Scribd
Upload
58 views3 pages

Numerical Integration Techniques Overview

This document discusses several numerical integration and differential equation solving methods: 1) Trapezoid rule and Simpson's 1/3 rule for numerical integration with errors of O(h^2) and O(h^4) respectively. 2) Gaussian quadrature which computes integrals as a weighted sum with an error of O(e^(-cn)) for some c. 3) Finite difference and finite element methods for discretizing differential equations including the Schrodinger equation. 4) Runge-Kutta and Gill's method for numerically solving ordinary differential equations with local truncation errors of O(h^5).

Uploaded by

Ming-Chien Hsu
Copyright
© Attribution Non-Commercial (BY-NC)
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
58 views3 pages

Numerical Integration Techniques Overview

This document discusses several numerical integration and differential equation solving methods: 1) Trapezoid rule and Simpson's 1/3 rule for numerical integration with errors of O(h^2) and O(h^4) respectively. 2) Gaussian quadrature which computes integrals as a weighted sum with an error of O(e^(-cn)) for some c. 3) Finite difference and finite element methods for discretizing differential equations including the Schrodinger equation. 4) Runge-Kutta and Gill's method for numerically solving ordinary differential equations with local truncation errors of O(h^5).

Uploaded by

Ming-Chien Hsu
Copyright
© Attribution Non-Commercial (BY-NC)
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

0.

1 Integration
0.1.1 Trapesoid rule
I
i
= [f(x
i
) + f(x
i+1
)]
h
2
I =
n1

i=1
I
i
= [f(x
1
) + 2f(x
2
) + 2f(x
3
) + + 2f(x
n1
) + f(x
n
)]
h
2
Truncation error E = O(h
2
)
0.1.2 Simpsons 1/3 rule
f
n

i
f(x
i
)l
i
(x) P
n1
(x)
I =
_
b
a
f(x)dx

n
i=1
_
f(x
i
)
_
b
a
l
i
(x)dx
_
=

n
i=1
A
i
f(x
i
)
A
i
=
_
b
a
l
i
(x)dx, i = 1, 2, , n
l
1
(x) =
(xx
2
)(xx
3
)
(x
1
x
2
)(x
1
x
3
)
l
2
(x) =
(xx
1
)(xx
3
)
(x
2
x
1
)(x
2
x
3
)
l
3
(x) =
(xx
1
)(xx
2
)
(x
3
x
1
)(x
3
x
2
)
For each interval b a = 2h, we can get
A
1
=
h
3
, A
2
=
4h
3
, A
3
=
h
3
Therefore, for each interval
I
i
=
3

i=1
A
i
f(x
i
) =
_
f(a) + 4f(
a + b
2
) + f(b)
_
h
3
The total integration
I =
_
b
a
f(x)dx [f(x
1
) + 4f(x
2
) + 2f(x
3
) + 4f(x
4
) + + 2f(x
n2
) + 4f(x
n1
+ f(x
n
)]
h
3
A total error of O(h
4
)
0.1.3 Gaussian Quadrature
I =
_
1
1
F(t)dt =
n

i=1
C
i
F(t
i
)
1
Table 1: Gaussian Quadrature Parameters
n t
i
c
i
order
2

3 1 3

3 1
3

0.6 5/9 5
0 8/9

0.6 5/9
0.2 Finite Element
0.2.1 Schr odinger Equation
Let | = a
n
| and H
mn

m
| H |
n
, and then from H = E we
can get

n
a
n

m
| H |
n

. .
H
mn
= Ea
n

nm
= Ea
m
The matrix form is as follows
_

_
H
11
H
12
H
13

H
21
H
22
H
23

.
.
.
.
.
.
.
.
.
.
.
.
_

_
= E
_

_
a
1
a
2
.
.
.
_

_
= E, where
_

_
a
1
a
2
.
.
.
_

_
Fourier basis
basis :
n
=
_
2
L
sin(
nx
L
) =
_
2
L
sin(k
n
x) dened within the range 0 L,
where k
n

n
L
, satisfying
m
|
n
=
mn
For T

2
2m

2
, we can get
m
| T |
n
=

2
n
2
2m

mn
0.3 Finite Dierence

x
2

i
=
1
a
2
[
i+1
2
i
+
i1
]
0.3.1 Schr odinger Equation
t
0


2
2ma
2
and U
n
U(x
n
)
i

n
dt
= [H]
x=x
n
= (U
n
+ 2t
0
)
n
t
0

n1
t
0

n+1
2
For sin(k

x), it is straightforward to show that

2
2m
_

x
2
_
x=x
n
= t
0
(k

a)
2
(x
n
)
while
t
0
[(x
n+1
) 2(x
n
) +
x
n1
] = 2t
0
(1 cos k

a)(x
n
)
Normalization
Analytical
_
dx|

(x)|
2
= 1
Numerical
N

n=1
|

(x
n
)|
2
= 1
Therefore,
|

(x
n
)|
2
. .
Numerical
= |

(x)|
2
a
2
. .
Analytical
0.4 ODE solving
0.4.1 Runge-Kutta method
y
n+1
= y
n
+
1
6
[k
1
+ 2k
2
+ 2k
3
+ k
4
] + O(h
5
)
where
k
1
= hf(x
n
, y
n
)
k
2
= hf(x
n
+
1
2
h, y
n
+
1
2
k
1
)
k
3
= hf(x
n
+
1
2
h, y
n
+
1
2
k
2
)
k
4
= hf(x
n
+ h, y
n
+ k
3
)
0.4.2 Gills method
y
n+1
= y
n
+
1
6
[k
1
+ (2

2)k
2
+ (2 +

2)k
3
+ k
4
] + O(h
5
)
where
k
1
= hf(x
n
, y
n
)
k
2
= hf(x
n
+
1
2
h, y
n
+
1
2
k
1
)
k
3
= hf(x
n
+
1
2
h, y
n
+
1
2
(1 +

2)k
1
+ (1
1
2

2)k
2
)
k
4
= hf(x
n
+ h, y
n

1
2

2k
2
+ (1 +
1
2

2)k
3
)
3

You might also like