Raymond Phan, Ph.D.

Candidate
Dept. of Electrical & Computer Engineering – Ryerson University
Multimedia and Distributed Computing Research Laboratory
rphan@ee.ryerson.ca

Monday, February 6th, 2012,

Eric Palin Hall (EPH) – Room 201
4 p.m. – 6 p.m.
Tutorial Topics – (1)
 1st Part of the Tutorial
 Introduction
 What MATLAB is (briefly) and where can I get it?
 Review of array and matrix operations
 Basic Math Operations
 Matrix Algebra vs. Point-by-Point Operations
 Modifying and extracting a portion of a matrix
 Creating and working with dynamic arrays
 Creating Function Handlers
 Creating equations of single or multiple variables
 Not that inline crap that you learned!
Tutorial Topics – (2)
 Plotting Graphs
 Creating an array of numbers following a pattern
 Plotting 2D graphs properly using function handlers
 Plotting multiple separate graphs on the same window
 Using MATLAB syntax to customize plots
 Plotting 3D plots: Surface and Mesh plots
 Some useful MATLAB commands and constructs
 Creating cell arrays
 Review of logical operators (AND, OR, NOT)
 The find command
 Locating values in an array or matrix meeting certain criteria

 Modifying matrices using logical operators

Tutorial Topics – (3)
 Creating Transfer Functions
 Both polynomial form and zero-pole-gain (ZPK) format

 Minimum Realizations

 Calculating the impulse and step response

 Finding the roots of equations
 Polynomial equations

 Non-linear equations

 Evaluating a polynomial equation at given points

 Using convolution to multiply polynomial equations
 Partial Fraction Decomposition
 Create a matrix by using copies of a smaller matrix
 Converting an array and reshaping it into a matrix
 Tutorial Topics – (4)
 2nd Part of the Tutorial:
 Advanced Mathematical Analysis
 Calculating Derivatives and Integrals Analytically using the
Symbolic Math Toolbox
 Using MATLAB to calculate with respect to “x”

 Evaluating derivatives at points

 Calculating the definite integral

 Finding the line of best fit via regression

 Linear regression, parabolic regression, etc.

 More commonly known as polynomial interpolation

 Calculating the area under a curve when the function is unknown,

but data points are known
 Calculating via trapezoidal rule
Tutorial Topics – (5)
 Numerical Solutions to Differential Equations
 Output is a set of points that is the solution at different points
in time  Not the actual equation
 Sorting arrays
Tutorial Topics – (1)
 1st Hour: 4:10 p.m. – 5:00 p.m.
 Introduction
 What MATLAB is (briefly) and where can I get it?
 Review of array and matrix operations
 Basic Math Operations
 Matrix Algebra vs. Point-by-Point Operations
 Modifying and extracting a portion of a matrix
 Creating and working with dynamic arrays
 Creating Function Handlers
 Creating equations of single or multiple variables
 Not that inline crap that you learned!
Introduction – (1)
 MATLAB  Stands for MATrix LABoratory
 Created by Cleve Moler @ Stanford U. in 1970
 Why? Makes linear algebra, numerical analysis and
optimization a lot easier
 MATLAB is a dynamically typed language
 Means that you do not have to declare any variables
 All you need to do is initialize them and they are created
 MATLAB treats all variables as matrices
 Scalar – 1 x 1 matrix. Vector – 1 x N or N x 1 matrix
 Why? Makes calculations a lot faster (will see later)
Introduction – (2)
 How can I get and/or use MATLAB?  3 ways:
 1) Find it on any of the ELCE departmental computers
 Log in to your account, go to Applications  Math 
MATLAB R2010b
 2) Use any Ryerson computer on to the ACS network
 Log in with your Matrix ID and Password, then go to Start 
MATLAB R2011
 3) Install it on your own laptop / computer
 Go to http://www.ee.ryerson.ca/matlab for more details
 You must be on the Ryerson network to sign up for an account
 After, you can download MATLAB from anywhere
Introduction – (3)
 A few things before we start
 This is an advanced tutorial
 I assume you have basic knowledge in MATLAB
 I will cover some concepts seen in 2nd year Math courses
 Don’t worry, a lot of you may have forgotten, or have not
seen this stuff before, and I will give brief intros
 If you have no exposure to MATLAB, then as long as you
have some programming experience, you will be able to
understand the basics part
 For each MATLAB line I show, assume I push ENTER
immediately after
 Enter the commands in the MATLAB Command Prompt
 The main window that opens when you open up MATLAB
 Push ENTER after each command you type in to execute
Tutorial Topics – (1)
 1st Hour: 4:10 p.m. – 5:00 p.m.
 Introduction
 What MATLAB is (briefly) and where can I get it?
 Review of array and matrix operations
 Basic Math Operations
 Matrix Algebra vs. Point-by-Point Operations
 Modifying and extracting a portion of a matrix
 Creating and working with dynamic arrays
 Creating Function Handlers
 Creating equations of single or multiple variables
 Not that inline crap that you learned!
Review: Array & Matrix Ops (1)
 Creating arrays:
 No need to allocate memory  Easy to create with the
numbers you want  Use square braces in between []
array = [1 5 8 7];  Row Vector
array = [1;5;8;7];  Column Vector
 Use a space to separate each element
 Use a semi-colon (;) to go down to the next row
array = [1 5 8 7]’;  Also a column vector
 ‘ means take the transpose (interchange rows & columns)
 After, MATLAB creates an array automatically, with the
values you specified, stored in MATLAB as array
Review: Array & Matrix Ops (2)
 Accessing elements in an array
 Use the name of the array, with round brackets ()
 Inside the brackets, specify the position of where you
want to access
 Remember, MATLAB starts at index 1, not 0 like in C
 Examples  array = [1 5 8 7];
 num = array(2);  num = 5
 num = array(4);  num = 7
 Accessing a range of values  Use colon (:) operator
 Style: val = array(begin:end);
 Begin & end are the starting & ending indices to access
Review: Array & Matrix Ops (3)
 Examples  array = [1 5 8 7];
 val = array(2:4);
 val  3 element array, containing elements 2 to 4 of array
 This is equivalent to val = [5 8 7];
 val = array(1:2);
 val  2 element array, containing elements 1 and 2 of array
 This is equivalent to val = [1 5];
 We can also access specific multiple elements
 val = array([1 4 2 3]);
 val  A 4 element array: 1st element is from index 1 of array,
2nd element is from index 4 of array, etc.
 This is equivalent to val = [1 7 5 8];
Review: Array & Matrix Ops (4)
 val = array([1 3 2]);
 val  A 3 element array: 1st element is from index 1 of
array, 2nd element is from index 3 of array, etc.
 This is equivalent to val = [1 8 5];
 Last thing about arrays
 To copy an entire array over, simply set another variable
equal to the array you want to copy
 i.e. array2 = array;
 This is equivalent to array2 = [1 5 8 7];
Review: Array & Matrix Ops (5)
 Let’s move onto matrices
 To create a matrix, very much like creating arrays
 Use square braces [], then use the numbers you want!
 Use spaces to separate between the columns
 Use semicolons (;) to go to each row
 Example:
M = [1 2 3 4;
M= 5 6 7 8;
9 10 11 12;
13 14 15 16];
Review: Array & Matrix Ops (6)
 How to access an element in a matrix
 Use the name of the matrix, with round braces ()
 We use two numbers, where each number references a
dimension in the matrix  Separate them by a comma
 1st number is the row, 2nd number is the column
 Examples:
B = M(2,3);  B = 7; 2nd row, 3rd col.
B = M(3,4);  B = 12; 3rd row, 4th col. M =
B = M(4,2);  B = 14; 4th row, 2nd col.
B = M(1,3);  B = 3; 1st row, 3rd col.
Review: Array & Matrix Ops (7)
 To access a range of values, use the colon operator just
like we did for arrays, but now use them for each
dimension
 Examples: M=
 B = M(1:3,3:4);  Rows 1-3, Cols 3-4
 This is equivalent to B = [3 4; 7 8; 11 12];
 B = M(3:4,3:4);  Rows 3-4, Cols 3-4
 This is equivalent to B = [11 12; 15 16];
 B = M(1,1:3);  Row 1, Cols 1-3
 This is equivalent to B = [1 2 3];
 B = M(3:4,2);  Row 3-4, Col 2
 This is equivalent to B = [10;14];
Review: Array & Matrix Ops (8)
 Last thing for accessing  Using : by itself
 Using : for any dimension means to access all values
 Examples:
 A = M(1:2,:);  Rows 1-2, All Cols. M =
 Equal to A = [1 2 3 4; 5 6 7 8];
 A = M(2,:);  Row 2, All Cols.
 This is equivalent to A = [5 6 7 8];
 A = M(:,3);  All Rows, Col 3
 This is equivalent to A = [3;7;11;15];
 A = M(:,1:3)  All Rows, Cols. 1 – 3
 Equal to A = [1 2 3; 5 6 7; 9 10 11; 13 14 15];
 A = M(:,:); or A = M;  Grab the entire matrix
Review: Array & Matrix Ops (9)
 Let’s get onto some basic math operations!
 Let’s assume the following:
 A and B are vectors / arrays, or matrices of compatible
dimensions
 We assume that A and B can be properly added, subtracted,
multiplied and divided together
 Also, assume that n is a scalar number
 Here’s a table that provides a good summary of all of
the basic operations you can perform on vectors /
array, and matrices
Review: Array & Matrix Ops (10)

