Interactive Computing with Matlab

Gerald W. Recktenwald Department of Mechanical Engineering Portland State University [email protected]

These slides are a supplement to the book Numerical Methods with Matlab: Implementations and Applications, by Gerald W. Recktenwald, c 20002006, Prentice-Hall, Upper Saddle River, NJ. These slides are copyright c 20002006 Gerald W. Recktenwald. The PDF version of these slides may be downloaded or stored or printed only for noncommercial, educational use. The repackaging or sale of these slides in any form, without written consent of the author, is prohibited. The latest version of this PDF le, along with other supplemental material for the book, can be found at www.prenhall.com/recktenwald or web.cecs.pdx.edu/~gerry/nmm/.

Version 1.1

August 21, 2006

page 1

Overview

(1)

Basic Matlab Operations Starting Matlab Using Matlab as a calculator Introduction to variables and functions Matrices and Vectors: All variables are matrices. Creating matrices and vectors Subscript notation Colon notation

NMM: Interactive Computing with Matlab

page 2

Overview
Additional Types of Variables Complex numbers Strings Polynomials Working with Matrices and Vectors Some basic linear algebra Vectorized operations Array operators Managing the Interactive Environment Plotting
NMM: Interactive Computing with Matlab

(2)

page 3

Starting Matlab
Double click on the Matlab icon, or on unix systems type matlab at the command line. After startup Matlab displays a command window that is used to enter commands and display text-only results. Enter Commands at the command prompt: >> EDU> for full version for educational version

Matlab responds to commands by printing text in the command window, or by opening a gure window for graphical output. Toggle between windows by clicking on them with the mouse.
NMM: Interactive Computing with Matlab page 4

Matlab Desktop
Command Prompt:
Enter typed commands here. Text results are displayed here.

Recent Directory Menu:

Used to change current working directory.

Launch Pad/Workspace:
Used to browse documentation, or view values of variables in the workspace Select Launch Pad tab or Workspace tab

Command History/Current Directory: View and re-enter

previously typed commands, or change directories Select Command History tab or Current Directory tab

NMM: Interactive Computing with Matlab

page 5

Matlab Desktop
The desktop provides dierent ways of interacting with Matlab Entering commands in the command window Viewing values stored in variables Editing statements in Matlab functions and scripts Creating and annotating plots

Watch an animated demonstration of the Matlab desktop by typing

playbackdemo(desktop)

at the command prompt.

NMM: Interactive Computing with Matlab

page 6

Matlab as a Calculator
Enter formulas at the command prompt
>> 2 + 6 - 4 ans = 4 >> ans/2 ans = 2 (press return after 4)

(1)

NMM: Interactive Computing with Matlab

page 7

Matlab as a Calculator
Dene and use variables
>> a = 5 a = 5 >> b = 6 b = 6 >> c = b/a c = 1.2000

(2)

NMM: Interactive Computing with Matlab

page 8

Built-in Variables
pi (= ) and ans are a built-in variables
>> pi ans = 3.1416 >> sin(ans/4) ans = 0.7071

Note: There is no degrees mode. All angles are measured in radians.

NMM: Interactive Computing with Matlab

page 9

Built-in Functions
Many standard mathematical functions, such as sin, cos, log, and log10, are built-in
>> log(256) ans = 5.5452 >> log10(256) ans = 2.4082 >> log2(256) ans = 8 log(x) computes the natural logarithm of x

log10(x) is the base 10 logarithm

log2(x) is the base 2 logarithm

NMM: Interactive Computing with Matlab

page 10

Ways to Get Help

Use on-line help to request info on a specic function
>> help sqrt

In Matlab version 6 and later the doc function opens the on-line version of the manual. This is very helpful for more complex commands.
>> doc plot

Use lookfor to nd functions by keywords

>> lookfor functionName

NMM: Interactive Computing with Matlab

page 11

On-line Help
Syntax:
help functionName

(1)

Example:
>> help log

produces
LOG Natural logarithm. LOG(X) is the natural logarithm of the elements of X. Complex results are produced if X is not positive. See also LOG2, LOG10, EXP, LOGM.

The help function provides a compact summary of how to use a command. Use the doc function to get more in-depth information.
NMM: Interactive Computing with Matlab page 12

On-line Help
>> doc plot

(2)

The help browser opens when you type a doc command:

NMM: Interactive Computing with Matlab

page 13

Looking for Functions

Syntax:
lookfor string

searches rst line of function descriptions for string. Example:

>> lookfor cosine

produces
ACOS ACOSH COS COSH Inverse cosine. Inverse hyperbolic cosine. Cosine. Hyperbolic cosine.

NMM: Interactive Computing with Matlab

page 14

Strategies for Interactive Computing

Use the command window for short sequences of calculations Later well learn how to build reusable functions for more complex tasks. The command window is good for testing ideas and running sequences of operations contained in functions Any command executed in the command window can also be used in a function. Lets continue with a tour of interactive computing.

