Numerical Integration Lecture
-
Upload
mohammed-hamdy -
Category
Documents
-
view
236 -
download
1
Transcript of Numerical Integration Lecture
-
7/30/2019 Numerical Integration Lecture
1/59
1
Advanced Computational MethodsNumerical IntegrationME753
-
7/30/2019 Numerical Integration Lecture
2/59
2
Introduction to NumericalIntegration
Definitions Upper and Lower Sums Trapezoid Method (Newton-Cotes Methods)Simpsons Rules Romberg Method Gauss Quadrature Examples
-
7/30/2019 Numerical Integration Lecture
3/59
3
Integration
Indefinite Integrals
Indefinite Integrals of afunction are functionsthat differ from eachother by a constant.
c
xdxx
2
2
Definite Integrals
Definite Integrals arenumbers.
2
1
2
1
0
1
0
2
xxdx
-
7/30/2019 Numerical Integration Lecture
4/59
4
Fundamental Theorem of Calculus
:forsolutionformclosedNo
:fortiveantiderivanoisThere
)f(x)(x)'F(i.e.,foftiveantiderivais
,intervalanoncontinuousisIf
2
2
b
a
x
x
b
a
dxe
e
F(a)F(b)f(x)dx
F
[a,b]f
-
7/30/2019 Numerical Integration Lecture
5/59
5
The Area Under the Curve
b
af(x)dxArea
One interpretation of the definite integral is:
Integral = area under the curve
a b
f(x)
-
7/30/2019 Numerical Integration Lecture
6/59
6
Upper and Lower Sums
a b
f(x)
3210 xxxx
iin
i
i
ii
n
i
i
iii
iii
n
xxMPfUsumUpper
xxmPfLsumLower
xxxxfM
xxxxfmDefine
bxxxxaPPartition
1
1
0
1
1
0
1
1
210
),(
),(
:)(max
:)(min
...
The interval is divided into subintervals.
-
7/30/2019 Numerical Integration Lecture
7/59
7
Upper and Lower Sums
a b
f(x)
3210 xxxx
2
2integraltheofEstimate
),(
),(
1
1
0
1
1
0
LUError
UL
xxMPfUsumUpper
xxmPfLsumLower
ii
n
i
i
ii
n
i
i
-
7/30/2019 Numerical Integration Lecture
8/59
8
Example
14
3
2
1
4
103,2,1,0
4
1
1,
16
9,
4
1,
16
1
16
9,
4
1,
16
1,0
)intervalsequalfour(4
1,4
3,
4
2,
4
1,0:
1
3210
3210
1
0
2
iforxx
MMMM
mmmm
n
PPartition
dxx
ii
-
7/30/2019 Numerical Integration Lecture
9/59
9
Example
14
3
2
1
4
10
8
1
64
14
64
30
2
1
32
11
64
14
64
30
2
1
integraltheofEstimate
64
301
16
9
4
1
16
1
4
1),(
),(
64
14
16
9
4
1
16
10
4
1),(
),(
1
1
0
1
1
0
Error
PfU
xxMPfUsumUpper
PfL
xxmPfLsumLower
ii
n
i
i
ii
n
i
i
-
7/30/2019 Numerical Integration Lecture
10/59
10
Upper and Lower Sums
Estimates based on Upper and LowerSums are easy to obtain for monotonicfunctions (always increasing or always
decreasing). For non-monotonic functions, finding
maximum and minimum of the function
can be difficult and other methods can bemore attractive.
-
7/30/2019 Numerical Integration Lecture
11/59
11
Newton-Cotes Methods
In Newton-Cote Methods, the functionis approximated by a polynomial oforder n.
Computing the integral of a polynomialis easy.
1
)(...
