0% found this document useful (0 votes)
54 viewsNumerical Integration
Uploaded by
Gyanendra RokaCopyright
© © 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)
54 viewsNumerical Integration
Uploaded by
Gyanendra RokaCopyright
© © 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
/ 29
Unit 5: Numerical Integration, (8 hours)
SA
5:2,
3.3
5.4
5.5
5.6
Integration
Definite integral and area computation
Trapezoidal rule
Estimation of errors
Simpson's rules
Estimation of errors Chapter
Numerical Integration
Definite Integral and Area Computation
‘The area bounded by regular plane figure such as triangle, quadrilateral,
cirele ete. can be calculated by using formulae. But the area bounded by the
curve of a function cannot be calculated by such method. The area bounded
by the curve can be calculated by using definite integral.
Lety = f(x) be a curve and if F(x) is an integral of f(x), ie, f f(x) dx = F(x),
then the difference of the value of integral F(x) for two different values of
independent variable x, ie. F(b) ~ F(a) for x=a to x= b is called the definite
.
integral of f{x) over the interval [a, b] and denoted by f f{x)dx , Le.
Jf fox)dx = FQ) — F(@), the values x = a and x = b are respectively
called the lower and upper limi
y
=f) 120 ...Applied Mathemati
“To calculate the approximate area bounded by CUrVe, We Use ditfeey,
Here, we discuss trapezoidal and Simpsons rules. Aig,
5.1 Trapezoidal Rule
Let f be defined on the closed interval [a, b]. Let us consider thy
enclosed by the curve f(x), the ordinates
= b and the x-axis ay
in the following figure. 8 show
y
fs) fx)
Hex 1055-1)
D co) 4x,
Al ax | ax | ax | ax | ax | ax, | dx | ax
Ls x
mora xy a Mob
Letus divide [a,b] into m equal parts, each of length Ax =
Hoy Ky Xap ons Xe ty Xe Her tx), fh
crite be pea nego a
= bythe pi
Hotei) fsa) se
a 1 Xe Tespectively. By joining tt
consecutive points f), fi), fs) On the graph, we get a trapennids. Te
‘area under the curve is approximated by the sum of the areas of these ®
trapezoids. e eta
Here, area of trapezoid ABCD (Ist trapezoid) in the figure
= 7 sum of parallel sides * Perpendicular dist. between parallel sid®S
1
7 Ul) + a] ax
‘Now, area under the curve = sum of n trapezoids
ji :
= laa) + x) dx + Fox) + A a ot Le 89
-z~A ‘Numerical Integration...121
Usa) + 20081) + 2864-1) Fxg] AR
bea
OE fle) + Aba) + + 20a) + fhe]
reef) x =F lc) + 2x1) + 202) +. + 2A 1) + AE)
h
anh aayan=iyor2y $292 +2 + Ya
Estimation of Error
The trapezoidal error or error relative to trapezoidal rule is the difference
between the actual value of the integral ffx) and. the coresponding
approximated value obtained from trapezoidal rule, With the application of
techniques of calculus, itis estimated in the following way:
Erin pei le! 20, here Mi he argent of Fn he
ceed traf,
Evaluate using Trapezoidal rule {- 42%z,n= 4. Also estimate error.
1
Solution:
Here,n=4,a=1,0=2
bea 2-1
25
Here, dividing the intervals (1, 2] into 4 equal sub-intervals of length
0.25, we get 1, 1.25} [1.25, 1.5}, (15, 1.75} (1.75, 21.
1
“The values ofthe funtion y= (8) =p a the end points ofthe sub-
intervals are given below:
End poins [gel pets [eS [ap=ts Tue?
w= 05 |, o.4444 | yo=0.4000 | y= 03636 | ye= 0333 122... Applied Mathematics
Now, by using trapezoidal rule, we have
2
SE Poor rapt 29,499
1
125,
TY (05+ 204444 + 04000 +03636 44355
= 0.125 (0.5 +2 * 1.208 +0333)
125 x 3.2493,
4061625,
2
Tex MPO =TE yt
M-= Maximum value of |" (x)
|erbap | iat ates nah
Tpeg?* 0.25 = 0.0013,
2
Boat ing acl ale negra f= tn
1
absolute error from the actual value,
Solutio
Here, n= 4 and length of each sub interval (h
'm order to divide the interval [1, 2] into four sub-intervals, we have!
