Scribd
Upload
44 views

Fundamentals of Programming For Chemical Engineers Methods in Solving Engg Problems

The document discusses different numerical methods for finding the roots of functions including bisection, regula falsi, secant, and Newton's method. It defines concepts like absolute error, relative error, and convergence rate. Examples are provided for each method.

Uploaded by

Mark Anthony Caro
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
44 views

Fundamentals of Programming For Chemical Engineers Methods in Solving Engg Problems

The document discusses different numerical methods for finding the roots of functions including bisection, regula falsi, secant, and Newton's method. It defines concepts like absolute error, relative error, and convergence rate. Examples are provided for each method.

Uploaded by

Mark Anthony Caro
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

4/14/2016

Fundamentals of Programming
for Chemical EngineersIntroductory Concepts
Leandro T. De Castro

Methods in Solving
Engineering Problems

4/14/2016

Definition: Absolute Error and Relative Error


Let p* be an approximation to p

: =

: =

Round-off Errors
To minimize loss of accuracy due to round-off
errors:
a. Reformulate the problem
b. Reduce the number of error-producing
computation
c. Use High-order digit arithmetic

4/14/2016

Convergence Rate
Suppose that a sequence {Xn} converges to a real
number X*. We say the convergence is of order l (where
k>=1) provided that for some positive number C, the
following inequality will eventually true:

In general, higher values of k give faster convergence.
If a sequence converges at k=1, it is lineartly convergent
If a sequence converges at k=2, it converges
quadratically.

Finding the root of f(x)=0


Methods:
Bisection Method
The Regula Falsi Method
The Secant Method
Newtons Method

4/14/2016

Bisection Method
- Based on the intermediate value theorem

Assumptions:
1. f(x) is continuous
2. On [a,b], f(a), f(b) have opposite signs
3. Tolerance = tol >0 is prescribed

Bisection Method

f(b)

root r
f(a)

4/14/2016

Bisection Method
Method will iteratively construct a sequence xn that
converges to the root r and will stop when it can be
guaranteed that |r-xn| < tol.
The difficulty though with the error expression, i.e.,
=
is that the roor of a function is rarely
known beforehand.
As the limiting criterion to the number of iterations, the
following may be used:
1.
<
2.
<
3.

<

4.

<

Algorithm for Bisection Method:


1.
2.
3.
4.
5.
6.

Choose an interval [a,b]


If f(a) and f(b) have opposite signs: start
Iterate: = +
/2
If
<
: stop
Set a=x1 if f(x1) has same sign as f(a), else set b=x1
Goto 3

4/14/2016

Example: Consider the function


=

a. Show that f(x) has a root on the interval [0,1]


b. Use the bisection method to approximate this root with an absolute
error <5e-3
c. How many iterations would it take the bisection method to guarantee
an absolute error < 1e-6

Regula Falsi

4/14/2016

Algorithm for Regula Falsi Method:


1. Choose an interval [xa, xb]
2. If f(xa) and f(xb) have opposite signs: start
3. Iterate:

4. If error < tol; : stop


5. Set
= if ( ) has same sign as (
6. Goto 3

) , else set

Example: Find the root of

4/14/2016

Secant Method

4/14/2016

Algorithm for Secant Method:


1. Choose an interval [xn-1, xb]
2. If f(xn-1) and f(xn) have opposite signs: start
3. Iterate:

4. If error
< tol; : stop
5. Goto 3 using the two most recently generated points

Example: Find the root of

4/14/2016

Newtons Method

10

4/14/2016

f(x) is a differentiable function that has a root


x= r which we wish to accurately approximate.

11

You might also like