0% found this document useful (0 votes)
25 views16 pagesNumerical Integration
Uploaded by
Soumajit DasCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
25 views16 pagesNumerical Integration
Uploaded by
Soumajit DasCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 16
Numerical Integration
§6.1. DEFINITION : NUMERICAL INTEGRATION/MECHANICAL
QUADRATURE
If we want to integrate a function within a definite limit we have to inspect
whether the function is continuous in the closed interval bounded by the limits
given. If however the integrand f(x) is not analytic then we have to find the
function from the values given between the limits and then integrate. Numerical
integration is the process of computing the value of a definite integral from a set
of numerical values of the integrand. This process is sometimes known as
mechanical quadrature.
Geometrically, we can interprate in the following manner the mechanical
quadrature formula as :
We first select a set of interpolating points or nodes x9, x), .. x, in the interval
of integration [a, b]. Then we calculate f(x,) (i = 0, ..., n). The curve is replaced
by the nth degree parabola y = (x) which passes through the (n + 1) points
n). Then the area bounded by the parabola, the x-axis, and
a, x = b is the value of the integration.
&» y) @=0, 1
the ordinates x
OS => ee
A general Numerical Integration formula
Suppose f(x) be known for a set of (n + 1) points xp, xj, ... x, and yj
(i = 0, 1, ..., n) be the corresponding values of f(x). If (x) be an interpolation
polynomial of degree less than or equal to n, which replaces f(x) on the set of
interpolating points x; (i = 0, 1, ..., n) ie. flx;) = O(x;) (i = 0, ..., m) and we can
write fx) = 0(x) + Ry).
Then integrating this with respect to x within the limits a and b, we get.
[Preode = fPo@d + PPR de... 6.1)
242
- CHAPTER SIX_O NUMERICAL INTEGRATION 243
Now the error term due to mechanical Quadrature is
De eee)
Sr an dk (a< & <5)
where w(x) = (x — X9) (W— x4)... (x = x,)
We replace the interpolating polynomial by Lagrange’s interpolating
polynomial as
_ S__w(x)y,
> Zoe)
b wy; pane)
Hence {? flx) dr = este d+ Prune mae ~ 6.12)
ea
wy, wh ay
ia Ce oH, Yi
a dx
eo iF @ =e
Opals 7(6¥1-3),
w(x)
Since we know that ead os
w(x) w(x)
= ie = W'G). eo | waa
w+ (6.1.4)
Now Epa = fw) oe dr... (6.1.5)
Now if we take f(x) = 1. Then, E=0
+ /™'(E) = 0 forn2 Land T= f’lde=b -a
The numerical values of the ordinates y; = f(x) (= 0, 1, .... n) are calculated
im such a manner the round off errors lie within the limits + 2 then’the error in
calculating the integral lies within the limits + 2 1H{"l, So to minimise the error
e i - * . (n)
we have to choose such a quadrature formula which will minimize S°1H{"l
Now it is evident that 1H) = [sue =b-a
& 244 NUMERICAL ANALYSIS AND STATISTICAL METHODS
Hence the least possible vaue of the left hand side of the above expression
will be b - a.
§6.2. NEWTON COTE’S FORMULA (open and closed type)
A mechanical Quadrature formula will be called closed or open according
as the limits of integration are used as interpolating points or not.
Sometimes it is more easier to calculate the value of integration by breaking
the whole interval (a, b] into number of subintervals. (a, ,, aj) (j . mJ then
we compute the values in each subintervals and add them.
Now for Newton cotes formula (closed type) we are taking gy X1 --) X, Which
x, %
0
are equally spaced i.e. xy = a, x, = b, = hand x,= x9 + th, i=0, ...n
Then, we have already find that
ffs) ax = ea
yu”
x)w Gy! ugh,
(nm) = (>__w@)dx_
Bae A Ge=sinics)
[o (x=39)(@-x,) F
90 = %)-%, = 19) —¥_ VO} = Fy) =%,)
Now if we set x = xp + sh. Then,
(% 4. Xp) «1. (= x) = AM s(s — 1)... (5 — 1).
