Artifact a2d828538f1d1dad5f15946d91376d111c2ae40edec1ccc87e1bd8bc5899b785:
- Executable file
r37/packages/linalg/linalg.tex
— part of check-in
[f2fda60abd]
at
2011-09-02 18:13:33
on branch master
— Some historical releases purely for archival purposes
git-svn-id: https://svn.code.sf.net/p/reduce-algebra/code/trunk/historical@1375 2bfe0521-f11c-4a00-b80e-6202646ff360 (user: arthurcnorman@users.sourceforge.net, size: 58539) [annotate] [blame] [check-ins using] [more...]
- Executable file
r38/packages/linalg/linalg.tex
— part of check-in
[f2fda60abd]
at
2011-09-02 18:13:33
on branch master
— Some historical releases purely for archival purposes
git-svn-id: https://svn.code.sf.net/p/reduce-algebra/code/trunk/historical@1375 2bfe0521-f11c-4a00-b80e-6202646ff360 (user: arthurcnorman@users.sourceforge.net, size: 58539) [annotate] [blame] [check-ins using]
\documentstyle[11pt,reduce,fancyheadings]{article} \title{A Linear Algebra package for \REDUCE{}} \author{Matt Rebbeck \\ Konrad-Zuse-Zentrum f\"ur Informationstechnik Berlin} \date{July 1994} \def\foottitle{The Linear Algebra Package} \pagestyle{fancy} \lhead[]{{\footnotesize\leftmark}{}} \rhead[]{\thepage} \setlength{\headrulewidth}{0.6pt} \setlength{\footrulewidth}{0.6pt} \cfoot{} \rfoot{\small\foottitle} %decided against using these. %\def\ltri{$\triangleleft$} %\def\rtri{$\triangleright$} %\newcommand{\tribound}[1]{\rtri#1\ltri} \def\exprlist {expr$_{1}$,expr$_{2}$, \ldots ,expr$_{{\tt n}}$} \def\lineqlist {lin\_eqn$_{1}$,lin\_eqn$_{2}$, \ldots ,lin\_eqn$_{n}$} \def\matlist {mat$_{1}$,mat$_{2}$, \ldots ,mat$_{n}$} \def\veclist {vec$_{1}$,vec$_{2}$, \ldots ,vec$_{n}$} \def\lazyfootnote{\footnote{If you're feeling lazy then the \{\}'s can be omitted.}} \renewcommand{\thefootnote}{\fnsymbol{footnote}} \begin{document} \maketitle \index{Linear Algebra package} \section{Introduction} This package provides a selection of functions that are useful in the world of linear algebra. These functions are described alphabetically in section 3 of this document and are labelled 3.1 to 3.51. They can be classified into four sections(n.b: the numbers after the dots signify the function label in section 3). Contributions to this package have been made by Walter Tietze (ZIB). \subsection{Basic matrix handling} \begin{center} \begin{tabular}{l l l l l l} add\_columns & \ldots & 3.1 & add\_rows & \ldots & 3.2 \\ add\_to\_columns & \ldots & 3.3 & add\_to\_rows & \ldots & 3.4 \\ augment\_columns & \ldots & 3.5 & char\_poly & \ldots & 3.9 \\ column\_dim & \ldots & 3.12 & copy\_into & \ldots & 3.14 \\ diagonal & \ldots & 3.15 & extend & \ldots & 3.16 \\ find\_companion & \ldots & 3.17 & get\_columns & \ldots & 3.18 \\ get\_rows & \ldots & 3.19 & hermitian\_tp & \ldots & 3.21 \\ matrix\_augment & \ldots & 3.28 & matrix\_stack & \ldots & 3.30 \\ minor & \ldots & 3.31 & mult\_columns & \ldots & 3.32 \\ mult\_rows & \ldots & 3.33 & pivot & \ldots & 3.34 \\ remove\_columns & \ldots & 3.37 & remove\_rows & \ldots & 3.38 \\ row\_dim & \ldots & 3.39 & rows\_pivot & \ldots & 3.40 \\ stack\_rows & \ldots & 3.43 & sub\_matrix & \ldots & 3.44 \\ swap\_columns & \ldots & 3.46 & swap\_entries & \ldots & 3.47 \\ swap\_rows & \ldots & 3.48 & \end{tabular} \end{center} \subsection{Constructors} Functions that create matrices. \begin{center} \begin{tabular}{l l l l l l} band\_matrix & \ldots & 3. 6 & block\_matrix & \ldots & 3. 7 \\ char\_matrix & \ldots & 3. 8 & coeff\_matrix & \ldots & 3. 11 \\ companion & \ldots & 3. 13 & hessian & \ldots & 3. 22 \\ hilbert & \ldots & 3. 23 & jacobian & \ldots & 3. 24 \\ jordan\_block & \ldots & 3. 25 & make\_identity & \ldots & 3. 