Consider ive pints x= 1, xy = 1.25, xp = 15,5 = 15,5472 a
Evaluating the values of fx) at these points we get the following wt
Exdpoinsot | xy=1 | x,=125] amis we?
sibineral
tw=3 | 190000 | os4oo0 | ossaas | 0.32650 | 028°
By using the trapezoidal rule, we have
UMxo)+ 2fx1)+ 209) + 2ft5) + faa] Numerical Integration...123
025,
[1.00000 + 0.64000 + 0 44444+0.32650) + 0.25000}
0.125 [4.07188] = 0.508985 = 0.509 (rounding off the digits.)
2
To estimate the error, we should find the actual value of integral f°.
Now, error
" rade
Using raezidal ae evaluate 8,
When n= 2, h= (0.5; and three points to be considered
are xp = 0, x; = 0.5 and x, = | and the values of the function at these
points are tabulated below:
end points [xo=0 |xi=05 | m=
1 08 05
1
yee
ee
S11 416405)
715 Ans
A
Approximate the valu using trapezoidal rule for fe dxyn=2.
1
Solution:
Here, h | 128 Applied Mathematics
Clearly, three points - 1,0, 1 divide the interval [1,1] ino
crrimerale cach of length h= 1 The values ofthe function se 2!
ona
1 are shown in the following table.
End points |xo=—1 x=0 x= 1
Vee el eat Pea
By trapezoidal rule, we have
J eaed
Je dx=F yet 291+ ya)
aflernied]
1
Lee+2+1
(e+1F
ales
| BITS
i
Using rapenada rate evaluate ix xn =2.
Solution:
Here, n=2,,
values of the function f{x) = sin at these points are given,
a ~
x! ai
End poins | xy=0
fix)=ysinx | 0 084 1
ing trapezoidal rule
Accor
2
JP? vane ox Bae) +20: + faa 32°F 0+2 «084+ 1)
Ease
= 0.3928 x 2.68
= 1.052 Ans
Using trapezoidal rule, evaluate f (Gx? 4x) dx,
Solutio
Here,n=3,a=0,b=3, 0)
Let us consider four points x9 = 0, x;
5X27 2, X53. The values of
the function y = 3x2 4x at these poins are tabulated below.
Endpoints | x9=0 | x
x53
y o [-1 | 4 15
“According to trapezoidal rule
2 b
J OX 4n) dx = (0291 +292 +9)
(0+2C1) 424415}
421 =105 Ans
oh 248415) 9521 =105 Ans.
Given 1=f' = ax, n= 4, estimate the value of [using Trapezoidal rule.
Solution:
Here, the length of each sub-interval is “Z = 1. In order to
Aivide the interval (0, 4 into four subintervals each of Tenth b
we have to consider five pons % = 0, x)= 1," 2,1" 3,34 4
The values of the function y = f{x) = x° at these points are tabulated
below,
End points_[s0=0
ysx 0 1 8 27 64
m2 [es [ued 1325. Applied Mathematls —
Non: bysing te topo rue, 6 Sve
1 few
aBiyot20rty:t yt ¥d
Biyyt2yt2ye+20 + yd
[0+2(1+8+27)+ 64]
1
ji+6
1
=7* 136=68 Ans,
PRI
ax,
Evaluate using trapezotdal rule the integral f jy,+=5-
Soluti
Here, n=5,a=0,
bea
Here, 4 sub-intervals of length 0.20 each should be considered, s0 we
hhave to consider five points x= 0, xi = 0.20, x2= 0.40, x5 = 0.60, 24
O80 andy 1. The al of finn y= 9 = heat es a
acti i
Endpoints | 3 =0 | x,=020 | x=040 ] x=060 | x=080 | =
0.71429 | o.aso0 | 0.5556 | 050000
Now, using trapezoidal rule, we have
dx oh
eiax 2 Dot 201+ ty +¥)+ 95]
029,
= "511.0000 + 2(0.83333 + 0.71429 + 0.62500 + 0.55556) + 0.50000]
0.1 [1.50000 +2 2.72818)
0.1 [1.50000 + 5.45636)
=011 x 6.95636 = 0.695636 Ans, ampate an approximate value of { (+x) by using the composite
trapezoid rule with three points, Them comparing with the actual value
afthe integral, find the error.