Wc) = he iG — 1) 223). n=)
= he 1)! in - a
Hence Hi") =~ yi
a: (cue Hee ve ae ”) as
= (b-a) H”
Now if we substitute &(”
(n) _ (Dt ns(s— oe 18(S—1)-+-(s—n)
then. k{”) = zolas 6 Oras
The error term will be
nh! (s—1)-+-(s —
Byay = hy =O putt Eds, CHAPTER SIX NUMERICAL INTEGRATION 245
The integration formula due to Newton Cote’s (closed type) formula is given
as
T= (b~ a) YKy, v6.2.1)
Fa
where k( are known as Cote’s coefficient. These co-efficients are independent
i
and can be calculated for different values of n and i. Now
ye a
i" t-a
Hf”) = Gso8
b-a
=0
Again we can easily prove that,
aD he (es0) 1. 1);
By using the above two relations we have the table for Ki) as.
Table for k\”)
n\i 0 1 2 3 4 3 6
ik ay
1 2 2.
sly
2 6 = t
ae a
: 8 8 3 3
4 gle 32 42 32 Te
30 90 90 90 90
19 7 50. a
eB 288 2 288 ae 288 288
Paley 21Gee p27 AL
o 840 840 840 8 840 Hg 840
(a) Deduction of Trapezoidal Formula
Now, if we apply Newton-cote’s formula with n = 1, we can deduce
Trapezoidal formula.
H™( =0, 1) n= 1 will give
<1) A s(5—1
H} = Girt eas = CALs ~ Ns = ¢
i!
Dee hel a(s=1) ss, 98 i
BY = Sor hb Gan = hhsds =A] =f. 246 NUMERICAL ANALYSIS AND STATISTICAL METHODS
: yy Mea ee es el
Hence Ip = Hf)¥9 + HiPy, = 590 + 31 = 300
= ep SE—DP'O __
The i of precision is 1.
Composite Trapezoidal Rule :
[Prode = fdr + fPfarde +. + f pode
‘0 i] Sn
b h
J fxdax =: 00+) + £0, + yo) + 40, +4).
ly $= Ly + 2yy + + 2yny + Yad
I= Alyy +204 + yp + +p) +99
Similarly Error term
h
ae =— 75 0-41 - 290 +) - £00 = 2y, + yp) -
h
~ 3 0n-3 ~ Waa + Yn) —
and ES = BE=-2b4 +9, — Oo + Ynav]
(b) Deduction of Simpon’s one-
ird Formula
+ y,)
h
{70r2 -
w
1 on -h
4 £6) = - TLV - 2% + 4]
a 40, ey,)
. (6.2.2)
2yn1 + Yn)
.. (6.23)
Now if we apply Newton-cote’s formula with n = 2, we can deduce
Simpson’s one-third formula.
H((i = 0, 1, 2) n = 2 will give
Hibs - Dds
‘Therefore Jefe dr = Hy, + Hy, +
5
E, =- &f"6) where a< € 4¥n + Ine)
=~ Ebua + Ynay — 400 + Yn) + 701 + Yn) — 802 + Y4 + Yn-2)
+ 8()3 + ys + + Yq-a) [for n 2 6] .. 6.2.5)
(c) Simpson’s Three-eighths formula :
-a
This is a four point formula, For n = 3 we can write h = 2=4 and can
deduce Simpson’s three-eighth formula,
HO n=3.
@)_ CIR bh B ae 3h
HP = SPB fp - Ue - 2(s - 3)ds = BI
(3) _ 9h 3)_ 9h 4403) _ 3h
Biter sg Ua
"3, = Ho @), 3) 3)
Hence Sif = HD) +H®y, +O yy +HYy,
3h 9h, , 9h 3h
=p yt gut gt gs
= 2, + 3y, + 3y) + y3] + 6.2.6)
and the error term can be written as,
5
B= ae AM (a< €)
py oe gee,
= Tp PSt%g + hI + 6 + ~ ST I”
ee gn OMe a
0
2 >: 6
Ou” Qu Gy OD
+ {yo + 2h) 9 + Meer
Cie
ty + Guy + Og) + boy + Gi) 96 + ae Ce
ome
+ S{yq + (Sh) yp +...) + (yg + COh)y! + YO
Be Tom 7 evi
=~ tap" %6 — a0" INGO)
{neglecting the eight and higher powers of h]
Ege pri) 2 AG,
MY 140" ‘0 Ti 0
J:
= me) where a =x) < €