\startcomponent context-matrices

%\showframe

\setupheadertexts[Matlab and Matrices]

%\starttext

 \start

 \setupalign[middle]

  \tfc Matlab and Matrices

  \tfa\setupinterlinespace

  \vskip 2ex

  Math 45---Linear Algebra

  David Arnold

  \vskip 2ex

 \stop

\midaligned{\currentdate}

  \vskip 2ex

\midaligned{\bf Abstract}

\startnarrower[3*middle]

\noindenting In this exercise you will learn how to enter and edit
matrices in Matlab. You will also experiment with some of
Matlab's built-in matrix builders. You will also learn how to
build matrices from vectors and blocks of matrices. {\em
Prequisites: None}

\stopnarrower

\section{Entering Matrices}

Entering matrices in Matlab is easy. Enter the following at the Matlab
prompt.
 \starttyping
 >> A=[1,2,3;4,5,6;7,8,9]
 \stoptyping
Delimit the entries in a row with commas and delimit the rows with
semicolons. Spaces can also be used to delimit the entries in a
row.
 \starttyping
 >> A=[1 2 3;4 5 6;7 8 9]
 \stoptyping

\subsection{Special Matrices in Matlab}

Matlab has a number of built-in routines to help create matrices
\footnote{For a list of these elementary matrices, type
\type{help elmat} and \type{help specmat} at the Matlab prompt.}.
You can create a matrix of zeros of any size \footnote{Remember,
help is readily available for any Matlab function. For example,
type \type{help zeros} to receive help on the \type{zeros}
function.}.
 \starttyping
 >> A=zeros(5)
 >> B=zeros(3,5)
 \stoptyping
You easily can create a matrix of zeros that is the same size as
a given matrix.
 \starttyping
 >> C=magic(5)
 >> D=zeros(size(C))
 \stoptyping
You can create matrices of ones in a similar manner.
 \starttyping
 >> A=ones(6)
 >> B=ones(2,10)
 >> C=hilb(5)
 >> D=ones(size(C))
 \stoptyping

When performing simulations in Matlab it is useful to build matrices of
random numbers. You can create a matrix of uniformly distributed random
numbers, each between 0 and 1 (0 inclusive), with the following commands.
 \starttyping
 >> A=rand(6)
 >> B=rand(5,3)
 \stoptyping

Scalar multiplication works with matrices in exactly the same manner as with
vectors.
 \starttyping
 >> C=10*rand(5)
 >> D=floor(10*rand(5))
 \stoptyping

The identity matrix has ones on its main diagonal and zeros
everywhere else.\footnote{The import of the identity matrix will
be investigated in later activities.}
 \starttyping
 >> I=eye(5)
 \stoptyping

Other types of diagonal matrices are possible with Matlab's
\type{diag} command.
 \starttyping
 >> E=diag([1,2,3,4,5])
 >> F=diag([1,2,3,4,5],-1)
 >> G=diag(1:5,1)
 \stoptyping

\subsection{The Transpose Operator}

The transpose operator (single apostrophe) plays the same role with matrices
as it does with vectors. Rows are changed to columns and columns are changed
to rows.
 \starttyping
 >> J=[1 2 3;4 5 6;7 8 9]
 >> J'
 \stoptyping

\subsection{Suppressing Output}

Remember, appending a semicolon to any Matlab command will suppress the
output. This is useful if the result is large and you have no desire to view
it.
 \starttyping
 >> K=rand(100);
 \stoptyping

\subsection{Matlab's Workspace}

Take a moment to examine Matlab's workspace with the \type{whos}
command.
 \starttyping
 >> whos
 \stoptyping
Note the size of each of the variables in your workspace is
given. Of course, you could also find the size of the matrix
\type{I} by entering
 \starttyping
 >> size(I)
 \stoptyping

\section{Indexing Matrices}

The following notation is used to denote a matrix with 3 rows and 3 columns.
 \placeformula[-]
 \startformula
 A=\pmatrix{%
 a_{11} & a_{12} & a_{13} \cr
 a_{21} & a_{22} & a_{23} \cr
 a_{31} & a_{32} & a_{33}}
 \stopformula
Note that $a_{12}$ refers to the entry \index{matrix
entries@matrix entries} in row 1, column 2. In general, $a_{ij}$
refers to the entry in row $i$, column $j$.

Matlab uses similar notation to identify the elements of a matrix.
 \starttyping
 >> A=pascal(5)
 >> A(1,2)
 >> A(3,4)
 \stoptyping
In general, \type{A(i,j)} refers to the element in row $i$, column
$j$ of matrix $A$.