27 \\ random\_matrix & \ldots & 3. 36 & toeplitz & \ldots & 3. 50 \\ Vandermonde & \ldots & 3. 51 & Kronecker\_Product & \ldots & 3. 52 \end{tabular} \end{center} \subsection{High level algorithms} \begin{center} \begin{tabular}{l l l l l l} char\_poly & \ldots & 3.9 & cholesky & \ldots & 3.10 \\ gram\_schmidt & \ldots & 3.20 & lu\_decom & \ldots & 3.26 \\ pseudo\_inverse & \ldots & 3.35 & simplex & \ldots & 3.41 \\ svd & \ldots & 3.45 & triang\_adjoint & \ldots & 3.51 \end{tabular} \end{center} \vspace*{5mm} There is a separate {\small NORMFORM}[1] package for computing the following matrix normal forms in \REDUCE. \begin{center} smithex, smithex\_int, frobenius, ratjordan, jordansymbolic, jordan. \end{center} \subsection{Predicates} \begin{center} \begin{tabular}{l l l l l l} matrixp & \ldots & 3.29 & squarep & \ldots & 3.42 \\ symmetricp & \ldots & 3.49 & \end{tabular} \end{center} \subsection*{Note on examples:} In the examples the matrix ${\cal A}$ will be \begin{flushleft} \begin{math} {\cal A} = \left( \begin{array}{ccc} 1 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 9 \end{array} \right) \end{math} \end{flushleft} \subsection*{Notation} Throughout ${\cal I}$ is used to indicate the identity matrix and ${\cal A}^T$ to indicate the transpose of the matrix ${\cal A}$. \section{Getting started} If you have not used matrices within {\REDUCE} before then the following may be helpful. \subsection*{Creating matrices} Initialisation of matrices takes the following syntax: {\tt mat1 := mat((a,b,c),(d,e,f),(g,h,i));} will produce \begin{flushleft} \begin{math} mat1 := \left( \begin{array}{ccc} a & b & c \\ d & e & f \\ g & h & i \end{array} \right) \end{math} \end{flushleft} \subsection*{Getting at the entries} The (i,$\,$j)'th entry can be accessed by: {\tt mat1(i,j);} \subsection*{Loading the linear\_algebra package} The package is loaded by: {\tt load\_package linalg;} \section{What's available} \subsection{add\_columns, add\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt add\_columns(${\cal A}$,c1,c2,expr);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ & :- & a matrix. \\ c1,c2 & :- & positive integers. \\ expr & :- & a scalar expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} \parbox[t]{0.95\linewidth}{{\tt add\_columns} replaces column c2 of ${\cal A}$ by expr $*$ column(${\cal A}$,c1) $+$ column(${\cal A}$,c2).} {\tt add\_rows} performs the equivalent task on the rows of ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \begin{math} \hspace*{0.16in} \begin{array}{ccc} {\tt add\_columns}({\cal A},1,2,x) & = & \left( \begin{array}{ccc} 1 & x+2 & 3 \\ 4 & 4*x+5 & 6 \\ 7 & 7*x+8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt add\_rows}({\cal A},2,3,5) & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ 4 & 5 & 6 \\ 27 & 33 & 39 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt add\_to\_columns}, {\tt add\_to\_rows}, {\tt mult\_columns}, {\tt mult\_rows}. \subsection{add\_rows} \hspace*{0.175in} see: {\tt add\_columns}. \subsection{add\_to\_columns, add\_to\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt add\_to\_columns(${\cal A}$,column\_list,expr);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ column\_list &:-& a positive integer or a list of positive integers. \\ expr &:-& a scalar expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt add\_to\_columns} adds expr to each column specified in column\_list of ${\cal A}$. {\tt add\_to\_rows} performs the equivalent task on the rows of ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} \begin{array}{ccc} {\tt add\_to\_columns}({\cal