NMM: Interactive Computing with Matlab

page 15

Suppress Output with Semicolon

(1)

Results of intermediate steps can be suppressed with semicolons. Example: Assign values to x, y, and z, but only display the value of z in the command window:
>> x = 5; >> y = sqrt(59); >> z = log(y) + x^0.25 z = 3.5341

NMM: Interactive Computing with Matlab

page 16

Suppress Output with Semicolon

(2)

Type variable name and omit the semicolon to print the value of a variable (that is already dened)
>> x = 5; >> y = sqrt(59); >> z = log(y) + x^0.25 z = 3.5341 >> y y = 7.6811

( = log(sqrt(59)) + 5^0.25 )

NMM: Interactive Computing with Matlab

page 17

Multiple Statements per Line

Use commas or semicolons to enter more than one statement at once. Commas allow multiple statements per line without suppressing output.
>> a = 5; b = sin(a), b = -0.9589 c = 74.2099 c = cosh(a)

NMM: Interactive Computing with Matlab

page 18

Matlab Variables Names

Legal variable names: Begin with one of az or AZ Have remaining characters chosen from az, AZ, 09, or Have a maximum length of 31 characters Should not be the name of a built-in variable, built-in function, or user-dened function Examples:
xxxxxxxxx pipeRadius widgets_per_box mySum mysum
NMM: Interactive Computing with Matlab

Note: mySum and mysum are dierent variables. Matlab is case sensitive.

page 19

Built-in Matlab Variables

Name ans eps pi realmax realmin Inf NaN Meaning

(1)

value of an expression when that expression is not assigned to a variable oating point precision

(3.141492 . . .)

largest positive oating point number smallest positive oating point number

, a number larger than realmax,

the result of evaluating 1/0. not a number, the result of evaluating 0/0

NMM: Interactive Computing with Matlab

page 20

Built-in Matlab Variables

(2)

Rule: Only use built-in variables on the right hand side of an expression. Reassigning the value of a built-in variable can create problems with built-in functions. Exception: i and j are preassigned to 1. One or both of i or j are often reassigned as loop indices. More on this later.

NMM: Interactive Computing with Matlab

page 21

Matrices and Vectors

All Matlab variables are matrices A vector is a matrix with one row or one column. A scalar is a matrix with one row and one column. A character string is a row of column vector of characters. Consequences: Rules of linear algebra apply to addition, subtraction and multiplication. Elements in the vectors and matrices are addressed with Fortran-like subscript notation, e.g.,, x(2), A(4,5). Usually this notation is clear from context, but it can be confused with a function call,
y = sqrt(5) z = f(3)
NMM: Interactive Computing with Matlab

sqrt is a built-in function Is f a function or variable?

page 22

Creating Matlab Variables

Matlab variables are created with an assignment statement
>> x = expression

where expression is a legal combinations of numerical values, mathematical operators, variables, and function calls. The expression can involve: Manual entry Built-in functions that return matrices Custom (user-written) functions that return matrices Loading matrices from text les or mat les

NMM: Interactive Computing with Matlab

page 23

Element-by-Element Creation of Matrices and Vectors

A matrix, a column vector, and a row vector: 3 2 A = 3 1 1 4 5 7 x= 9 2 v = 9 3 4 1
NMM: Interactive Computing with Matlab

(1)

As Matlab variables:
>> A = [3 2; 3 1; 1 4] A = 3 2 3 1 1 4 >> x = [5; 7; 9; 2] x = 5 7 9 2 >> v = [9 -3 4 1] v = 9 -3 4 1

page 24

Element-by-Element Creation of Matrices and Vectors

(2)

For manual entry, the elements in a vector are enclosed in square brackets. When creating a row vector, separate elements with a space.
>> v = [7 3 9] v = 7 3 9

Separate columns with a semicolon

>> w = [2; 6; 1] w = 2 6 1

NMM: Interactive Computing with Matlab

page 25

Element-by-Element Creation of Matrices and Vectors

(3)

When assigning elements to matrix, row elements are separated by spaces, and columns are separated by semicolons
>> A = [1 2 3; 5 7 11; 13 17 19] A = 1 2 3 5 7 11 13 17 19

NMM: Interactive Computing with Matlab

page 26

Transpose Operator

(1)

Once it is created, a variable can be transformed with other operators. The transpose operator converts a row vector to a column vector (and vice versa), and it changes the rows of a matrix to columns.
>> v = [2 4 1 7] v = 2 4 1 >> w = v w = 2 4 1 7

NMM: Interactive Computing with Matlab

page 27

Transpose Operator
>> A = [1 2 3; 4 5 6; 7 8 9 ] A = 1 2 3 4 5 6 7 8 9 >> B = A B = 1 2 3

(2)

4 5 6

7 8 9

NMM: Interactive Computing with Matlab

page 28

Overwriting Variables
Once a variable has been created, it can be reassigned
>> x = 2; >> x = x + 2 x = 4 >> y = [1 2 3 4] y = 1 2 3 >> y = y y = 1 2 3 4
NMM: Interactive Computing with Matlab page 29

Using Functions to Create Matrices and Vectors

Create vectors with built-in functions: linspace and logspace Create matrices with built-in functions: ones, zeros, eye, diag, . . . Note that ones and zeros can also be used to create vectors.