 Let’s look at addition and subtraction first

 With addition or subtraction, A & B must be same size
 The result will be a matrix of the same size, where each
element is added or subtracted with its corresponding
location
 For vectors, we do this for each corresponding element
Review: Array & Matrix Ops (11)
 Example:
 A = [1 2; 3 4];
 B = [5 6; 7 8];
 E = A + B;
 This is equivalent to E = [6 8; 10 12];
 F = A – B;
 This is equivalent to F = [-4 -4; -4 -4];
 G = C + D;
 This is equivalent to G = [5 7 9];
 H = C – D;
 This is equivalent to H = [-3 -3 -3];
Review: Array & Matrix Ops (12)

 * here means matrix multiplication

 Division is a bit more complicated
 We have left division and right division
 Left Division: A-1B, and Right Division: AB-1
 ’ means transpose, as we saw earlier
 A^n means to multiply a matrix n times
Review: Array & Matrix Ops (13)
 Examples:
 C = A * B;
 This is equivalent to C = [19 22; 43 50];
 D = A \ B;
 This is equivalent to D = [-3 -4; 4 5];
 E = A / B;
 This is equivalent to E = [-3 -2; 2 -1];
 F = A^2;
 This is equivalent to F = [7 10; 15 22];
 G = B^3;
 This is equivalent to G = [881 1026; 1197 1394];
Review: Array & Matrix Ops (14)
 Special case: +,-,* or / every element in a vector or
matrix by a constant number
 You use an operation, with the constant number
 Examples: A = [1 2; 3 4]; B = [5 6 7 8];
 C = 2*A;
 This is equivalent to C = [2 4; 6 8];
 D = B/3;
 This is equivalent to D = [1.6666 2 2.3333 2.6666];
 E = A-2;
 This is equivalent to E = [-1 0; 1 2];
 F = B+4;
 This is equivalent to F = [9 10 11 12];
Review: Array & Matrix Ops (15)
 So far, we’ve done operations with matrix algebra
 MATLAB also has called point-by-point operations
 Point-by-point operations are applied to
multiplication (*), division (\,/), & exponentiation (^)
 To use point-by-point operators, you put a dot (.)
before any of the operations above: .* or ./ or .\ or .^
 Point-by-point operators work by taking the operation
and performing it on corresponding elements only
 The 1st element of A, is *, /, \ or * with 1st element of
 ...don’t get it? Here’s an example
Review: Array & Matrix Ops (16)
 Examples:
 C = A * B;
 This is equivalent to C = [19 22; 43 50];
 D = A .* B;
 This is equivalent to D = [5 12; 21 32];
 E = A \ B;
 This is equivalent to E = [-3 -4; 4 5];
 F = A .\ B;
 This is equivalent to F = [5 3; 2.3333 2];
 Same as taking every element from B and dividing by its
corresponding element in A
Review: Array & Matrix Ops (17)
 Examples:
 G = A / B;
 This is equivalent to G = [-3 -2; 2 -1];
 H = A ./ B;
 This is equivalent to H = [0.2 0.3333; 0.4286 0.5];
 Same as taking every element from A and dividing by its
corresponding element in B
 I = A^2;
 This is equivalent to I = [7 10; 15 22];
 J = A.^2;
 This is equivalent to J = [1 4; 9 16];
Review: Array & Matrix Ops (18)
 Now on to modifying elements in arrays and matrices
 Do same thing with accessing elements or in groups
 Flip the order of what’s on which side of the equals sign
 Examples, using the A matrix previously:
 A(3,4) = 7;  Store 7 in 3rd row, 4th col
 A(1:2,3:4) = [5 6; 7 8];
 Store this 2 x 2 matrix in Rows 1-2, and Cols 3-4
 A(2,:) = [8 19 2 20];
 Store this row vector in Row 2