2
)()()(
...)(
1122
10
10
n
aba
abaabadxxf
dxxaxaadxxf
nn
n
b
a
b
a
n
n
b
a
-
7/30/2019 Numerical Integration Lecture
12/59
12
Newton-Cotes Methods
Trapezoid Method (First Order Polynomials are used)
Simpson 1/3 Rule (Second Order Polynomials are used)
b
a
b
a
b
a
b
a
dxxaxaadxxf
dxxaadxxf
2210
10
)(
)(
-
7/30/2019 Numerical Integration Lecture
13/59
13
Trapezoid Method
Derivation-One Interval Multiple Application Rule Estimating the Error Recursive Trapezoid Method
-
7/30/2019 Numerical Integration Lecture
14/59
14
Trapezoid Method
)()()(
)( axab
afbfaf
f(x)
ba
2
)()(
2
)()(
)()()(
)()()(
)(
)(
2
afbfab
x
ab
afbf
xab
afbfaaf
dxaxab
afbfafI
dxxfI
b
a
b
a
b
a
b
a
-
7/30/2019 Numerical Integration Lecture
15/59
15
Trapezoid MethodDerivation-One Interval
2
)()(
)()(2
)()()()()(
2
)()()()()(
)()()()()(
)()()()()(
22
2
afbfab
abab
afbfabab
afbfaaf
x
ab
afbfx
ab
afbfaaf
dxxab
afbf
ab
afbfaafI
dxaxab
afbfafdxxfI
b
a
b
a
b
a
b
a
b
a
-
7/30/2019 Numerical Integration Lecture
16/59
16
Trapezoid Method
)(bf
f(x)
ba
)()(2 bfafab
Area
)(af
-
7/30/2019 Numerical Integration Lecture
17/59
17
Trapezoid MethodMultiple Application Rule
a b
f(x)
3210 xxxx
strapezoidtheof
areastheofsum)(
...segmentsintodpartitione
isb][a,intervalThe
210
b
a
n
dxxf
bxxxxan
1212
2
)()(xx
xfxfArea
x
-
7/30/2019 Numerical Integration Lecture
18/59
18
Trapezoid MethodGeneral Formula and Special Case
)()(21
)(
...
equal)ynecessaril(notsegmentsintodividedisintervaltheIf
11
1
0
210
iiii
n
i
b
a
n
xfxfxxdxxf
bxxxxa
n
1
1
0
1
)()()(2
1)(
allfor
points)basespacedEqualiy(CaseSpecial
n
i
in
b
a
ii
xfxfxfhdxxf
ihxx
-
7/30/2019 Numerical Integration Lecture
19/59
19
Example
Given a tabulated
values of the velocity of
an object.
Obtain an estimate of
the distance traveled in
the interval [0,3].
Time (s) 0.0 1.0 2.0 3.0
Velocity (m/s) 0.0 10 12 14
Distance = integral of the velocity
3
0)(Distance dttV
-
7/30/2019 Numerical Integration Lecture
20/59
20
Example 1
Time (s) 0.0 1.0 2.0 3.0
Velocity
(m/s)
0.0 10 12 14
29)140(2
112)(101Distance
)()(2
1
)(
1
0
1
1
1
n
n
ii
ii
xfxfxfhT
xxh
MethodTrapezoid
3,2,1,0arepointsBaselssubinterva3into
dividedisintervalThe
-
7/30/2019 Numerical Integration Lecture
21/59
21
Error in estimating the integralTheorem
)(''max12
)(12
then)(eapproximat
tousedisMethodTrapezoidIf:Theorem
)(widthintervalsEqual
oncontinuousis)('':Assumption
],[
2
''2
a
xfhab
Error
[a,b]wherefhab
Error
dxxf
h
[a,b]xf
bax
b
-
7/30/2019 Numerical Integration Lecture
22/59
22
Estimating the ErrorFor Trapezoid Method
?accuracydigitdecimal5to
)sin(computetoneeded
areintervalsspacedequallymanyHow
0
dxx
-
7/30/2019 Numerical Integration Lecture
23/59
23
Example
intervals719
00437.0
)(
00437.0106
102
1
121)(''
)sin()('');cos()(';0;
)(''max12
102
1error,)sin(
52
52
],[
2
5
0
h
abn
hh
hErrorxf
xxfxxfab
xfh
ab
Error
thatsohf inddxx
bax
-
7/30/2019 Numerical Integration Lecture
24/59
24
Example
)()(2
1)(),(
,allfor:
)()(2
1),(
)(:computetomethodTrapezoidUse
0
1
1