NMM: Interactive Computing with Matlab

page 30

Creating vectors with linspace

(1)

The linspace function creates vectors with elements having uniform linear spacing. Syntax:
x = linspace(startValue,endValue) x = linspace(startValue,endValue,nelements)

Examples:
>> u = linspace(0.0,0.25,5) u = 0 0.0625 0.1250 >> u = linspace(0.0,0.25);

0.1875

0.2500

Remember: Ending a statement with semicolon suppresses the output.

NMM: Interactive Computing with Matlab page 31

Creating vectors with linspace

(2)

Column vectors are created by appending the transpose operator to linspace

>> v = linspace(0,9,4) v = 0 3 6 9

NMM: Interactive Computing with Matlab

page 32

Example: A Table of Trig Functions

>> x = linspace(0,2*pi,6); >> y = sin(x); >> z = cos(x); >> [x y z] ans = 0 0 1.0000 1.2566 0.9511 0.3090 2.5133 0.5878 -0.8090 3.7699 -0.5878 -0.8090 5.0265 -0.9511 0.3090 6.2832 0 1.0000 (note transpose)

The expressions y = sin(x) and z = cos(x) take advantage of vectorization. If the input to a vectorized function is a vector or matrix, the output is often a vector or matrix having the same shape. More on this later.
NMM: Interactive Computing with Matlab page 33

Creating vectors with logspace

The logspace function creates vectors with elements having uniform logarithmic spacing. Syntax:
x = logspace(startValue,endValue) x = logspace(startValue,endValue,nelements)

creates nelements elements between 10startValue and 10endValue. The default value of nelements is 100. Example:
>> w = logspace(1,4,4) w = 10 100

1000

10000

NMM: Interactive Computing with Matlab

page 34

Functions to Create Matrices

Name diag eye ones rand zeros linspace logspace Operation(s) Performed

(1)

create a matrix with a specied diagonal entries, or extract diagonal entries of a matrix create an identity matrix create a matrix lled with ones create a matrix lled with random numbers create a matrix lled with zeros create a row vector of linearly spaced elements create a row vector of logarithmically spaced elements

NMM: Interactive Computing with Matlab

page 35

Functions to Create Matrices

Syntax:
A = ones(nrows,ncols) A = zeros(nrows,ncols)

(2)

Use ones and zeros to set intial values of a matrix or vector.

Examples:
>> D = ones(3,3) D = 1 1 1 1 1 1 1 1 1 >> E = ones(2,4) E = 1 1 1 1 1 1
NMM: Interactive Computing with Matlab

1 1
page 36

Functions to Create Matrices

(3)

ones and zeros are also used to create vectors. To do so, set either nrows or ncols to 1.
>> s = ones(1,4) s = 1 1 1 >> t = zeros(3,1) t = 0 0 0

NMM: Interactive Computing with Matlab

page 37

Functions to Create Matrices

(4)

The eye function creates identity matrices of a specied size. It can also create non-square matrices with ones on the main diagonal. Syntax:
A = eye(n) A = eye(nrows,ncols)

Examples:
>> C = eye(5) C = 1 0 0 1 0 0 0 0 0 0

0 0 1 0 0

0 0 0 1 0

0 0 0 0 1
page 38

NMM: Interactive Computing with Matlab

Functions to Create Matrices

(5)

The optional second input argument to eye allows non-square matrices to be created.
>> D = eye(3,5) D = 1 0 0 1 0 0

0 0 1

0 0 0

0 0 0

where Di,j = 1 whenever i = j.