 A(:,3) = [5;4;3;2];
 Store this column vector in Col 3

 A(1:2,:) = 5;  Set Row 1-2 and all Cols. to 5

Review: Array & Matrix Ops (19)
 Last thing for our review: Dynamic Arrays
 Good for not knowing size before hand
 We can create arrays that grow in size
 We can extend the size and add new elements
 You simply set a variable to a set of square braces []
 array = [];
 After, you just assign numbers to the array
 array(1:4) = 5;
 array(5:7) = [8 6 8];
 array(end+1) = 7;  end goes to the end of array,
and +1 puts 7 after the end and extends the size by 1.
 Equivalent to array = [5 5 5 5 8 6 8 7];
Creating Function Handlers (1)
 We can use MATLAB to create complex equations
 These equations are in terms of whatever variable you
want: x, y, z, w, t, etc.
 We can create single variable, or multiple variables
 These are called function handlers
 Once we create function handlers, we simply supply
the function handlers with the values we want to
substitute into the variables
 The function handlers will evaluate the function at these
values, and give us outputs
Creating Function Handlers (2)
 To create function handlers, you use the @ sign, then
surround it with round braces ()
 In the braces, place the variables relevant to equation
 After, you then write out the equation in C style
 Make sure that all multiplication, division and
exponentiation operations are point-by-point
 Why? That way, we can supply a vector or matrix into
the function handler, and it’ll evaluate the function for
every element on its own, and send that to the output
 Built-in MATLAB functions like sin, cos, sqrt, etc. are
all function handlers too!
Creating Function Handlers (3)
 We don’t need to do point-by-point if we’re performing
these operations on constants
 Function Handler Examples:

 f = @(x) 5*sin(x) – x.^2 + 3;

 f = @(x) sqrt(x.^3 + exp(-2*x) – 5);

 g = @(x,y)(25 – x.^2 – y.^2);
  i is the imaginary variable
 h = @(w,t) cos(w.*t) + i*sin(w.*t);
Creating Function Handlers (4)
 We can also nest function handlers within other ones
 Examples:

 f = @(x) 2*x.^2 + 3;
 g = @(x,f) 2*x.*f(x-3) + 5*x;
 Make sure you specify f as a variable to function handler g
 To use function handlers, use them like any other
function in MATLAB, using how you named them
 You put in values that you want the function to be
evaluated at  Vectors / Arrays or Matrices
 Output: Same size as input  Function evaluated on
each element
Creating Function Handlers (5)
 Examples:
 A = [1 2; 3 4]; B = [2.5 6.7 9];

 f = @(x) 5*sin(x) – x.^2 + 3;

 g = @(x,y) 25 – x.^2 – y.^2;
 a = f(3);
 Equivalent to
 In MATLAB: a = -5.2944  sin is in radians

 b = f(A);
 Using same principle, we get:
b = [6.2074 3.5465; -5.2944 -16.7840];
 c = g(B,B-2);
 Using same principle, this is equivalent to:
c = [18.5 -41.98 -105];
Tutorial Topics – (2)
 Plotting Graphs
 Creating an array of numbers following a pattern
 Plotting 2D graphs properly using function handlers
 Plotting multiple separate graphs on the same window
 Using MATLAB syntax to customize plots
 Plotting 3D plots: Surface and Mesh plots

 Some useful MATLAB commands and constructs

 Creating structures and cell arrays
 Review of logical operators (AND, OR, NOT)
 The find command
 Locating values in an array or matrix meeting certain
criteria
 Finding minimum and maximum values
Plotting Graphs – (1)
 This is not properly taught in any courses that I’ve
TAed  Here to remedy this situation
 MATLAB makes it very easy to plot data, and makes
the graphs look visually pleasing
 The basic run through is this:
 (a) Provide an array of points for each set of axes
 (b) Run a function that plots things for you
 (c) Run a few more commands that customize the
appearance of the plot  grid, legend, axes labels
Plotting Graphs – (2)
 Let’s start with 2D plotting:
 Assuming that we have two variables, x and y and are
both the same size
 To produce the most basic graph in MATLAB that plots
y vs. x, you just do:
plot(x,y);
 Let’s start off with a basic example: y = x
 We need to generate x and y values first
 x = [1 2 3 4 5];
 y = [1 2 3 4 5]; or y = x;
 plot(x,y);... and we’ll get...
Plotting Graphs – (3)
Plotting Graphs – (4)
 This is great... but what if we wanted to generate a
large set of numbers?
 A pain to manually input all these points into an array!
 If the numbers follow a pattern then this is easy
 Use the following syntax:
 array = first : increment : last;

 first – Value to start at

 increment – step size
 last – Final value to stop
 Can also do array = first : last;
 Leaving the increment out defaults to a step size of 1
Plotting Graphs – (5)
 Examples:
 array = 0 : 2 : 10;
 Equivalent to array = [0 2 4 6 8 10];
 array = 3 : 3 : 30;
 Equivalent to array = [3 6 9 12 15 18 ... 30];
 array = 1 : 0.1 : 2.2;
 Equivalent to array = [1 1.1 1.2 1.3 ... 2.2];
 array = 3 : -1 : -3;
 Equivalent to array = [3 2 1 0 -1 -2 -3];
 array = 5 : 15;
 Equivalent to array = [5 6 7 8 ... 14 15];
Plotting Graphs – (6)
 Can also perform this using the linspace command
 array = linspace(first, last, N);
 array = linspace(first, last);
 The first style generates an array of size N, with equally
spaced points between first and last
 The second style generates an array of the default size of
100 equally spaced points between first and last
 For non-linear equations, try using logspace
 array = logspace(first, last, N);
 array = logspace(first, last);
 Same as linspace, but on a logarithmic scale instead
Plotting Graphs – (7)
 Let’s move to more advanced equations for plotting