1
11
1
0
3
1
n
n
i
i
ii
iiii
n
i
xfxfxfhPfT
ixxhCaseSpecial
xfxfxxPfTTrapezoid
dxxf
x 1.0 1.5 2.0 2.5 3.0
f(x) 2.1 3.2 3.4 2.8 2.7
-
7/30/2019 Numerical Integration Lecture
25/59
25
Example
9.5
7.21.2218.24.32.35.0
)()(2
1)()( 0
1
1
3
1
n
n
i
i xfxfxfhdxxf
x 1.0 1.5 2.0 2.5 3.0
f(x) 2.1 3.2 3.4 2.8 2.7
-
7/30/2019 Numerical Integration Lecture
26/59
ExampleUse the trapezoidal rule to estimate
8.0
0
5432 400900675200252.0 dxxxxxx
Solution
%5.89
1728.0
2
232.02.08.0
2
)8.0()0()08.0(
2
)()()(
t
ff
bfafabI
-
7/30/2019 Numerical Integration Lecture
27/59
ExampleUse the multiple-application trapezoidal rule for n = 2 toestimate
8.0
0
5432 400900675200252.0 dxxxxxx
Solution
%9.34
0688.1
4
232.0)456.2(22.08.0
4
)8.0()4.0(2)0()08.0(
)2(2
)()(2)()( 210
t
fff
xfxfxfabI
-
7/30/2019 Numerical Integration Lecture
28/59
-
7/30/2019 Numerical Integration Lecture
29/59
Computer Algorithms for the
Trapezoidal Rule
-
7/30/2019 Numerical Integration Lecture
30/59
30
Recursive Trapezoid Method
f(x)
haa
)()(2
)0,0(R
:intervaloneonbasedEstimate
bfafab
abh
-
7/30/2019 Numerical Integration Lecture
31/59
31
Recursive Trapezoid Method
f(x)
hahaa 2
)()0,0(21
)0,1(
)()(2
1)(
2)0,1(R
2
:intervals2onbasedEstimate
hafhRR
bfafhafab
abh
Based on previous estimate
Based on new point
-
7/30/2019 Numerical Integration Lecture
32/59
32
Recursive Trapezoid Method
f(x)
)3()()0,1(2
1)0,2(
)()(2
1
)3()2()(
4
)0,2(R
4
hafhafhRR
bfaf
hafhafhafab
abh
hahaa 42 Based on previous estimate
Based on new points
-
7/30/2019 Numerical Integration Lecture
33/59
33
Recursive Trapezoid MethodFormulas
n
k
abh
hkafhnRnR
bfafab
n
2
)12()0,1(2
1)0,(
)()(2
)0,0(R
)1(2
1
-
7/30/2019 Numerical Integration Lecture
34/59
34
Recursive Trapezoid Method
)1(
2
2
1
2
1
3
2
12
1
1
)12()0,1(2
1)0,(,
2
..................
)12()0,2(
2
1)0,3(,
2
)12()0,1(2
1)0,2(,
2
)12()0,0(2
1)0,1(,
2
)()(2
)0,0(R,
n
kn
k
k
k
hkafhnRnRab
h
hkafhRRab
h
hkafhRRab
h
hkafhRRab
h
bfafab
abh
-
7/30/2019 Numerical Integration Lecture
35/59
Example on Recursive Trapezoid
35
errortheestimatethenR(3,0)computingby)sin(
:estimatetomethodTrapezoidRecursiveUse
2/
0
dxx
n h R(n,0)
0 (b-a)=/2 (/4)[sin(0) + sin(/2)]=0.785398
1 (b-a)/2=/4 R(0,0)/2 + (/4) sin(/4) = 0.948059
2 (b-a)/4=/8 R(1,0)/2 + (/8)[sin(/8)+sin(3/8)] = 0.987116
3 (b-a)/8=/16 R(2,0)/2 + (/16)[sin(/16)+sin(3/16)+sin(5/16)+sin(7/16)] = 0.996785
Estimated Error = |R(3,0) R(2,0)| = 0.009669
-
7/30/2019 Numerical Integration Lecture
36/59
36
Advantages of Recursive Trapezoid
Recursive Trapezoid:
Gives the same answer as the standardTrapezoid method.
Makes use of the available information toreduce the computation time.
Useful if the number of iterations is not
known in advance.
-
7/30/2019 Numerical Integration Lecture
37/59
37
Simpsons Rules
Simpsons 1/3 Rule
Simpsons 3/8 Rule
-
7/30/2019 Numerical Integration Lecture
38/59
SIMPSONS RULES
More accurate estimate of an integral is obtained if a high-order polynomial is used to connect the points.