You can easily change an entry in a matrix.
 \starttyping
 >> A(3,3)=11111
 \stoptyping

\subsection{Advanced Indexing Techniques}

When indexing a matrix, your subscripts can also be vectors. This is a
powerful tool which allows you to easily select a submatrix of a given
matrix.
 \starttyping
 >> A=magic(6)
 >> A([1,2],[3,4,5])
 \stoptyping
The notation \type{A([1,2],[3,4,5])} references the submatrix
formed using the elements that appear in rows 1 and 2 {\em and}
in columns 3,4, and 5 of the matrix \type{A}.

The command
 \starttyping
 >> A([1,3,5],[1,2,3,4,5,6])
 \stoptyping
produces a submatrix composed of rows 1, 3, and 5 of matrix
\type{A}.

If you recall that the notation \type{1:6} represents the vector
\type{[1,2,3,4,5,6]}, so that \type{A(1,1:6)} is equivalent to
\type{A(1,[1,2,3,4,5,6])}.
 \starttyping
 >> A(1,1:6)
 \stoptyping
Using a colon by itself in place of a subscript denotes {\em all}
of the corresponding row or column. Thus,
 \starttyping
 >> A(:,1)
 \stoptyping
produces the first column of the matrix \type{A}, and
 \starttyping
 >> A(3,:)
 \stoptyping
produces the third row of the matrix \type{A}. In a sense, you
can read the notation \type{A(3,:)} as follows: ``Third row,
every column.'' The command
 \starttyping
 >> A(1:3,:)
 \stoptyping
produces a submatrix composed of the first three rows of the
matrix \type{A}, while the command
 \starttyping
 >> A(:,1:2:6)
 \stoptyping
produces a submatrix composed of columns 1, 3, and 5 of matrix
\type{A}.

\section{Building Matrices}

Matlab allows you to build complex matrices out of vectors and matrices.

\subsection{Building Matrices From Vectors}

Create three row vectors with the following commands.
 \starttyping
 >> v1=1:3
 >> v2=4:6
 >> v3=7:9
 \stoptyping
The command
 \starttyping
 >> M=[v1;v2;v3]
 \stoptyping
builds a matrix with the vectors \type{v1}, \type{v2}, and
\type{v3} each forming a row of the matrix \type{M}.

The command
 \starttyping
 >> N=[v1,v2,v3]
 \stoptyping
produces a significantly different but logical result.

Change the vectors \type{v1}, \type{v2}, and \type{v3} into
column vectors with the transpose operator.
 \starttyping
 >> v1=v1'
 >> v2=v2'
 >> v3=v3'
 \stoptyping
The command
 \starttyping
 >> P=[v1,v2,v3]
 \stoptyping
builds a matrix with the vectors \type{v1,v2}, and \type{v3} each
forming a column of the matrix \type{P}. Of course, you could
have achieved the same result by taking the transpose of the
matrix \type{M}.
 \starttyping
 >> M'
 \stoptyping

\subsection{The Dimensions Need To Match}

When building matrices, be sure that each row and column has the same number
of elements. For example, the following sequence of commands will produce an
error.
 \starttyping
 >> w1=1:3,w2=4:6,w3=7:10
 >> Q=[w1;w2;w3]
 \stoptyping

\subsection{Building Matrices From Other Matrices}

It is a simple matter to augment a row or column vector to a matrix.
Remember, each row and column of a matrix must have the same number of
elements. Thus,
 \starttyping
 >> A=[1,2,3,4;5,6,7,8;9,10,11,12]
 >> b=[1,1,1]'
 >> M=[A,b]
 \stoptyping
is valid, but
 \starttyping
 >> M=[A;b]
 \stoptyping
is not.

You can also augment two or more matrices. Remember, each row and column of
a matrix must have the same number of elements. Thus,
 \starttyping
 >> A=magic(3),B=ones(3,4)}
 >> M=[A,B]
 \stoptyping
is valid, but
 \starttyping
 >> N=[A;B]
 \stoptyping
is not.

\subsection{Your Imagination Is Your Only Limit}

Matlab's matrix building capability is extremely flexible. Here is an
interesting example.
 \starttyping
 >> A=zeros(3),B=ones(3),C=2*ones(3),D=3*ones(3)
 >> M=[A,B;C,D]
 \stoptyping

Here is something called a {\em Vandermonde} matrix.
 \starttyping
 >> x=[1,2,3,4,5]'
 >> N=[ones(size(x)),x,x.^2,x.^3,x.^4]
 \stoptyping
And here is one final piece of chicanery.
 \starttyping
 >> B=zeros(8)
 >> B(1:3,1:3)=[1 2 3;4 5 6;7 8 9]
 >> B(4:8,4:8)=magic(5)
 \stoptyping

%\stoptext

\stopcomponent