 f = @(x) 5*sin(x) – x.^2 + 3;
 First, let’s generate our x points
 x = -5:0.01:5;
 y = f(x);
 plot(x,y);
 We use a step size of 0.01, because the plot command
draws graphs by “connecting the dots”
 The smaller the step size, the more accurate the plot is
 ... So! We get...
Plotting Graphs – (8)
Plotting Graphs – (9)
 To plot more than one graph on the same window, you
use the plot command by specifying pairs of x and y
arrays in the same command
 i.e. plot(x1,y1,x2,y2,...,xN,yN);
 Each pair of arrays generates a plot in a different colour
 Let’s try plotting 3 graphs on the same plot
 Using the previous example:
x = -5:0.01:5;
y = f(x);
y2 = 1 + f(x);  Shift up graph by 1
y3 = 2 + f(x);  Shift up graph by 2
plot(x,y,x,y2,x,y3);
Plotting Graphs – (10)
Plotting Graphs – (11)
 Now, we can customize the graph to make it look nice
 To label the x-axis, use the xlabel command
xlabel(‘Title’);
 Replace ‘Title’ with whatever you want for the x-axis
 For the y-axis, use the ylabel command
ylabel(‘Title’);
 Replace ‘Title’ with whatever you want for the y-axis
 To put a legend, use the legend command
legend(‘Graph1’,’Graph2’,’Graph3’...’);
 Change ‘Graph1’, ‘Graph2’, ... with the graphs’ names
 Caution: Label the graphs in the same order as how you
placed them in the plot command
Plotting Graphs – (12)
 To name the plot, use the title command
title(‘Title’);
 Change ‘Title’ to whatever you want to name the plot
 To add gridlines, use the grid command
 Using the same graph that we just plotted, we can do
the following  Ensure that graph window is still open
 xlabel(‘X’);
 ylabel(‘Y’);
 title(‘Three Graphs’);
 grid;
 legend(‘y=f(x)’,‘y=1+f(x)’,‘y=2+f(x)’);
Plotting Graphs – (13)
Plotting Graphs – (14)
 By default, MATLAB takes each pair of points in the
arrays and connects them  “connects the dots”
 Also has its own way of colouring each plot
 Is there a way that we can control this behaviour? Yes!
 When using the plot command, you specify arrays of x
and y points, with an additional 3rd parameter: A string
of up to 3 characters surrounded by single quotes
 plot(x,y,’cst’);
 For the 3rd parameter, the 1st character controls the colour
 The 2nd and optionally the 3rd character control the style of the
plot
Plotting Graphs – (15)
 Supported colours and their character codes:
 b – blue, g – green, r – red, c – cyan, m – magenta, y –
yellow, k – black
 Supported plot styles:
Plotting Graphs – (16)
 Examples:
 x = 0 : 10;
 y = x;
 plot(x,y,’g.’);
 Plots a green line with dots at each point
 plot(x,y,’bx’);
 Plots a blue line with an x at each point
 plot(x,y,’ko’);
 Plots a black line with circles at each point
 plot(x,y,’m.-’);
 Plots a magenta line with a solid line and a dot at each point
Plotting Graphs – (17)
 Like how we saw before, you can put multiple plots on
a single graph, each with their own colour and plot
styles with the following:
plot(x1, y1, ’line_style1’, x2, y2,
’line_style2’,..., xN, yN,
’line_styleN’);
 N is the number of plots you want to appear on the
single graph
 xi and yi are the points to the ith graph you want
plotted on the single graph
 line_stylei is the plot style and colour of that ith
graph
 We can also use xlabel, ylabel, legend,
grid to customize as well
Plotting Graphs – (18)
 Let’s say we wanted to plot multiple graphs in separate
plots on the same window
 We use the subplot command
 Command treats the window as having multiple slots
 Each slot takes in a graph
 How do we use the command?
subplot(m,n,p);
 m and n you need to know before hand
 These determine the number of rows (m) and columns (n) for
the amount of graphs you want
 p chooses which location in the window the plot should go to

 The order is from left to right, top to bottom

Plotting Graphs – (19)
 In order to properly use subplot, you must call this
function first
 After, you code the syntax to plot something normally
 Here’s a small example:
 To make a window that has 4 plots: 2 rows and 2 columns
 Do subplot(221) Specify the top left corner
 Code the syntax to plot normally. The plot will appear on the top left
 Do subplot(222) Specify the top right corner
 Code the syntax to plot normally. The plot will appear on the top
right
 Do subplot(223)  Specify the bottom left corner
 Code the syntax to plot normally. The plot will appear on bottom
left
 Do subplot(224)  Specify the bottom right corner
 Code the syntax to plot normally. The plot will appear on bottom
right
Plotting Graphs – (20)
 With this, let’s do another example
 Let’s make 4 graphs: 2 rows and 2 columns
 Each graph will have one plot.
 Let’s make each plot the following:
1. y1 = sin(x);
2. y2 = cos(x);
3. y3 = 3*x;
4. y4 = 6*x;
 Let’s make the range of the plot go from:
x = -10 : 0.1 : 10;… now what?
 Create a blank window, then start adding each graph
using the subplot command
 Customize a graph before moving to the next one
Plotting Graphs – (21)
x = -10:0.01:10;
y1=cos(x); y2=sin(x); y3=3*x; y4=6*x;
figure;  Generate blank window
subplot(2,2,1);  Top left corner
plot(x,y1);
xlabel(‘X’); ylabel(‘Y’); title(‘Plot 1’);
subplot(2,2,2);  Top right corner
plot(x,y2);
xlabel(‘X’); ylabel(‘Y’); title(‘Plot 2’);
Plotting Graphs – (22)
subplot(2,2,3);  Bottom left corner
plot(x,y3);
xlabel(‘X’); ylabel(‘Y’); title(‘Plot 3’);
subplot(2,2,4);  Bottom right corner
plot(x,y4);
xlabel(‘X’); ylabel(‘Y’); title(‘Plot 4’);
Plotting Graphs – (23)
Plotting Graphs – (24)
 One last thing before we move onto more advanced
stuff
 We can customize which values along the x or y axis we
want to display
 Use the set command
set(gca, ‘xtick’, array);
set(gca, ‘ytick’, array);
 gca  Get Current Axis
 array: The values we wish to display for either axis,
stored as an array
 Plotting Graphs – (25)
 Example:
x = 0:2:10;
y = x;
plot(x,y);
set(gca,‘xtick’,0:2:10);
set(gca,’ytick’,0:10);
Plotting Graphs – (26)
 Let’s move onto 3D plots!
 MATLAB can draw 3D plots in a very nice way
 Here are the steps:
 1) Create a function header of two variables
 2) Figure out the range of x and y values you want, then
create a grid of these x and y points
 3) Use the function handler to create the output points
in the 3rd dimension (z)
 4) Use a function that plots this triplet of arrays in 3D
Plotting Graphs – (27)
 How do we generate a grid of x and y values?
 Use the meshgrid function
 This function will output two matrices, X and Y
 X contains what the x co-ordinates are in the grid of
points we want
 Y contains what the y co-ordinates are in the grid of
points we want
 You need to specify the range of x and y values we want to
plot
[X,Y] = meshgrid(x1:x2,y1:y2);
[X,Y] = meshgrid(x1:inc:x2,y1:inc:y2);
Plotting Graphs – (28)
 x1,x2 determines the beginning and ending x values
respectively
 y1,y2 determines the beginning and ending y values
respectively
 We can optionally specify inc which specifies the step
size, just like what we’ve seen earlier
 Example: [X,Y] = meshgrid(0:4,0:5);
 X: Range is from 0 to 4
 Y: Range is from 0 to 5
 Plotting Graphs – (29)
X= Y=

0 1 2 3 4 0 0 0 0 0
0 1 2 3 4 1 1 1 1 1
0 1 2 3 4 2 2 2 2 2
0 1 2 3 4 3 3 3 3 3
0 1 2 3 4 4 4 4 4 4
0 1 2 3 4 5 5 5 5 5

