Scribd
Upload
124 views

Numerical Methods With Excel/VBA

The document discusses numerical methods for solving problems that cannot be solved analytically. It introduces the trapezoid rule for numerical integration and provides an example VBA module to implement this method. It also discusses other numerical integration techniques like Simpson's rule and compares different methods on sample integrals. The document notes limitations of numerical methods related to computer accuracy and approximations used. It also mentions other numerical techniques like solving equations can be implemented in Excel using tools like Goal Seek and Solver.

Uploaded by

Hisham Maddah
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
124 views

Numerical Methods With Excel/VBA

The document discusses numerical methods for solving problems that cannot be solved analytically. It introduces the trapezoid rule for numerical integration and provides an example VBA module to implement this method. It also discusses other numerical integration techniques like Simpson's rule and compares different methods on sample integrals. The document notes limitations of numerical methods related to computer accuracy and approximations used. It also mentions other numerical techniques like solving equations can be implemented in Excel using tools like Goal Seek and Solver.

Uploaded by

Hisham Maddah
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Numerical Methods with Excel/VBA:

Many problems in Mathematics, Physics, Economics, etc can


only be solved in very idealized situations in an exact analytical
fashion. Even solvable problems can often only be tackled with
great effort.
Numerical methods often lead to solutions which are extremely
close to the correct answers. They lead very often quickly to some
insights.
Especially with the advance in computer technology, in terms of
speed and storage capacity the limits of what can be computed
are permanently pushed.
Here we only have a glimpse at some methods to get an idea what
could be possible and to apply the programming structures we
have learned so far.
23

! Numerical Integration

Recall:

- Idea: approximate the integral by sums over trapezoidal areas :

24

Programming Excel/VBA Part II (A.Fring)


- Take the subdivision of the domain [a,b] to be evenly spaced:

Trapezoid rule for integration:

Let us write a module (program) for this:


- Input: a lower integration limit
b upper integration limit
n number of subdivisions
some function f(x) which we want to integrate

- Output: approximate value for the integral

25

Sub Nint()
a=0
b=5
n = 100
h = (b - a) / n
I = h * (f(a) + f(b)) / 2
For m = 2 To n
I = I + f(a + h * (m - 1)) * h
Next
Range("B3").Value = I
End Sub
Put the result onto the Excel sheet into the cell B3
Function f(x)
f=x^4
End Function
26

Programming Excel/VBA Part II (A.Fring)


Example 1:

- The program gives:

Example 2:

- Generate the by 4 Arctan(1). In VBA this is written as 4 *Atn(1).


- The program yields:

27

- So far we could have solved the intgrals also analytically, but not the
next integral.
Example 3:

- How do we deal with the lower bound a=0? This is well defined
analytically, but the computer can not handle 0/0, if we dont
specify how to do that. Recipe: Just take the lower bound a to be a
very small number, e.g.a=0.0000001.
- The program yields:

28

Programming Excel/VBA Part II (A.Fring)


Example 4:

- How do we deal with infinity? Introduce a cut-off at some value


large enough such that the mistake is small. This is possible because
the integrant falls off sharply after certain values:

- Compute instead:
- The program gives:

29

Different types of methods:


- Simpsons 1/3 rule (based on three adjacent points):

- Simpsons 3/8 rule (based on four adjacent points):

- Runge-Kutta methods, Monte Carlo integration,


Here we do not derive these rules, but just take them as facts.
See a different course on numerical methods for details.
- Let us implement the Simpsons 3/8 rule as a user defined function.
- Implement the Simpsons 1/3 rule in Labsession 3.
30

Programming Excel/VBA Part II (A.Fring)


Function Nintff(a, b, n)
h = (b - a) / n
I=0
For m = 1 To n - 2 Step 3
I = I + (f(a+h*(m-1)) +3* f(a+h* m) +3* f(a+h*(m+1))+f(a+h*(m+2)))
Next
Nintff = I * h * 3 / 8
End Function
Example 1:

A problem here is to find a good cut-off for the upper limit.

b=10 ?

31

- Compare different integration methods:

i) open a worksheet

In this example we introduce an additional error though the


cut-off.
When the subdivision of the interval is large enough the three
methods are almost equally good.
32

Programming Excel/VBA Part II (A.Fring)


Limitations:
- The speed of the computer.
- The accuracy of the numerical method used. (In principle there
exist procedures to estimate the errors made.)
- The accuracy of the computer, i.e. in computing functions used
in the program and variables employed such as single or double
precision.
- Various other approximations such as division by zero, cut-offs
for lower and upper bounds etc.
There exist different types of numerical methods for other
mathematical problems, such as solving polynomial equations,
solving differential equations etc.
Some methods are implemented in Excel as Built-in functions:

33

! Goal Seek
Goal seek is a numerical routine implemented in Excel in form
of a built-in function. It canbe used to solve equations.
Usage: select Tools Goal Seek a dialog window opens
- Set cell contains the left hand side of
an equation you want to solve
- To value contains the RHS of the equation
- By changing cell contains the variable
of the equation
Disadvantage: You have to guess a value near the answer.

Example: Solve the equation: 2x2-9x-5=0


(We expect to find: x1=-1/2 and x2=5 )
- Type into the cell C3: =2*B3^2-9*B3-5
- Type into the cell C4: =2*B4^2-9*B4-5 34

Programming Excel/VBA Part II (A.Fring)


- Type into the cell B3 some starting value, e.g. -10
- open the Goal Seek dialog box and fill in

- OK

- The cell B3 and C3 have changed to -0.5 and 1.912E-07


- Repeat this process for the cells C4 and B3 to find the other
solution. (You need a new guess for the starting value.)
A more sophisticated method is the Excel Solver.
35

Programming Excel/VBA Part II (A.Fring)

You might also like