Computational Methods for

Engineers

Lecture 2

Reuel C. Pallugna
instructor
Topics
• Roots of equations
• Open Methods
• Fixed point iteration Method
• Newton Raphson Method
• Secant Method
• Polynomials using Matlab
• Solutions to Linear Algebraic Equations
• Curve Fitting
• Numerical Differentiation and Integration
• Numerical Solutions to Ordinary Differential Equations
• Numerical Solution to Partial Differential Equations
 Fixed Point Iteration Method
• This is called a method of Example:
successive substitution Solve for the roots of
• A function f(x) is rearranged
such that the x is on the left
side of the equation Solution:

• The auxiliary equation is then

cast into its iterative form
A given function

Rearrange into an auxiliary equation

Cast into its iterative form

Iterated until its tolerance value

Newton-Raphson Method
• One of the most widely used root location
method
• Needs only one “good” initial guess
• Needs exact derivative of the function
• Fast method, easily converge
• May not converge quickly for multiple roots and
extremely nonlinear or polynomial curves
 Basic Concept of Newton-Raphson Method

The equation of the tangent line is

The Basic Concept of NR:

1. Assume an initial estimate of the root x1
2. Take the derivative of the curve at that initial point This is now the
3. The derivative is the slope of a line tangent to the curve NR formula
4. Locate the point where the tangent line crosses the x-axis , x2
5. Use x2 as the second estimate of the root
6. Repeat the process until desired accuracy is reached
Newton-Raphson Method: example

Solution:
 Newton-Raphson Method Algorithm and Matlab Code

NR Algorithm NR Code in Matlab Matlab Graph

• % This is a Newton Raphson Code to solve

% the equation
% f(x) =e^x -x, evaluated between 0<x<1
% df(x) = -e^x-1
xr = 0;
tol =1e-6;
dx = inf;
while abs(dx) > tol
dx = (exp(1).^-xr-xr)/((-exp(1).^-xr) -1);
xr = xr-dx
end

Matlab Output Result from Textbook

NewtonRaphsonCode

xr = 0.5000

xr = 0.5663

xr = 0.5671

xr = 0.5671
 Secant Method
Basic Concept of the Secant Method
• Is an open method ( no need to bracket the root)
• Similar to the Newton Raphson method but
requires two initial guesses to determine the slope
of the secant line
• Whereas the NR uses the derivative to determine
the slope of the tangent line, the secant method
uses the two initial guesses to determine the
slope of the secant line
• The point in the secant line that crosses the x-axis
is used to estimate of the root of the curve Substituting the value of
• A numerical equivalent to the NR

Derivation of the Secant Method Formula Equation of a line connecting

Points 2 and 3 is: The general equation for the
Initial guesses Slope of line that passes Secant Method
Point 1 and point 2 is:
Points y-coordinates
on curve at initial guesses
but:
Secant Method (alternate derivation)
The Basic Concept of the Secant Method Initial guesses
can be derived from the Newton-Raphson
method
Points y-coordinates
on curve at initial guesses

The general equation for the

Secant Method
Secant Method sample problem

Solution:
Initial guesses

Solving for y-coordinates

Solving for the next guess

 Secant Method Algorithm and Matlab Code
Secant Method Algorithm Secant Method Code in Matlab Result from Matlab Code
% Matlab Code sample for the secant method
% this is taken from CEM Computational Methods for Electrical
x2 = 0.6127
% Engineers and written by Dr. Ramf
% this is put into the slides as an illustration of Secant Code
% This is a more complete code
x2 = 0.5638
% INITIALIZE MATLAB
close all; % Closes all figure window
clc; % Clears command window
x2 = 0.5672
clear all; % Clears workspace
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%DASHBOAARD
x2 = 0.5671
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% DEFINE THE FUNCTION
func =@(x) exp(1)^-x - x; % a handle to the function
x2 = 0.5671 ( in red dot )
% INITIAL GUESS
x1 = 0;
x2 = 1;
%TOLERANCE
tol = 1e-6;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%IMPLEMENT SECANT METHOD
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% EVALUAT FUNCTION AT X1
f1 = func(x1);
%MAIN LOOP
dx =inf;
iter = 0;
while abs(dx)>tol
% count iterations
iter = iter + 1;
f2 = func(x2);
Note: dx = (x2-x1)*f2/(f2-f1);
% the second point now becomes new first point
Videos of this topic can be viewed x1 = x2;
f1 = f2;
On youtube at CEM Lectures, Topic 4b, %update position of the second point
x2 = x2-dx
Numerical Methods for Electrical Engineers end
By Dr. Ramf
Roots of Polynomials
A polynomial of the form
Introduction
• Can be expressed in matrix form. The coefficient
Matrix , c , of which is as shown highlighted

The eigenvalue algorithm can then be used to

solve
for the roots of the equation
The Matlab function to solve the given equation is

>> x = roots(c) Where c is the coefficient of f(x)

Roots of Polynomials solution using Matlab roots
function
Solve for the roots of the equation using Matlab roots function

Solution:
Matlab m-file using roots function Matlab Output
% this m-file is an illustration of using
the Matlab roots function
% to solve for the roots of the equation x=
% f(x) = x^5 - 3.5x^4 + 2.75x^3 + 2.125x^2
-3.875x +1.25 2.0000 + 0.0000i
c = [1 -3.5 2.75 2.125 -3.875 1.25]; -1.0000 + 0.0000i
x = roots(c) 1.0000 + 0.5000i
1.0000 - 0.5000i
0.5000 + 0.0000i
Activity 2
• Solve problems 6.1 to 6.6 of Applied Numerical Methods with Matlab,
2nd Ed., by S.C. Chapra