Solution:
‘There are three end points ofthe intervals, son=2.
Here, a=0,b=1,
‘ are 0, 0.5, | and values ofy a these points are
given in the following table:
x 0 [os [a
y= [1 | 08 | os
Using the composite trapezoidal re, we Have
Sites B fyo+2yity2}
95 eox08+0s}
=ors
Actual value= fuses (tax), =
171854 ~ 0.775 = 0.104 Ans.
Error=
Using the trapezoidal rule, compute J2 ex 1) axwith 4 intervals. Find
{he absolute error of approximation from its actual value,
Solution:
Herea=0,b=2,0=4
5 and 5 points to be considered are Xo = 0, X1 =
La 4
0.5, x2= 1.0, x3 = 1.5, X42
‘The value of integrand corresponding to each points are tabulated
below.
fesiees | ocn|nes wpa [ras
[y=2e-1 Sparse 2 128 Applied Mathemat
‘Now, using trapezoidal rule, we have
fee) éx=4pot201+ ty) +d
°
Bp 142C05+1+35)+7]
0.25[-1+2*4+7]
0.25% 14=35 Ans.
eed
°
10
2353-35-07
lane sing compote traperidal ule the integral sn xin =4
Solution:
Sie, n 4g ofeach biter) = P=2 BO
the fv oats io be coe 9
3/4, xq = 1 which divide the interval (0, x} into four sub-intervals
valu
the values ofthe function, we get the following table
‘We can divide the interval [0, n] into 4 sub-intervals ‘of equal length
cach F namely [0, 14), [n,n], [2,34], [3n/4,
End points of Jxs=0 =n [uaa [or
subinterval
Ax) = sin x
int Pua
0.00000 |infa [1.00000 [raja [0.00000
Using wapenoidal rue, we have
eoten
Pais Us) +2) + ft) + ft) + fd]
=F) to00000 +2
. @)wi+ 1)= 1.896 Ans,
) + 0.000001 een Nett
Apply rapezoldal rule to evaluate the integral oT aforn= 4.Give
the answer correct up-to 3 places of decimal,
Solution:
Since n= 4, length of each sub-interval (h) = ne) 205 we
have to consider five points x) =-0, x1 = 0.5, x2 = 1.0, x5 = 15,
X4 = 2.0 which divide the interval [0, 2}, into four sub-intervals,
Evaluating the values ofthe function, we will gt the following table:
Endpoint ]x=0 | x1=05]xg=10 |x 15 | x=20
1.00000 | 0.94117 | 0.50000 | 0.16494 | 0.05662
981 momraset710 5000 +0165 +085
= (0.25) [427104]
= 1.06776 ~ 1.067 (by chopping off extra di
106776 = 1.068 Ans.
‘Approximation from its actual value.
Solution:
Here, n=4,a=0,b=1
Hrs subintervals of eth 025 each sold be cose we
have to considered five points x= 0, x1 = 025, x2 = 0:50, 33, 130... Applied Mathematics
y= 100. Te value of the function y= £0) Tx at hese points ae
{abulated below.
Endpoint [9=0 _[x1=025 [2-05 |x9=075 |xy=1
7 }1.00000 [0.80000 [0.66666 }0.57143 [0.50000
YA Tex,
Tow, by using vapezoial le, we have
:
iB abiyg tay + 2904295 +94
2B yrayitntytyd
=0.125{1.00000 + 2(0.80000 +0.6666640.57143) + 0.50000]
0.125 [1.00000 + 2 x 2.03809 + 0.50000]
128» 557618 = 0.69702
piey ~ log +)
~og (141) —Hog (+0)
=log2-0= 0.9314
Truc ror = actual vale pprorimate vale)
=|0.69314--0.69702| = 0.00388 Ans.