NMM: Interactive Computing with Matlab

page 39

Functions to Create Matrices

(6)

The diag function can either create a matrix with specied diagonal elements, or extract the diagonal elements from a matrix Syntax:
A = diag(v) v = diag(A)

Example: Use diag to create a matrix

>> v = [1 2 3]; >> A = diag(v) A = 1 0 0 2 0 0

0 0 3

NMM: Interactive Computing with Matlab

page 40

Functions to Create Matrices

(7)

Example: Use diag to extract the diagonal of a matrix

>> B = [1:4; 5:8; 9:12] B = 1 2 3 4 5 6 7 8 9 10 11 12 >> w = diag(B) w = 1 6 11

NMM: Interactive Computing with Matlab

page 41

Functions to Create Matrices

(8)

The action of the diag function depends on the characteristics and number of the input(s). This polymorphic behavior of Matlab functions is common. Refer to the on-line documentation for the possible variations.
>> A = diag([3 2 1]) A = 3 0 0 0 2 0 0 0 1 >> B = [4 2 2; 3 6 9; 1 1 7]; >> v = diag(B) v = 4 6 7
NMM: Interactive Computing with Matlab

Create a matrix with a specied diagonal

Extract the diagonal of a matrix

page 42

Subscript Notation

(1)

If A is a matrix, A(i,j) selects the element in the ith row and jth column. Subscript notation can be used on the right hand side of an expression to refer to a matrix element.
>> A = [1 2 3; 4 5 6; 7 8 9]; >> b = A(3,2) b = 8 >> c = A(1,1) c = 1

NMM: Interactive Computing with Matlab

page 43

Subscript Notation

(1)

Subscript notation is also used to assign matrix elements

>> A(1,1) = c/b A = 0.2500 2.0000 4.0000 5.0000 7.0000 8.0000

3.0000 6.0000 9.0000

Referring to elements beyond the dimensions the matrix results in an error

>> A = [1 2 3; 4 5 6; 7 8 9]; >> A(1,4) ??? Index exceeds matrix dimensions.

NMM: Interactive Computing with Matlab

page 44

Subscript Notation

(1)

Assigning an element that is beyond the existing dimensions of the matrix causes the matrix to be resized!
>> A = [1 2 3; 4 5 6; 7 8 9]; A = 1 2 3 4 5 6 7 8 9 >> A(4,4) = 11 A = 1 2 4 5 7 8 0 0

3 6 9 0

0 0 0 11

In other words, Matlab automatically resizes matrices on the y.

NMM: Interactive Computing with Matlab page 45

Colon Notation

(1)

Colon notation is very powerful and very important in the eective use of Matlab. The colon is used as both an operator and as a wildcard. Use colon notation to: create vectors refer to or extract ranges of matrix elements

NMM: Interactive Computing with Matlab

page 46

Colon Notation
Syntax:
startValue:endValue startValue:increment:endValue

(2)

Note: startValue, increment, and endValue do not need to be integers

NMM: Interactive Computing with Matlab

page 47

Colon Notation
Creating row vectors:
>> s = 1:4 s = 1 2

(3)

>> t = 0:0.1:0.4 t = 0 0.1000

0.2000

0.3000

0.4000

NMM: Interactive Computing with Matlab

page 48

Colon Notation
Creating column vectors:
>> u = (1:5) u = 1 2 3 4 5 >> v = 1:5 v = 1 2

(4)

v is a row vector because 1:5 creates a vector between 1 and the transpose of 5.
NMM: Interactive Computing with Matlab page 49

Colon Notation

(5)

Use colon as a wildcard to refer to an entire column or row

>> A = [1 2 3; 4 5 6; 7 8 9]; >> A(:,1) ans = 1 4 7 >> A(2,:) ans = 4

NMM: Interactive Computing with Matlab

page 50

Colon Notation

(6)

Or use colon notation to refer to subsets of columns or rows

>> A(2:3,1) ans = 4 7 >> A(1:2,2:3) ans = ans = 2 3 5 6

NMM: Interactive Computing with Matlab

page 51

Colon Notation

(7)

Colon notation is often used in compact expressions to obtain results that would otherwise require several steps. Example:
>> A = ones(8,8); >> A(3:6,3:6) = zeros(4,4) A = 1 1 1 1 1 1 1 1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 1 1 1 1 1 1
NMM: Interactive Computing with Matlab

1 1 0 0 0 0 1 1

1 1 0 0 0 0 1 1

1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1
page 52

Colon Notation

(8)

Finally, colon notation is used to convert any vector or matrix to a column vector. Example:
>> x = 1:4; >> y = x(:) y = 1 2 3 4

NMM: Interactive Computing with Matlab

page 53

Colon Notation

(9)

Colon notation converts a matrix to a column vector by appending the columns of the input matrix
>> A = rand(2,3); >> v = A(:) v = 0.9501 0.2311 0.6068 0.4860 0.8913 0.7621 0.4565 Note: The rand function generates random elements between zero and one. Repeating the preceding statements will, in all likelihood, produce dierent numerical values for the elements of v.
NMM: Interactive Computing with Matlab page 54

Additional Types of Variables

Matrix elements can either be numeric values or characters. Numeric elements can either be real or complex (imaginary). More general variable types are available: n-dimensional arrays (where n > 2), structs, cell arrays, and objects. Numeric (real and complex) and string arrays of dimension two or less will be sucient for our purposes. Consider some simple variations on numeric and string matrices: Complex Numbers Strings Polynomials

NMM: Interactive Computing with Matlab

page 55

Complex Numbers
Matlab automatically performs complex arithmetic
>> sqrt(-4) ans = 0 + 2.0000i >> x = 1 + 2*i x = 1.0000 + 2.0000i >> y = 1 - 2*i y = 1.0000 - 2.0000i >> z = x*y z = 5
NMM: Interactive Computing with Matlab page 56

(or,

x = 1 + 2*j)

Unit Imaginary Numbers

(1)

i and j are ordinary Matlab variables preassigned with the value

>> i^2 ans = -1

1.