 Top-left corner: X = 0, Y = 0
 Top-right corner: X = 4, Y = 0
 Middle: X = 2, Y = 2
 Each matching location for both the X and Y matrices
corresponds to the (x,y) location in Cartesian co-ordinates
Plotting Graphs – (30)
 Now, how do we create 3D plots? Use either the surf
or mesh commands
 surf(x,y,z);  The z values all have different
colours assigned to them and is a closed plot
 mesh(x,y,z);  The z values all have different
colours assigned to them, but only draws lines through
the points
 For each function, you specify a triplet of x, y and z
values of all the same size
 We can also customize the graph using the commands
seen previously
 Plotting Graphs – (31)
 Example: Let’s plot the following function

f = @(x,y) x.*exp(-x.^2 – y.^2);

[X,Y] = meshgrid(-2:0.2:2,-2:0.2:2);
Z = f(X,Y);  X,Y: Range is from -2 to 2 in steps of 0.2
surf(X,Y,Z);  Previous line gets Z points & now plot
xlabel(‘x’); ylabel(‘y’); zlabel(‘z’);
figure;  Create a new window
mesh(X,Y,Z);  Place mesh graph in new window
xlabel(‘x’); ylabel(‘y’); zlabel(‘z’);
Plotting Graphs – (32)

surf plot mesh plot

Tutorial Topics – (2)
 Plotting Graphs
 Creating an array of numbers following a pattern
 Plotting 2D graphs properly using function handlers
 Plotting multiple separate graphs on the same window
 Using MATLAB syntax to customize plots
 Plotting 3D plots: Surface and Mesh plots
 Some useful MATLAB commands and constructs
 Creating cell arrays
 Review of logical operators (AND, OR, NOT)
 The find command
 Locate values in arrays / matrices meeting certain criteria

 Modifying matrices using logical operators

Useful MATLAB Commands (1)
 We now turn to cell arrays / matrices. What are these?
 Cell arrays are arrays where each element can store any
type of variable
 An example: The 1st element could be a 4 x 4 matrix, the
2nd element could be a structure, 3rd element could be a
single number, etc.
 Cell arrays are powerful, as when we’re reading in data
during multiple experiments, each experiment might
have a different total number of points
 How do we create cell arrays?
 A = cell(r,c);  r and c tell us how many rows
and columns this cell matrix will have
Useful MATLAB Commands (2)
 To create a dynamic cell array, we do:
 A = {};
 After, we populate accordingly
 To write to a cell array or access the actual contents in a
location, we use curly braces {}
 To copy elements of a cell array, we use round braces ()
 Examples:
A = cell(1,3);
A{1}=3; A{2}=ones(3,3); A{3}=zeros(4,4);
 ones(3,3) creates a 3 x 3 matrix full of 1s
 zeros(4,4) creates a 4 x 4 matrix full of 0s
Useful MATLAB Commands (3)
b = A{2};
 Equivalent to b = [1 1 1; 1 1 1; 1 1 1];
C = A(1:2);
 Copies elements 1 and 2 from cell array into C
 i.e. C{1}=3; C{2}=ones(3,3);
d = A{1};
 Equivalent to d = 3;
e = A{3};
 Equivalent to e = [0 0 0 0; 0 0 0 0; 0 0 0 0;
0 0 0 0; 0 0 0 0];
Useful MATLAB Commands (4)
 Review of Logical Operators
 These will be useful if we want to write a function that
will behave differently based on the inputs we give it
 Also useful if we want to modify elements of an array or
matrix that meet certain criteria
 Operators produce
1 if expression is
true
 Produce 0 if
expression is false
Useful MATLAB Commands (5)
 Let’s take a look at the find command
 The find command returns locations of an array or
matrix that satisfy a logical expression
 How do we use? index = find(expr);
 expr: Logical operator used to find certain values
 index: The locations in the array / matrix where
expr is true
 Example: A = [1 3 6 2 4];
index1 = find(A >= 4);
 Gives: index1 = [3 5];  Locations 3 and 5
 Tells us that locations 3 – 5 satisfy this logical expression
 Useful MATLAB Commands (6)
index = find(A >= 2 & A <= 4);
 Gives: index = [2 4 5];
index = find(A == 3);
 Gives: index = 2;
index = find(A ~= 2);
 Gives: index = [1 2 3 5];
 Using these indices, we can modify arrays / matrices when
certain criteria is met
 Examples: A = [1 3 6 2 4]; B = [1 2 3; 4 5 6];
index = find(A >= 2 & A <= 4);
A(ind) = 0;  Finds those values between 2 & 4, and set to 0
Useful MATLAB Commands (7)
 Now, A = [1 0 6 0 0];
index2 = find(~(B >= 1 & B <= 3));
 Find those values that are NOT greater than or equal to
1, and less than or equal to 3.
B(index2) = [9 10 11];
 Now, B = [1 2 3; 9 10 11];
 One more before we get onto better things
index3 = find(B == 2 | B == 5)
B(index3) = 10;
 Now, B = [1 10 3; 4 10 6];
Tutorial Topics – (3)
 Creating Transfer Functions
 Both polynomial form and zero-pole-gain (ZPK) format

 Minimum Realizations

 Calculating the impulse and step response

 Finding the roots of equations
 Polynomial equations

 Non-linear equations

 Evaluating a polynomial equation at given points

 Using convolution to multiply polynomial equations
 Partial Fraction Decomposition
 Create a matrix by using copies of a smaller matrix
 Converting an array and reshaping it into a matrix
Useful MATLAB Commands (8)
 Creating Transfer Functions
 We have seen these in a variety of courses: Signals and
Systems, Differential Equations, Control Systems and
other Mathematics Courses
 Transfer Functions are dealt in the Laplace domain, s.
 We commonly see two forms of Transfer Functions
 Polynomial Format:

 Zero-Pole-Gain (ZPK) Format:

b0,b1,b2… Poles, K: Gain
a0,a1,a2… Zeroes
Useful MATLAB Commands (9)
 To create the first style of transfer function, use the tf
command
 G = tf(num,den);
 num is an array of coefficients in descending order of
power for the numerator (top-half) of the equation
 den is an array of coefficients in descending order of
power for the numerator (bottom-half) of the equation
 The output, G, is a transfer function object that can be
used in a variety of different signal analysis functions
seen in MATLAB
Useful MATLAB Commands (10)
 Example: Let’s say I wanted to create these TFs:

 G1(s): Numerator: [2 1], Denominator: [1 2 3 1]