The formulas that result from taking the integrals undersuch polynomials are called Simpsons rules.
-
7/30/2019 Numerical Integration Lecture
39/59
Simpsons 1/3 Rule
This rule results when a second-order interpolatingpolynomial is used.
2
where)()(4)(
3
)())((
))((
)())((
))(()(
))((
))((
,andLet
)()(
210
2
1202
10
1
2101
200
2010
21
20
2
2
0
abhxfxfxf
hI
dxxfxxxx
xxxx
xfxxxx
xxxxxf
xxxx
xxxxI
xbxa
dxxfdxxfI
x
x
b
a
b
a
After integration,
-
7/30/2019 Numerical Integration Lecture
40/59
Simpsons 3/8 Rule
This rule results when a third-order interpolatingpolynomial is used.
3
where)()(3)(3)(8
3
,yieldsThis
)()(
3210
3
abhxfxfxfxf
hI
dxxfdxxfI
b
a
b
a
-
7/30/2019 Numerical Integration Lecture
41/59
41
Romberg Method
Motivation
Derivation of Romberg Method
Romberg Method
Example
When to stop?
-
7/30/2019 Numerical Integration Lecture
42/59
42
Motivation for Romberg Method
Trapezoid formula with a sub-interval h gives anerror of the order O(h2).
We can combine two Trapezoid estimates with
intervals h and h/2 to get a better estimate.
-
7/30/2019 Numerical Integration Lecture
43/59
43
Romberg Method
First column is obtained
using Trapezoid Method
dxf(x)ofionapproximattheimprovetocombinedare
...h/8h/4,h/2,h,sizeofintervalsmethodTrapezoidusingEstimates
b
a
R(0,0)
R(1,0) R(1,1)
R(2,0) R(2,1) R(2,2)
R(3,0) R(3,1) R(3,2) R(3,3)The other elements
are obtained using
the Romberg Method
-
7/30/2019 Numerical Integration Lecture
44/59
44
First ColumnRecursive Trapezoid Method
n
k
abh
hkafhnRnR
bfafab
n
2
)12()0,1(2
1)0,(
)()(2
)0,0(R
)1(2
1
-
7/30/2019 Numerical Integration Lecture
45/59
45
Derivation of Romberg Method
.. .)0,1()0,(43
1)(
2*41
)2(.. .64
1
16
1
4
1)0,()(
by R(n,0)obtainedisestimateaccurateMore
)1(...)0,1()(
2methodTrapezoid)()0,1()(
66
44
66
44
22
66
44
22
12
hbhbnRnRdxxf
giveseqeq
eqhahahanRdxxf
eqhahahanRdxxf
abhwithhOnRdxxf
b
a
b
a
b
a
n
b
a
-
7/30/2019 Numerical Integration Lecture
46/59
46
Romberg Method
1,1)1,1()1,(414
1),(
)12()0,1(2
1)0,(
,
2
)()(2
)0,0(R
)1(2
1
mnmnRmnRmnR
hkafhnRnR
abh
bfafab
m
m
k
n
n
R(0,0)
R(1,0) R(1,1)
R(2,0) R(2,1) R(2,2)
R(3,0) R(3,1) R(3,2) R(3,3)
-
7/30/2019 Numerical Integration Lecture
47/59
47
Property of Romberg Method
)(),()(
Theorem
22 m
b
a
hOmnRdxxf
R(0,0)
R(1,0) R(1,1)
R(2,0) R(2,1) R(2,2)
R(3,0) R(3,1) R(3,2) R(3,3)
)()()()( 8642 hOhOhOhOError Level
-
7/30/2019 Numerical Integration Lecture
48/59
48
Example
3
1
2
1
8
34
3
1)0,0()0,1(4
14
1)1,1(
1,1)1,1()1,(414
1),(
8
3
4
1
2
1
2
1
2
1))(()0,0(
2
1)0,1(,
2
1
5.010
2
1)()(
2
)0,0(,1
Compute
1
1
0
2
RRR
mnformnRmnRmnR
hafhRRh
bfafab
Rh
dxx
mm
0.5
3/8 1/3
-
7/30/2019 Numerical Integration Lecture
49/59
49
Example (cont.)