Both or either i and j can be reassigned

>> i = 5; >> t = 8; >> u = sqrt(i-t) u = 0 + 1.7321i >> u*u ans = -3.0000

(i-t = -3,

not -8+i)

NMM: Interactive Computing with Matlab

page 57

Unit Imaginary Numbers

(2)

The i and j variables are often used for array subscripting. Set i (or j) to an integer value that is also a valid array subscript.
>> A = [1 2; 3 4]; >> i = 2; >> A(i,i) = 1 A = 1 2 3 1 >> x = A(2,j) ??? Subscript indices must either be real positive integers or logicals.

Note: When working with complex numbers, it is a good idea to reserve either i or j for the unit imaginary value 1.
NMM: Interactive Computing with Matlab page 58

Euler Notation

(1)

Euler notation represents a complex number by a phaser

z = e

imaginary iy x real z = ei

x = Re(z) = |z| cos() = cos()

y = iIm(z) = i|z| sin() = i sin()

NMM: Interactive Computing with Matlab

page 59

Functions for Complex Arithmetic

Function abs angle exp Operation Compute the magnitude of a number

(1)

abs(z) is equivalent to sqrt( real(z)^2 + imag(z)^2 ) Angle of complex number in Euler notation If x is real, If z is complex, conj imag real exp(x) = ex exp(z) = eRe(z)(cos(Im(z) + i sin(Im(z))

Complex conjugate of a number Extract the imaginary part of a complex number Extract the real part of a complex number

NMM: Interactive Computing with Matlab

page 60

Functions for Complex Arithmetic

Examples:
>> zeta = 5; theta = pi/3; >> z = zeta*exp(i*theta) z = 2.5000 + 4.3301i >> abs(z) ans = 5 >> sqrt(z*conj(z)) ans = 5 >> x = real(z) x = 2.5000 >> y = imag(z) y = 4.3301 >> angle(z)*180/pi ans = 60.0000

(2)

Remember: There is no degrees mode in Matlab. All angles are in radians.

NMM: Interactive Computing with Matlab page 61

Strings
Strings are matrices with character elements. String constants are enclosed in single quotes Colon notation and subscript operations apply Examples:
>> first = John; >> last = Coltrane; >> name = [first, ,last] name = John Coltrane >> length(name) ans = 13 >> name(9:13) ans = trane

NMM: Interactive Computing with Matlab

page 62

Functions for String Manipulation

Function char findstr length num2str str2num strcmp strmatch strncmp sprintf Operation

(1)

Converts an integer to the character using ASCII codes, or combines characters into a character matrix Finds one string in another string Returns the number of characters in a string Converts a number to string Converts a string to a number Compares two strings Identies rows of a character array that begin with a string Compares the rst n elements of two strings Converts strings and numeric values to a string

NMM: Interactive Computing with Matlab

page 63

Functions for String Manipulation

num2str converts a number to a string

(2)

>> msg1 = [There are ,num2str(100/2.54), inches in a meter] msg1 = There are 39.3701 inches in a meter

For greater control over format of the number-to-string conversion, use sprintf
>> msg2 = sprintf(There are %5.2f cubic inches in a liter,1000/2.54^3) msg2 = There are 61.02 cubic inches in a liter

The Matlab sprintf function is similar to the C function of the same name, but it uses single quotes for the format string.

NMM: Interactive Computing with Matlab

page 64

Functions for String Manipulation

The char function can be used to combine strings
>> both = char(msg1,msg2) both = There are 39.3701 inches in a meter There are 61.02 cubic inches in a liter

(3)

or to refer to individual characters by their ASCII codes1

>> char(49) ans = 1 >> char([77 65 84 76 65 66]) ans = MATLAB
1

See e.g.,, www.asciicodes.com or wikipedia.org/wiki/ASCII.

page 65

NMM: Interactive Computing with Matlab

Functions for String Manipulation

(4)

Use strcmp to test whether two strings are equal, i.e., if they contain the same sequence of characters.
>> msg1 = [There are ,num2str(100/2.54), inches in a meter]; >> msg2 = sprintf(There are %5.2f cubic inches in a liter,1000/2.54^3); >> strcmp(msg1,msg2) ans = 0