 G2(s): Numerator: [5], Denominator: [1 -3 2]
 G3(s): Numerator: [10 5 -3], Denominator: [1 5 -4 0 6]
 Get it? The numbers that go beside each power of s go
in decreasing order in the square brackets
 To specify subtraction, use a negative sign for the #!
Useful MATLAB Commands (11)
 Let’s code these transfer function objects
 G1 = tf([2 1], [1 2 3 1]);
 G2 = tf(5, [1 -3 2]);
 G3 = tf([10 5 -3], [1 5 -4 0 6]);
 How about ZPK format? Pretty much the same thing
This is the case where we specifically know the what
the poles, zeroes and gain are
 Use the zpk command to invoke this style of TF
G = zpk([b0 b1 b2…], [a0 a1 a2…], K);
 b0, b1, b2, … are the zero locations
 a0, a1, a2, … are the pole locations, and K the gain
Useful MATLAB Commands (12)
 Note: We can convert a TF in tf form into zpk form
 Example:
G_TF = tf([2 1], [1 2 3 1]);
G_ZPK = zpk(G_TF);
 Example: Let’s say I wanted to
create these TFs:

 G1(s): zeros: [-2 -3], poles: [-1 -5 -6], gain: 3

 G2(s): zeros: [3], poles: [1-i 1+i], gain: 5
 G3(s): zeros: [-1], poles: [1 -2], gain: 1
 Get it? We can specify imaginary poles / zeroes too!
 Factors with addition are negative poles, and vice-versa
Useful MATLAB Commands (13)
 Let’s code these transfer function objects
 G1 = zpk([-2 -3], [-1 -5 -6], 3);
 G2 = zpk(3, [1-i 1+i], 5);
 G3 = zpk(-1, [1 -2], 1);
 There may be some poles or zeroes that are deemed
insignificant
 Their contribution to the overall system will not affect
the output as much
 Criteria: If poles and zeroes are very close to each other,
or if any pole or zero is far away from the imaginary axis
 To get the best TF, we should remove these
Useful MATLAB Commands (14)
 How do we remove this? Use the minreal function
 Gout = minreal(G);
 Input, G, is a transfer function object created either by
tf, or zpk
 What are transfer functions useful for?
 Transfer functions basically represent the input / output
relationship of a system
 We can feed any input into this system, and we’ll see
what the output is
 The two most common inputs that are used for
simulating behaviour are the step and impulse inputs
Useful MATLAB Commands (15)
 Step Input is defined as:

 Impulse Input (in MATLAB) is defined as:

 We can find the step and impulse response, which is

the response of the system to a step or impulse input
 The output is a set of x and y arrays
 x  A set of time values
 y  What the output amplitude is at each value in x
Useful MATLAB Commands (16)
 Something to note: We are specifying the transfer
function in Laplace domain, yet output is time domain
 This is normal. Laplace is used as a tool to get the time-
response more efficiently  Time is what we want
 To calculate the impulse response, use impulse
 impulse(sys);
 Makes new window plotting impulse resp. of TF object, sys
 impulse(sys, Tfin);
 Makes new window plotting impulse resp. of TF object, sys
with a specified ending time, Tfin
 impulse(sys, time_vector);
 Makes a new window plotting impulse resp. of TF object, sys,
at specific time points, given as an array, time_vector
Useful MATLAB Commands (17)
 We can also repeat the same style of invocation by:
 [y,t] = impulse(sys);
 [y,t] = impulse(sys, Tfin);
 [y,t] = impulse(sys, time_vector);
 In this way, we don’t generate a plot, but outputs two
arrays: y, the impulse response values, and t, the time
points at each of these values in y
 This way is useful if you want to plot more than one
impulse response on the same graph
 Use the plot and subplot tools that we’ve seen earlier
Useful MATLAB Commands (18)
 To calculate the step response, use the step
command, and we call it exactly in the same style as
impulse
 step(sys);
 step(sys, Tfin);
 step(sys, time_vector);
 [y,t] = step(sys);
 [y,t] = step(sys, Tfin);
 [y,t] = step(sys, time_vector);
 Useful MATLAB Commands (19)
 Example: Let’s try finding impulse and step response with
these two transfer functions

G1 = tf(5, [1 3 2]);
G2 = tf(3, [1 0 25]);
step(G1,7); xlabel(‘Time’); title(‘Step’);
figure; impulse(G2,4);
xlabel(‘Time’); title(‘Impulse’);
Useful MATLAB Commands (20)
Useful MATLAB Commands (21)
 Let’s try something a bit more advanced
[y1,t1] = step(G1, 0:0.1:7);
[y2,t2] = impulse(G2, 0:0.1:7);
plot(t1,y1,’r-’,t2,y2,’gx-’);
xlabel(‘Time’);
title(‘Step and Impulse Responses’);
legend(‘G1(s)’, ‘G2(s)’);
Useful MATLAB Commands (22)
Useful MATLAB Commands (23)
 MATLAB is a great numerical analysis tool
 We sometimes encounter polynomial equations that are
of 3rd order or higher
 Most conventional calculators can only find up to 3 roots
 MATLAB is able find the roots of any equation you want
 We will look at finding roots for two very popular forms
of equations
 First form is the standard polynomial form:

 Here the equation will have n roots

Useful MALTAB Commands (24)
 The next form is non-linear equations
 Equations that don’t factor nicely like the first form
 We’ll have to use methods like Newton’s method, or any
other numerical root finding tool to find roots
 To find roots using the first form, use roots
 A = roots(array);
 array: is an array of coefficients in descending order,
much like how we saw in the tf command
 Output is an n x 1 array, where n is the total # of roots
Useful MATLAB Commands (25)
 Examples:

 P1 = roots([3 2.5 1 -1 3]);

 P2 = roots([5 0 3 -2 0 -1]);
 … and we thus get:
 P1 = [-0.9195 + 0.8263i; -0.9195 -
0.8263i; 0.5028 + 0.6337i; 0.5028 -
0.6337i];
 P2 = [0.7140; -0.0394 + 0.7440i; -0.0394
- 0.7440i; -0.3176 + 0.6354i; -0.3176 -
0.6354i];
Useful MATLAB Commands (26)
 For non-linear equations, we use fzero
 This finds where the root of a non-linear equation,
based on an initial guess, x0.
 fzero uses a variety of root finding methods to find the
root we want (Newton’s Method, Bisection Method, etc.)
 y = fzero(func, x0);
 func is a function handler of a single variable
 x0 is the initial guess of the root
 Warning: If the non-linear equation has more than 1
root, then we have to specify as more than 1 initial guess
to get all of the roots!
Useful MATLAB Commands (27)
 Example: Let’s try to find some roots of sin(x)
f = @(x) sin(x);
y = fzero(f, 0.5);
y2 = fzero(f, 2.5);
y3 = fzero(f, -4);
 In MATLAB, this will give us:
 y = -1.8428e-28;  ~0 because of approximation
 y2 = 3.1416;  Corresponds to pi
 y3 = -3.1416;  Corresponds to -pi
Useful MATLAB Commands (28)
 Another useful function: Evaluating polynomial
equations at a point, or several points
 You’ll see that this is very useful later when we get into
regression
 Given: A polynomial equation, where its coefficients
are stored in an array  Like what we’ve seen in tf
 val = polyval(coeffs, X);
 coeffs – The array of coefficients in descending order
 X – A single value / vector / matrix of points to evaluate
 val – The same size as X with the equation evaluated at
each corresponding point
Useful MATLAB Commands (29)
 Example: Let’s evaluate at a variety of
different points
coeffs = [3 5 7];
X = [5 7 9];
val = polyval(coeffs,X);
X2 = [1 2; 3 4];
val2 = polyval(coeffs,X2);
 Our outputs are:
 val = [107 189 295];
 val2 = [15 29; 49 75];
Useful MATLAB Commands (30)
 If we have two polynomial equations, with their
coefficients stored in separate arrays, we can use the
convolution operator to multiply the two equations to
get a resultant one
 Use the conv command
 Y = conv(A,B);
 A and B are arrays of coefficients in the style of tf
 Y gives you an array of coefficients that would result if
you multiplied the two polynomial equations together
Useful MATLAB Commands (31)
 Example:

 When we multiply p, with q, we should get:

 When we run this in MATLAB, we’ll see:

P = [1 0 2 4]; Q = [2 0 -1];
A = conv(P,Q);
 Now, we will get: A = [2 0 3 8 -2 -4];
Useful MATLAB Commands (32)
 We can also use MATLAB to perform Partial Fraction
Decomposition (think MTH 240 / MTH 312 / ELE 532)
 We can decompose a fraction that has a set of factored
poles into a sum of each pole scaled by a factor
 What we mean is the following, assuming equation is
in terms of s:

 We have n roots in the equation, and if the equation

has a higher greatest power on the numerator than the
denominator, we have K(s) as well
Useful MATLAB Commands (33)
 We call partial fractions by the residue function
 [R,P,K] = residue(B,A);
 P – the decomposed pole locations (p1,p2,p3…)
 R – The scale factors for each of the fractions (R1,R2,R3…)
 K – The coefficients of the polynomial equation when
the order of the numerator is greater than the
denominator
 If we have repeated roots, then the P and R arrays are
arranged like so:
Useful MATLAB Commands (34)
 Which means:
 R = [… R(j) R(j+1) … R(j+m-1)…];
 P = [… P(j) P(j+1) … P(j+m-1)…];
 Example: Let’s find the PFD of
B = [1 -4];
A = [1 6 11 6];
[R,P,K] = residue(B,A);
 K = [];
 R = [-3.5;6;-2.5];
 P = [-3;-2;-1];
Useful MATLAB Commands (35)
 Couple of things before we take a break:
 If we ever wanted to take a small matrix, and repeat it
horizontally and vertically for a finite amount of times,
we can use the repmat command
 B = repmat(A,M,N);
 M - 1 and N – 1 are the amount of times we want to
repeat the matrix vertically and horizontally respectively
 A is the matrix we want to use to repeat
 Example:
B =

1 2 1 2 1 2
A = [1 2; 3 4]; 3 4 3 4 3 4
1 2 1 2 1 2
B = repmat(A,2,3); 3 4 3 4 3 4
Useful MATLAB Commands (36)
 If we wanted to a column vector, and reshape it into a
matrix, we can use the reshape command
 A = reshape(X,M,N);
 X – A column vector of size MN
 M – The desired columns we want
 N – The desired rows we want
A =
 Example: 1 4 7 10
2 5 8 11
X = (1:12)’; 3 6 9 12
B =
A = reshape(X,4,3);
1 5 9
B = reshape(X,3,4); 2
3
6
7
10
11
4 8 12
 Tutorial Topics – (4)
 2nd Hour: 5:10 p.m. – 6:00 p.m.
 Advanced Mathematical Analysis
 Calculating Derivatives and Integrals Analytically using the
Symbolic Math Toolbox
 Using MATLAB to calculate with respect to “x”

 Evaluating derivatives at points

 Calculating the definite integral

 Finding the line of best fit via regression

 Linear regression, parabolic regression, etc.

 Finding the line of best fit via regression

 Linear regression, parabolic regression, etc.

 More commonly known as polynomial interpolation

 Calculating the area under a curve when the function is unknown,

but data points are known
 Calculating via trapezoidal rule
Advanced Math – (1)
 We can use MATLAB as a tool of analytically
calculating what the integrals and derivatives are
 Can do this in closed form (with respect to “x”)
 Can also calculate what they are at points for derivatives,
or what the definite integral is using MATLAB
 Done by using the Symbolic Math Toolbox
 Here are the basic steps you need to do:
 (a) Tell MATLAB what variables you want to use to
create your functions to integrate / differentiate
 (b) Code your equations in C-style syntax, just like what
we’ve been doing so far
Advanced Math – (2)
 (c) Use these equations and put them into functions
that differentiate or integrate for you
 To do step (a), we use the syms command
 This tells MATLAB what variables are symbolic which
will be used in creating equations
 You can specify more than one variable with spaces
 Examples:
syms x t w alpha;
syms y x;
syms w;
Advanced Math – (3)
 Next, you code equations normally – no function
handlers this time, and no need for point-by-point ops
 Example:
syms x y;
F = 5*sin(x) – x^2 + 3;
G = 25 – x^2 – y^2;
 Now that we’re finished, let’s find the derivative:
 A = diff(F);  Differentiates F(x) wrt the
single variable by default
 B = diff(G,’x’);  Differentiates G(x) wrt x
 C = diff(G,’y’);  Differentiates G(x) wrt y
Advanced Math – (4)
 D = diff(F,n);  Differentiates F(x) n times wrt
single variable by default
 E = diff(G,’x’,n);  Does G(x) wrt x n times
 F = diff(G,’y’,n);  Does G(x) wrt y n times
 Examples:
A = diff(F); A = 5*cos(x) - 2*x
B = diff(G,’x’); B = -2*x

C = diff(G,’y’); C = -2*y

D = diff(F,2); D = -5*sin(x) - 2

E = diff(G,’x’,2); E = -2
Advanced Math – (5)
 We can repeat the same thing with integration
 Use the int command
 Here are all the possible ways you can run this
command
 A = int(F);  Finds the indefinite integral wrt the
single variable by default
 B = int(F,x);  Finds the indefinite integral wrt
x, and leaves other variables constant
 No single quotes here!
 C = int(F,a,b);  Finds the definite integral wrt
single variable between the values a and b
 D = int(F,y,a,b);  Finds the definite integral
wrt y between a & b, & leaves other variables constant
Advanced Math – (6)
 Here are some good ‘ol examples, using the functions
we defined earlier:
A = int(F); A = 3*x - 5*cos(x) - x^3/3

B = int(G,x); B = - x^3/3 + (25 - y^2)*x

C = int(G,y); C = - y^3/3 + (25 - x^2)*y

D = int(F,0,1); D = 23/3 - 5*cos(1)