A},\{1,2\},10) & = & \left( \begin{array}{ccc} 11 & 12 & 3 \\ 14 & 15 & 6 \\ 17 & 18 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.175in} \begin{math} \begin{array}{ccc} {\tt add\_to\_rows}({\cal A},2,-x) & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ -x+4 & -x+5 & -x+6 \\ 7 & 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt add\_columns}, {\tt add\_rows}, {\tt mult\_rows}, {\tt mult\_columns}. \subsection{add\_to\_rows} \hspace*{0.175in} see: {\tt add\_to\_columns}. \subsection{augment\_columns, stack\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt augment\_columns(${\cal A}$,column\_list);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ column\_list &:-& either a positive integer or a list of positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt augment\_columns} gets hold of the columns of ${\cal A}$ specified in column\_list and sticks them together. {\tt stack\_rows} performs the same task on rows of ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt augment\_columns}({\cal A},\{1,2\}) & = & \left( \begin{array}{cc} 1 & 2 \\ 4 & 5 \\ 7 & 8 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt stack\_rows}({\cal A},\{1,3\}) & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ 7 & 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt get\_columns}, {\tt get\_rows}, {\tt sub\_matrix}. \subsection{band\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt band\_matrix(expr\_list,square\_size);} \hspace*{0.1in} \begin{tabular}{l l l} expr\_list \hspace*{0.088in} &:-& \parbox[t]{.72\linewidth} {either a single scalar expression or a list of an odd number of scalar expressions.} \end{tabular} \vspace*{0.04in} \hspace*{0.1in} \begin{tabular}{l l l} square\_size &:-& a positive integer. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt band\_matrix} creates a square matrix of dimension square\_size. The diagonal consists of the middle expr of the expr\_list. The exprs to the left of this fill the required number of sub\_diagonals and the exprs to the right the super\_diagonals. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt band\_matrix}(\{x,y,z\},6) & = & \left( \begin{array}{cccccc} y & z & 0 & 0 & 0 & 0 \\ x & y & z & 0 & 0 & 0 \\ 0 & x & y & z & 0 & 0 \\ 0 & 0 & x & y & z & 0 \\ 0 & 0 & 0 & x & y & z \\ 0 & 0 & 0 & 0 & x & y \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt diagonal}. \subsection{block\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt block\_matrix(r,c,matrix\_list);} \hspace*{0.1in} \begin{tabular}{l l l} r,c &:-& positive integers. \\ matrix\_list &:-& a list of matrices. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt block\_matrix} creates a matrix that consists of r by c matrices filled from the matrix\_list row wise. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\cal B} = \left( \begin{array}{cc} 1 & 0 \\ 0 & 1 \end{array} \right), & {\cal C} = \left( \begin{array}{c} 5 \\ 5 \end{array} \right), & {\cal D} = \left( \begin{array}{cc} 22 & 33 \\ 44 & 55 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.175in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt block\_matrix}(2,3,\{{\cal B,C,D,D,C,B}\}) & = & \left( \begin{array}{ccccc} 1 & 0 & 5 & 22 & 33 \\ 0 & 1 & 5 & 44 & 55 \\ 22 & 33 & 5 & 1 & 0 \\ 44 & 55 & 5 & 0 & 1 \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{char\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt char\_matrix(${\cal A},\lambda$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a square matrix. \\ $\lambda$ &:-& a symbol or algebraic expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt char\_matrix} creates