IE
‘Now, actual value of integral
Using trapezadal ue evaluate [75 for n= 4 correct 3 plas of
decimal,
Solution:
Since n= 4, length of each sub-interval(h) = ©=2) 22
have 10 coir five pins x4 = 0, x = 0, x= 10, x = 15
X4 = 20 which divide the interval [0,2], into four subintervals
Evaluating the values ofthe funtion, we will get the following able:
End point | xo=0
05,we
x= 10 x=20
fa)=—1, | 100000] 0.94117 | 0.50000 [0.16894 | 0.05882
Tt Pees
By the trapez il rule,
‘2_dx_ oh
Joa Fa +2) +209) +20) +
7 1.00000+2(0.941 1740.50000 + 0.16494) + 0.05882}
= (025) (4.27104)
(06776 = 1.067 (by chopping off extra digit)
By rounding off the
pe
OT 4d 106776 = 1.068 Ans.
erro
Solution:
Here, n =4 and length of each sub-interval (h)
Inorder to divide the interval [1,2] into four subintervals, we have to
consider five points xo = 1, X1 = 1.25, Xp = 1.5, 43 = 1.75, x4 = 2.
Evaluating the values of f{x) at these points we get the following
table
Endpomwor | w=! funt2s [w=15]u-175] =2
sxb-inerval
1 Fro0000 | oto00 | oaases | 022650 | 025000
fore.
By using the trapezoidal rule,
2 tsa) + 2x) + 280) + 2F 09) + x]
9.25 1 pongo + 2.64000 + 0444444032650) + 0.25000]
125 [4.07188] = 0.508985 ~ 0.509 (rounding off the digits)
‘To estimate the error, we should find the actual value of integral
LP atei=05
rf
F 5 so eat ane oneal | $F -[-
Now, error =0.509—0.5]=0.009 Ans. 10,
setpoint a
Using trapeznidal rule compute xa with 4 intervals Find
absolute eror of approximation from its actual value,
Using trapezoidal rule compute fs with 3 intervals, Find yy
absolute error of approximation from is actual value,
Using trapezoidal rule compute fea with 4 intervals, Find the
absolute err of approximation from its actual value.
Using tapezia rae compute J Gx? 4x) dx with 3 iterate Fag
the absolute error of approximation from is actual value,
Using trapezoidal rule, evaluate
dx
with n
- Estimate the error of
Approximation from its actual value.
alte sng pez lth intepral sin xd, n=
‘Approximate the value using trapezoidal rule for Set dxn=2
Fvaluate using Trapezoidal le f 2S y= 4, Alo estimate erat
1
Using trapezoidal rule evaluate f
THES for n= 4 comes to 3 plac of
decimal. : we
‘Numerical Integration..133,
NSWER SHER!
2.375, 0 3.42, 0.667
4.105, 15
jt
2e
7, 425, 0.25
5, 0469702, 0.00388 6.1.896 7.1082
49,0.4061625, 0 10. 1.068
52 simpsons nue
To approximate the integral “9
b
J fs) &x5 we take a set of three
points and fit with a part of the
yarabola whose equation is of the
formy= Ax? + Bx +C.
The area under the curve from
tox his given by
4
, ‘as Be oy]
Seats per ode -[AE BE sox] |
= Fl2ah’ + 6C].
Since the eurve passes the points (-h (0, and (h 2) $0 We have
‘Yo= Ah?~Bh+C, yi ‘and y2= Ah? + Bh +C
yeoyi= Aly Bh and ys ys= Ah? + Bh
ra
Adding, yo* ya 2ys
Hence, area = Hit ys—2)+ 69
b
Hoot anit]
Proceeding similarly to the successive pieces of the curve y = f(x) between
x=.and x= taking the sum ofall, we get the Simpson's ule as|
Yoel) #202 + Ye oct Yor]
®
St) dx Ago ty +401 +54
>.
where nis even and h rule its to be noted that n should be even, Lethe
Ma PrP megaton must be divided info an even number of equal gp,
intervals.