Compare the rst n characters of two strings with strncmp

>> strncmp(msg1,msg2,9) ans = 1

The rst nine characters of both strings are There are, so strncmp(msg1,msg2,9) returns 1, or true.
NMM: Interactive Computing with Matlab page 66

Functions for String Manipulation

(5)

Locate occurances of one string in another string with findstr

>> findstr(in,msg1) ans = 19 26 >> msg1(19:20) ans = in

NMM: Interactive Computing with Matlab

page 67

Polynomials
Matlab polynomials are stored as vectors of coecients. The polynomial coecients are stored in decreasing powers of x Pn(x) = c1xn + c2xn1 + . . . + cnx + cn+1 Example: Evaluate x3 2x + 12 at x = 1.5 Store the coecients of the polynomial in vector c:
>> c = [1 0 -2 12];

Use the built-in polyval function to evaluate the polynomial.

>> polyval(c,1.5) ans = 12.3750
NMM: Interactive Computing with Matlab page 68

Functions for Manipulating Polynomials

Function conv deconv poly polyder polyval polyfit roots

Operations performed Product (convolution) of two polynomials Division (deconvolution) of two polynomials Create a polynomial having specied roots Dierentiate a polynomial Evaluate a polynomial Polynomial curve t Find roots of a polynomial

NMM: Interactive Computing with Matlab

page 69

A Quick Overview of Linear Algebra in Matlab

The name Matlab is a shortened form of MATrix LABoratory. Matlab data types and syntax make it easy to perform standard operations of linear algebra including addition, subtraction, and multiplication of vectors and matrices. Chapter 7 provides a detailed review of linear algebra. Here we provide a simple introduction to some operations that are necessary for routine calculation. Vector addition and subtraction Inner and outer products Vectorization Array operators
NMM: Interactive Computing with Matlab page 70

Vector Addition and Subtraction

Vector and addition and subtraction are element-by-element operations. Example:
>> u = [10 9 8]; >> v = [1 2 3]; >> u+v ans = 11 11 11 >> u-v ans = 9 (u and v are row vectors)

NMM: Interactive Computing with Matlab

page 71

Vector Inner and Outer Products

The inner product combines two vectors to form a scalar = u v = u v T = ui vi

The outer product combines two vectors to form a matrix A = uT v ai,j = ui vj

NMM: Interactive Computing with Matlab

page 72

Inner and Outer Products in Matlab

Inner and outer products are supported in Matlab as natural extensions of the multiplication operator
>> u = [10 9 8]; >> v = [1 2 3]; >> u*v ans = 52 >> u*v ans = 10 9 8 (u and v are row vectors) (inner product)

(outer product) 20 18 16 30 27 24

NMM: Interactive Computing with Matlab

page 73

Vectorization
Vectorization is the use of single, compact expressions that operate on all elements of a vector without explicitly writing the code for a loop. The loop is executed by the Matlab kernel, which is much more ecient at evaluating a loop in interpreted Matlab code. Vectorization allows calculations to be expressed succintly so that programmers get a high level (as opposed to detailed) view of the operations being performed. Vectorization is important to make Matlab operate eciently2.

Recent versions of Matlab have improved the eciency for some non-vectorized code.
page 74

NMM: Interactive Computing with Matlab

Vectorization of Built-in Functions

Most built-in function support vectorized operations. If the input is a scalar the result is a scalar. If the input is a vector or matrix, the output is a vector or matrix with the same number of rows and columns as the input. Example:
>> x = 0:pi/4:pi x = 0 0.7854 >> y = cos(x) y = 1.0000 0.7071 (dene a row vector) 1.5708 2.3562 3.1416

(evaluate cosine of each x(i)) 0 -0.7071 -1.0000

NMM: Interactive Computing with Matlab

page 75

Contrast with FORTRAN Implementation

The Matlab statements
x = 0:pi/4:pi; y = cos(x);

are equivalent to the following FORTRAN code

real x(5),y(5) pi = 3.14159624 dx = pi/4.0 do 10 i=1,5 x(i) = (i-1)*dx y(i) = sin(x(i)) continue

10

No explicit loop is necessary in Matlab.

NMM: Interactive Computing with Matlab page 76

Vectorized Calculations
More examples
>> A = pi*[ 1 2; 3 4] A = 3.1416 6.2832 9.4248 12.5664 >> S = sin(A) S = 0 0 0 0

(6)

>> B = A/2 B = 1.5708 4.7124 >> T = sin(B) T = 1 0 -1 0

3.1416 6.2832

NMM: Interactive Computing with Matlab

page 77

Array Operators
Array operators support element-by-element operations that are not dened by the rules of linear algebra. Array operators have a period prepended to a standard operator. Symbol .* ./ .\ .^ Operation element-by-element multiplication element-by-element right division element-by-element left division element-by-element exponentiation

Array operators are a very important tool for writing vectorized code.