3
1
3
1
3
16
15
1)1,1()1,2(4
14
1)2,2(
3
1
8
3
32
114
3
1)0,1()0,2(4
3
1)1,2(
)1,1()1,(414
1),(
32
11
16
9
16
1
4
1
8
3
2
1))3()(()0,1(
2
1)0,2(,
4
1
2
2
RRR
RRR
mnRmnRmnR
hafhafhRRh
m
m
0.5
3/8 1/3
11/32 1/3 1/3
-
7/30/2019 Numerical Integration Lecture
50/59
50
When do we stop?
at R(4,4)STOPexample,for
steps,ofnumbergivenaAfteror
)1,(),( nnRnnR
ifSTOP
-
7/30/2019 Numerical Integration Lecture
51/59
51
Gauss Quadrature
Motivation General integration formula
-
7/30/2019 Numerical Integration Lecture
52/59
52
Motivation
nandih
nihcwhere
xfcdxxf
xfxfxfhdxxf
i
n
i
ii
b
a
n
i
ni
b
a
05.0
1,...,2,1
)()(
:asexpressedbecanIt
)()(2
1)()(
:MethodTrapezoid
0
1
1
0
-
7/30/2019 Numerical Integration Lecture
53/59
53
General Integration Formula
integral?theofionapproximatgoodagivesformulaethat thsoselectwedoHow
:Problem
::
)()(
0
ii
ii
n
i
ii
b
a
xandc
NodesxWeightsc
xfcdxxf
-
7/30/2019 Numerical Integration Lecture
54/59
54
Lagrange Interpolation
b
aii
n
i
ii
b
a
b
ai
n
i
i
b
an
b
a
n
n
b
an
b
a
dxxcwherexfcdxxf
dxxfxdxxPdxxf
xxx
xPwhere
dxxPdxxf
)()()(
)()()()(
,...,,:nodesat the
f(x)esinterpolatthatpolynomialais)(
)()(
0
0
10
-
7/30/2019 Numerical Integration Lecture
55/59
Example
Determine the Gauss Quadrature Formula of
If the nodes are given as (-1, 0 , 1)
Solution: First we need to find l0(x), l1(x), l2(x)
Then compute:
55
2
2
)( dxxf
2
2
22
2
2
11
2
2
00 )(,)(,)( dxxlcdxxlcdxxlc
-
7/30/2019 Numerical Integration Lecture
56/59
Solution
56
)1(3
8)0(
3
4)1(
3
8)(forFormulaQuadratureGaussThe
3
8
2
)1(,
3
4)1)(1(,
3
8
2
)1(
2
)1(
)12)(02(
)1)(0()(
)1)(1()21)(01(
)2)(0()(
2)1(
)20)(10()2)(1()(
2
2-
2
2
2
2
2
1
2
2
0
2
1
0
fffdxxf
dxxx
cdxxxcdxxx
c
xx
xxxx
xxxxxl
xxxxxx
xxxxxl
xxxxxx
xxxxxl
-
7/30/2019 Numerical Integration Lecture
57/59
Using the Gauss Quadrature Formula
57
answerexactsametheiswhich,3
16)1(
3
8)0(
3
4)1(
3
8
)1(3
8)0(
3
4)1(
3
8FormulaQuadratureGaussThe
3
16)(foreexact valuThe
)(Let:1Case
222
2
2-
2
2
2-
2
fff
dxxdxxf
xxf
-
7/30/2019 Numerical Integration Lecture
58/59
Using the Gauss Quadrature Formula
58
answerexactsametheiswhich,0)1(3
8)0(
3
4)1(
3
8
)1(3
8)0(
3
4)1(
3
8FormulaQuadratureGaussThe
0)(foreexact valuThe
)(Let:2Case
333
2
2-
3
2
2-
3
fff
dxxdxxf
xxf
-
7/30/2019 Numerical Integration Lecture
59/59
59
Improper Integrals
1
022
12
1
1 2a
1
111:
function.newtheonmethodapply theand
)0assuming(,11)(
followingthelikemationa transforUse
)orislimitstheof(oneintegralsimpropereapproximat
todirectlyusedbecannotearlierdiscussedMethods
dt
t
tdx
xExample
abdtt
ft
dxxf a
b
b