Estimation of Error
‘Simpson's error relative to Simpson's 3 1 rule is the difference between actual
b
value of the integral f(x) dx and the corresponding approximated value
a
chained from Simpson's rule
Error (Simpson's 3 1 rule
where M is the largest value of f"()
inthe closed in [01
Note: Simpson's re canbe used only when the numberof sub-inerval is
4
‘even, But the Simpson's’ rule can be used when umber of sub-interval i odd
as well as even which is given as follows:
Sf) d= Poy 39.3924 90) b
Evaluate using Simpson's rule f° ‘3, n= 4, Also, estimate error.
ee
Solution:
Since n= 4,length ofeach sub-interval = "8 = 3=1 _ 95,
‘Here, dividing the interval [1, 3] into 4 ‘sub-ir ls, we get
UWS} 015.2, 2,25),(25,3), | ee nee
‘The value of
End Points
1
‘¥F atthe end points of sub
METS] ued | wes] med
fed= zt] F601] fe) 0.4444 m3) =0.25| ts) = 016] x) = 0-1"
By Simpson's rule, ee ie
Here, (9) =P
‘M= maximum value of| fx) |in [1,3] occurs atx:
ie Melon
Now,a=1,b=3,n=4
penta Gan 20 _ 32120
T80n? =" 1804)" “180% 256
f
compa F274 Simp ro A compart
0
ie
ee,
n=4,
Then, h=P=2=45%
125
“The five points xy = 0, X; = 0.25, x2 = 0.5, x3 ='0.75 and x,~ 1 divide
the interval [0, 1] into four sub-intervals of length 0.25 each,
bt gto nal ie gin
The values of y =
following table
End Point [%=0 [x=025 [m=05 | x5=0.75 | xy=1
1 oss [08 064 [os
‘Simpson's rule,
h
3
9251 44x94 4208 +4% 064405)
Lot 4yi + 292+ Aya + Yad
<2 9424-0785 An 136 .. Applied Mathematics.
1
Actual value= 3-32 dx
0
at
={un"x
* = tan! ~tan0
rating Pasan2
oo rash ppesinta vale wig Sings rule wp 3 pl of
decimals.
(i Calculate an error bound for the approximation. Determine the
error. Is it within the bounds?
Solution:
@ Here,
In order to partion the interval {- 1, 4] into two subintervals we
Ihave to conider three 1,31 1.5 and x5=4, The values
ofthe funtion fx) =x’ at these points ae tabulated below.
End points aad
fx) =x 64
By Simpson's rule, we have
J
a
pe
We = fl) + fC) + 2)
25
Be i+4x3.375 +64)
= 73 x165 =63.750
f'Q@)eae
£"G)= 6x
£"@)
£%(@):
M=|f"(3)|=0, forall xin (1,4), Rarer tegration 7
Ani ‘bour (b—
‘pper bound for enor is M PAP 9, 4+ DE
0.
180 (2) -
‘The error =| exact value ~. -Approximates value| _
=|63.75 -63.25|=0
Thus the error = 0= error bound. Hence, the errors within the bound. Ans.
‘Simpson's rule to, f
pps Simpson's rule o approximate the value of fe" dx dx with n=3.
Solution:
Here, f(x) =e énx
End point maa
y=e" fax 75.6894
So gang
‘Now, by using Simpson's rule, we have
f eeoxds = By tay. 43+)
3
[0 +3 x 5.1214-+3 x 22.0659 + 75.6894]
= F115.3642 + 66.1977 + 75.6894]
3 cis.asia S896 Am,
Using Simpsor
Solution:
i oul SR acu
02.
2n=2 eho
x; = 0.1, x2= | divide the interval [0, 0.2] into
two sub-intervals. The values ofthe integrand y =_[T—2x? at each of
the end points ofthe sub-intevals are given below.
End points [=O] x50.