NMM: Interactive Computing with Matlab

page 78

Using Array Operators

(1)

Examples: Element-by-element multiplication and division

>> u = [1 2 3]; >> v = [4 5 6];

Use .* and ./ for element-by-element multiplication and division

>> w = u.*v w = 4 10 >> x = u./v x = 0.2500

18

0.4000

0.5000

NMM: Interactive Computing with Matlab

page 79

Using Array Operators

(1)

Examples: Element-by-element multiplication and division

>> u = [1 2 3]; >> v = [4 5 6]; >> y = sin(pi*u/2) .* cos(pi*v/2) y = 1 0 1 >> z = sin(pi*u/2) ./ cos(pi*v/2) Warning: Divide by zero. z = 1 NaN 1

NMM: Interactive Computing with Matlab

page 80

Using Array Operators

Examples: Application to matrices
>> A = [1 >> B = [8 >> A.*B ans = 8 20 2 3 4; 5 6 7 8]; 7 6 5; 4 3 2 1];

(2)

14 18

18 14

20 8

>> A*B ??? Error using ==> * Inner matrix dimensions must agree.

The last statement causes an error because the number of columns in A is not equal to the number of rows in B a requirement for A and B to be compatible for matrix multiplication.
NMM: Interactive Computing with Matlab page 81

Using Array Operators

>> A = [1 >> B = [8 >> A*B ans = 60 164 2 3 4; 5 6 7 8]; 7 6 5; 4 3 2 1];

(3)

20 60

The number of columns in A is equal to the number of rows in BT , so A*B is a legal matrix-matrix multiplication. Array operators also apply to matrix powers.
>> A.^2 ans = 1 25

4 36

9 49

16 64
page 82

NMM: Interactive Computing with Matlab

The Matlab Workspace

(1)

All variables dened as the result of entering statements in the command window, exist in the Matlab workspace. At the beginning of a Matlab session, the workspace is empty. Being aware of the workspace allows you to Create, assign, and delete variables Load data from external les Manipulate the Matlab path

NMM: Interactive Computing with Matlab

page 83

The Matlab Workspace

(2)

The clear command deletes variables from the workspace. The who command lists the names of variables in the workspace
>> clear >> who (Delete all variables from the workspace) (No response, no variables are dened after clear) >> a = 5; b = 2; c = 1; >> d(1) = sqrt(b^2 - 4*a*c); >> d(2) = -d(1); >> who Your variables are: a b c d

NMM: Interactive Computing with Matlab

page 84

The Matlab Workspace

(3)

The whos command lists the name, size, memory allocation, and the class of each variables dened in the workspace.
>> whos Name a b c d Size 1x1 1x1 1x1 1x2 Bytes 8 8 8 32 Class double double double double array array array array (complex)

Grand total is 5 elements using 56 bytes

NMM: Interactive Computing with Matlab

page 85

The Matlab Workspace

(4)

The whos command returns the array dimensions, memory requirement and class for each variable in the workspace. Built-in variable classes are double, char, sparse, struct, and cell. The class of a variable determines the type of data that can be stored in it. We will be dealing primarily with numeric data, which is the double class, and occasionally with string data, which is in the char class.

NMM: Interactive Computing with Matlab

page 86

Working with External Data Files

Write data to a le
save fileName save fileName variable1 variable2 ... save fileName variable1 variable2 ... -ascii

Read in data stored in matrices

load fileName load fileName matrixVariable

NMM: Interactive Computing with Matlab

page 87

Loading Data from External File

Example: Load data from a le and plot the data
>> >> >> >> load wolfSun.dat; xdata = wolfSun(:,1); ydata = wolfSun(:,2); plot(xdata,ydata)

NMM: Interactive Computing with Matlab

page 88

The Matlab Path

Matlab will only use those functions and data les that are in its path. To add N:\IMAUSER\ME352\PS2 to the path, type
>> p = path; >> path(p,N:\IMAUSER\ME352\PS2);

Matlab version 5 and later has an interactive path editor that makes it easy to adjust the path. The path specication string depends on the operating system. On a Unix/Linux computer a path setting operation might look like:
>> p = path; >> path(p,~/matlab/ME352/ps2);
NMM: Interactive Computing with Matlab page 89

Plotting

Plotting (x, y) data Axis scaling and annotation 2D (contour) and 3D (surface) plotting

NMM: Interactive Computing with Matlab

page 90

Plotting (x, y) Data

(1)

Two dimensional plots are created with the plot function Syntax:
plot(x,y) plot(xdata,ydata,symbol) plot(x1,y1,x2,y2,...) plot(x1,y1,symbol1,x2,y2,symbol2,...)

Note: x and y must have the same shape, x1 and y1 must have the same shape, x2 and y2 must have the same shape, etc.