E = int(G,x,0,1); E = 74/3 – y^2
Advanced Math – (7)
 Next, let’s take a look at finding the line of best fit
 Also known as regression
 Regression minimizes the error between the optimal line,
and with the data points
 Error is represented in terms of least squares
 More commonly known as polynomial interpolation
 Won’t go into the math as it’s too detailed for this kind of
tutorial, but I can tell you what function to use
 For this function, we are given a set of x and y points
 It is now our job to figure out the best polynomial equation
to fit these set of points
Advanced Math – (8)
 In other words, for our pair of x and y points, we want
to find the co-efficients a0,a1,a2 etc., such that:

 How do we do this? Use the polyfit command

 A = polyfit(x,y,N);
 x and y are our data points  Must both be same size!
 N is represents the type of polynomial equation we want
 N = 1  Linear
 N = 2  Parabolic / Quadratic
 N = 3  Cubic, etc. etc.
A = [ ];
Advanced Math – (9)
 A has the coefficients in decreasing order, just like
what we have seen in tf
 Let’s do an example. Let’s say we had the following
data
x = [10 15 20 25 40 50 55 60 75];
y = [5 20 18 40 33 54 70 60 78];
 Let’s try fitting a straight line, and a quadratic through
this data by regression
A = polyfit(x,y,1);
B = polyfit(x,y,2);
Advanced Math – (10)
 Now, let’s plot the points on the graph, as well as what
each line looks like:
xp = 10 : 0.1 : 75;  Range is from 10 to 75
Yline = polyval(A,xp);  Find y pts. along line
Ypara = polyval(B,xp);  Find y pts. along parab.
plot(x,y,’bo’,xp,Yline,’g’,xp,Ypara,’r’);
xlabel(‘x’); ylabel(‘y’);
title(‘Linear vs. Quadratic Regression’);
legend(‘Data’, ‘Linear’, ‘Quadratic’);
Advanced Math – (11)
Advanced Math – (12)
 Going with the set of x,y points, let’s say we wanted to
calculate the area / integral
 We don’t have a closed-form function, so we can’t do it
analytically!
 Instead, we’ll turn to numerical methods
 What we’ll do here is calculate the area under the
curve for a set of x,y points using the trapezoidal rule
 Basically, between each neighbouring pair of points, we
form a trapezoid, and calculate the best fitting area
under this trapezoid
 Add up all these trapezoids together and you get the
total area under the curve
Advanced Math – (13)
 How do we do this? You use the trapz routine!
 How do we call?
area = trapz(x,y);
 x and y are the data points that we know
 area  Gives you the total estimated area underneath
the curve bounded by these x,y points
 Note of caution:
 This is the estimated area
 Forming trapezoids around pairs of points will
inevitably lead to over estimation and underestimation
of the real area
Tutorial Topics – (5)
 Numerical Solutions to Differential Equations
 Output is a set of points that is the solution at different
points in time  Not the actual equation
 Sorting arrays
Advanced Math – (14)
 Last advanced topic before moving on
 Calculating numerical solutions to differential equations
 Most computer scientists and mathematicians now rely
on calculating the solutions to differential equations
numerically
 The output is no longer a function of a variable (x, y,…)
 The output is an array of points
 Each point in the array corresponds what the actual solution /
data point / y-value is for a particular point in time
 There are many methods to solve these numerically, but
what I will show you today deals with matrix algebra
Advanced Math – (15)
 Here are the steps:
 (a) Represent the derivatives in terms of their numerical
estimates
 If we let tk represent the time at index k, and let y(t) be
the actual function / constraint that we know of
 Therefore, for succinctness, y(tk) = yk
 From literature, the numerical estimates for the first and
second derivatives are:

 h represents the step size  The increment in between

successive time points  We will assume uniformly
spaced points here
Advanced Math – (16)
 (b) Calculate step size and time points
 To calculate step size: h = (b – a) / n;
 To calculate time points, do t = a + i*h;
 i is the index in time that we want
 (c) Substitute the numerical approximations of the
derivatives into the differential equations
 (d) Determine a system of equations from i = 0 to i = n
by substituting i = 0, 1, 2, … n into step (c)
 (e) Form into a matrix equation and solve
Advanced Math – (17)
 Example:
 Let’s solve for time values between 0 to 1
 Let’s also assume 100 equally spaced points
 First, let’s figure out what the equation is in terms of I

 For t = 0 = t0, y(0) = 0 = y0

 For t = 1 = tn, y(n) = 0 = yn
Advanced Math – (18)
 So we finally have:

 This creates the following system: (blanks are zeroes)

X Y = F
Advanced Math – (19)
 To solve for the y values, we simply find the inverse of
that system  Y = X-1F
 How do we code this?
% Define constants
n = 100;
a = 0; b = 1; h = (b – a)/n;
t = a + (1:100)’*h; % Define time values
f = @(t) 125*t; % Define constraint
F = 2*(h^2)*f(t); % Create RHS vector
X = zeros(n,n); % Create LHS matrix
Advanced Math – (20)
% Create first and last row
X(1,1) = -4;
X(1,2) = 2-h;
X(n,n-1) = 2+h;
X(n,n) = -4;
% Create the rest of the rows
for i = 2 : n-1
X(i,i-1:i+1) = [2+h -4 2-h];
end
Advanced Math – (21)
% Find the inverse of the system
Y = X \ F;
% Plot the graph
plot(t,Y);
xlabel(‘Time (s)’);
ylabel(‘Amplitude’);
title(‘Solution to Diff. Eqn.);
grid;
Advanced Math – (22)
Advanced MATLAB – (23)
 Next advanced function  Sorting
 Very useful!
 If we have an array that we want to sort, use the sort
command
 Y = sort(A,’ascend’);
 Y = sort(A,’descend’);
 [Y,I] = sort(A,’ascend’);
 [Y,I] = sort(A,’descend’);
 First two sort the array in ascending or descending
order & descending order then place in Y
Advanced MATLAB – (24)
 Next two not only give you the sorted array, but it gives
you the indices of where the original values came from
for each corresponding position
 Example: A = [1 5 7 2 4];
Y = sort(A,’ascend’);
 Y = [1 2 4 5 7];
Y = sort(A,’descend’);
 Y = [7 5 4 2 1];
[Y,I] = sort(A,’ascend’);
 Y = [1 2 4 5 7]; I = [1 4 5 2 3];
[Y,I] = sort(A,’descend’);
 Y = [7 5 4 2 1]; I = [3 2 5 4 1];
END!
 In this tutorial, I’ve tried to show you some advanced
concepts that you may use later on in your courses /
thesis
 This is obviously not an exhaustive tutorial! Check
MathWorks forums and Google  They’re your friends
for MATLAB coding
 You’re more than welcome to contact me for MATLAB
help  If I have the time, I will help you
 Last but not least, thanks for coming out!