YouT| yi =0.989040494 | ya = 0959166305
y=Vi-2e
Here a= 0, b: ou
‘Three points x9
x02 138 ..Applied Mathematics
1
By Simpson's 5 rule,
: h
i =x? da=3 [yo + Ay + Yad
= +4 x 0989949494 + 0.959166305)
=x 5918964281 =0.197298809 Ans,
02
Fi he approsinate vate of f YTB? dn =2, wig Simp
o
rule,
Solution:
Herea=0,
Thre points xy
x4 = 0.1, x2= 1 divide the interval [0, 0.2} into
‘wo sub-intervals. The values ofthe integrand y =-J1—2x? at each of
the end points of the sub- intervals are given below.
End points [xy=0] _xi=0.1 x02
y=Vi-2¢ |y=0] y= 0.989949404 V2= 0959166305 |
Dimon
f T= Ae dx =F Lye ys ya]
1
="7 Ul +4 0,989949494 + 0.959166305]
% 5.918964281 = 0.197298809 Ans.
Using Simpson's rule, evaluate 2.
Solution:
Here,n=4,a=0, ‘Numerical Integration...139,
‘The end points of 4 subvintervals of [0, 1} are x= 0, x1 = 0.25, x2
055, x3 = 0.75, x4 = 1.00, The values of y =
1
Tegat each of the end
points ofthe intervals are given below:
End point |x=0 [xj =025 [xp
15] x= 075 ]xy= 10
yay | 100000] 80000" 0.68667 Jo37143 fos0000
By using Simpson's rule, we have
3 U0 + Ay + yp + ys + yg)
025
“5 [1.00000+4x 0.80000 + 2 x 0,66667+ 4 x 0.57143) + 0.50000]
693255 Ans.
:
Aerie 4x ing Simp leith =
Solution:
Here, a
b=2,y=fy=2,h
‘The end points are 0, 0.5, 1, 1.5 and 2.
x Jo| os [1] 15 [2
ye2 | 1 | 14142 [2 | 2.8284 | 4
1
Now, using Simpson's 3 rule, we have
2
fre Fouran 42, tay tyd)
95 (1 + ay. 4142-+ 200 + 450.8288)
05
25,9704 = 4.3284 Ans
Using Simpson's rule, evaluate ((2x—1)*dxym= 4.
.
Solution:
Here, a=0,b=2,n=4 fed Mathematics
sare 0, :
ofthe fancion are shown nthe following tbe
me oO 05 i 1s 2
x
yeax-¥| 1 & i 4 9
1
‘Now, by using Simpson's 3 rule, we have
h
festa By ay ty) +2 + yd
oy saorgeaxt es)
8 s16+249
'n order to patton the interval 1,5] into four sub-intervals, each of length
* we have to consider the five points x= 1,
Bvaluting the values of function y_
End points | x9=1
1
1
By Simpson's rule, we have
Sta
Be 4 +29 49,494
1
S3U44* 164.2% 81 44x 256+ 625]
BU +64 162+ 1024 +625)
1
* 1876 = 625.33 Ans, _———
an
Using Simpson's 4
le, eva ate: | ae ds, h=025,
Solution:
Hoe a0, 6~1,h=025,
Ths the endpoint ofthe intervals ae 0.25, 050, 0.75, 1
The values of the integrand at each end point of the intervals are given below:
= oT 025 [050 | 075] 1
y=Vi+2x? t 1.06066 | 1.22474 [1.45773 | 1.73205
7
Al, TERK 3 bo + 40H + yo) + 290-494)
=9F5 1+4(.06066-+1.4573)42m1.2247441.73205]
= 0.08333{1 + 10.07356 + 2.44948 + 1.73205]
2712 Ans.
Ie
igen trac Paaclac
ae
on jeaghitica htc gy bEPOE=O
pape pa eng uae
length b=, we fav o conser sve points x= 0, 2.33 =
aie
Basu Fae n ts noi) tte
points are tabulated below:
End point | x)= 0 xan
3 [272
4)=sinx | 0.0000 0.000] 8 1.0000] 2 | os000 | 0.0000
By Simpson's rule, we have
SGsins. dbx= Bs) + 4081) + 21082) + 4105) + 20a) + 46x) + A(x.)
=
