated
values of equispaced data, To interpolate the values of a function nearer the middle of the tabulated values
Central difference formulas are developed which ae Gauss forward {interpolation formula, Gauss backward
interpolation formule, Sterling's formula and Besse's formula.
pendent variable °x takes the value wt ~ 2hxo— #0, 0% xp + 2h.
m2, Yt» Yor Ya
Suppose th
nd the corresponding values of an unknovin function y= f(x) are
"respectively, then the difference table is.
tw 1 Central FP Central |
Value | Valueof | gi ‘2° Central oh cone
ofe | ances | ainsi Diary | iter’
Xg-2h| Y-2
Yar Y-2=
aye
xy-h | %4 i
Yo~Y-1 =|
oy!
x | %
3
90 =
am yy
xpth | % oy
8-8 Yo~
bys — dys OP ys
oy
xotan| % 7
Page 33Gauss Forward Interpolation Formula:
“To interpolate the value of a function in the middle of the tabular -values for equi-spaced data we use Gauss
forward interpolation formula or Gauss forward difference formula
Gauss forward interpolation formula ot Gauss forward difference formula is
OEEDED 54h
1-4) 52 y 4 @+)PO=2)
+ BOND 52 yy PED 9% yy
yf) = Yo + POY:
Here p
®
h= interval of difference
x = where we have to estimate the value of the function
fas
Note: For Gauss forviard interpolation formula, 0