NMM: Interactive Computing with Matlab

page 91

Plotting (x, y) Data

Example: A simple line plot
>> x = linspace(0,2*pi); >> y = sin(x); >> plot(x,y);
1 0.8 0.6 0.4 0.2 0 0.2 0.4 0.6 0.8 1 0 1 2 3

(2)

NMM: Interactive Computing with Matlab

page 92

Line and Symbol Types

(1)

The curves for a data set are drawn from combinations of the color, symbol, and line types in the following table.
Color y m c r g b w k yellow cyan red green blue white black . x + * s d v point circle x-mark plus star square diamond triangle (down) magenta o Symbols ^ < > p h triangle (up) triangle (left) triangle (right) pentagram hexagram : -. -Line solid dotted dashdot dashed

To choose a color/symbol/line style, chose one entry from each column.

NMM: Interactive Computing with Matlab page 93

Line and Symbol Types

Examples: Put yellow circles at the data points:
plot(x,y,yo)

(2)

Plot a red dashed line with no symbols:

plot(x,y,r--)

Put black diamonds at each data point and connect the diamonds with black dashed lines:
plot(x,y,kd--)

NMM: Interactive Computing with Matlab

page 94

Alternative Axis Scaling

(1)

Combinations of linear and logarithmic scaling are obtained with functions that, other than their name, have the same syntax as the plot function. Name loglog plot semilogx semilogy Axis scaling log10(y) versus log10(x) linear y versus x linear y versus log10(x) log10(y) versus linear x

Note: As expected, use of logarithmic axis scaling for data sets with negative or zero values results in a error. Matlab will complain and then plot only the positive (nonzero) data.
NMM: Interactive Computing with Matlab page 95

Alternative Axis Scaling

Example:
>> x = linspace(0,3); >> y = 10*exp(-2*x); >> plot(x,y);
10 8 6 4 2 0 0
1

(2)

>> semilogy(x,y);

10 10 10 10

-1

-2

NMM: Interactive Computing with Matlab

page 96

Multiple plots per gure window

(1)

The subplot function is used to create a matrix of plots in a single gure window. Syntax:
subplot(nrows,ncols,thisPlot)

Repeat the values of nrows and ncols for all plots in a single gure window. Increment thisPlot for each plot

NMM: Interactive Computing with Matlab

page 97

Multiple plots per gure window

Example:
>> x = linspace(0,2*pi); >> subplot(2,2,1); >> plot(x,sin(x)); axis([0 2*pi -1.5 1.5]); >> subplot(2,2,2); >> plot(x,sin(2*x)); >> subplot(2,2,3); >> plot(x,sin(3*x)); >> subplot(2,2,4); >> plot(x,sin(4*x));

(1)

title(sin(x));

axis([0 2*pi -1.5 1.5]);

title(sin(2x));

axis([0 2*pi -1.5 1.5]);

title(sin(3x));

axis([0 2*pi -1.5 1.5]);

title(sin(4x));

(See next slide for the plot.)

NMM: Interactive Computing with Matlab

page 98

Multiple plots per gure window

sin(x) 1.5 1 0.5 0 -0.5 -1 -1.5 0 2 sin(3x) 1.5 1 0.5 0 -0.5 -1 -1.5 0 2 4 6 1.5 1 0.5 0 -0.5 -1 -1.5 0 2 4 4 6 1.5 1 0.5 0 -0.5 -1 -1.5 0 2 sin(4x) 4 sin(2x)

(2)

NMM: Interactive Computing with Matlab

page 99

Plot Annotation

Name axis grid gtext legend text xlabel ylabel title

Operation(s) performed Reset axis limits Draw grid lines at the major ticks marks on the x and y axes Add text to a location determined by a mouse click Create a legend to identify symbols and line types when multiple curves are drawn on the same plot Add text to a specied (x, y) location Label the x-axis Label the y -axis Add a title above the plot

NMM: Interactive Computing with Matlab

page 100

Plot Annotation Example

>> D = load(pdxTemp.dat); >> >> >> >> >> >> m = D(:,1);

(1)

T = D(:,2:4);

plot(m,t(:,1),ro,m,T(:,2),k+,m,T(:,3),b-); xlabel(Month); ylabel(Temperature ({}^\circ F)); title(Monthly average temperature at PDX); axis([1 12 20 100]); legend(High,Low,Average,2);

Note: The pdxTemp.dat le is in the data directory of the NMM toolbox. Make sure the toolbox is installed and is included in the Matlab path.
(See next slide for the plot.)
NMM: Interactive Computing with Matlab page 101

Plot Annotation Example

100 90 80 Temperature (degrees F) 70 60 50 40 30 20

(2)

Monthly average temperature for Portland International Airport High Low Average

6 Month

10

12

NMM: Interactive Computing with Matlab

page 102