the characteristic matrix ${\cal C}$ of ${\cal A}$. This is ${\cal C} = \lambda * {\cal I} - {\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt char\_matrix}({\cal A},x) & = & \left( \begin{array}{ccc} x-1 & -2 & -3 \\ -4 & x-5 & -6 \\ -7 & -8 & x-9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt char\_poly}. \subsection{char\_poly} %{\bf How to use it:} \hspace*{0.175in} {\tt char\_poly(${\cal A},\lambda$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a square matrix. \\ $\lambda$ &:-& a symbol or algebraic expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt char\_poly} finds the characteristic polynomial of ${\cal A}$. This is the determinant of $\lambda * {\cal I} - {\cal A}$. \end{addtolength} {\bf Examples:} \hspace*{0.175in} {\tt char\_poly({\cal A},$x$) $= x^3-15*x^2-18*x$} {\bf Related functions:} \hspace*{0.175in} {\tt char\_matrix}. \subsection{cholesky} %{\bf How to use it:} \hspace*{0.175in} {\tt cholesky(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a positive definite matrix containing numeric entries. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt cholesky} computes the cholesky decomposition of ${\cal A}$. It returns \{${\cal L,U}$\} where ${\cal L}$ is a lower matrix, ${\cal U}$ is an upper matrix, \\ ${\cal A} = {\cal LU}$, and ${\cal U} = {\cal L}^T$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal F} = \left( \begin{array}{ccc} 1 & 1 & 0 \\ 1 & 3 & 1 \\ 0 & 1 & 1 \end{array} \right) \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} ${\tt cholesky}$({\cal F}) & = & \left\{ \left( \begin{array}{ccc} 1 & 0 & 0 \\ 1 & \sqrt{2} & 0 \\ 0 & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} \end{array} \right), \left( \begin{array}{ccc} 1 & 1 & 0 \\ 0 & \sqrt{2} & \frac{1}{\sqrt{2}} \\ 0 & 0 & \frac{1}{\sqrt{2}} \end{array} \right) \right\} \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt lu\_decom}. \subsection{coeff\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt coeff\_matrix(\{\lineqlist{}\});} \lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \lineqlist &:-& \parbox[t]{.435\linewidth}{linear equations. Can be of the form {\it equation $=$ number} or just {\it equation}.} \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt coeff\_matrix} creates the coefficient matrix ${\cal C}$ of the linear equations. It returns \{${\cal C,X,B}$\} such that ${\cal CX} = {\cal B}$. \end{addtolength} {\bf Examples:} \begin{math} \hspace*{0.175in} {\tt coeff\_matrix}(\{x+y+4*z=10,y+x-z=20,x+y+4\}) = \end{math} \vspace*{0.1in} \begin{flushleft} \hspace*{0.175in} \begin{math} \left\{ \left( \begin{array}{ccc} 4 & 1 & 1 \\ -1 & 1 & 1 \\ 0 & 1 & 1 \end{array} \right), \left( \begin{array}{c} z \\ y \\ x \end{array} \right), \left( \begin{array}{c} 10 \\ 20 \\ -4 \end{array} \right) \right\} \end{math} \end{flushleft} \subsection{column\_dim, row\_dim} %{\bf How to use it:} \hspace*{0.175in} {\tt column\_dim(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt column\_dim} finds the column dimension of ${\cal A}$. \hspace*{0.175in} {\tt row\_dim} finds the row dimension of ${\cal A}$. {\bf Examples:} \hspace*{0.175in} {\tt column\_dim}(${\cal A}$) = 3 \subsection{companion} %{\bf How to use it:} \hspace*{0.175in} {\tt companion(poly,x);} \hspace*{0.1in} \begin{tabular}{l l l} poly &:-& a monic univariate polynomial in x. \\ x &:-& the variable. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt companion} creates the companion matrix ${\cal C}$ of poly. This is the square matrix of dimension n, where n is the degree of poly w.r.t. x. The entries of ${\cal C}$ are: ${\cal C}$(i,n) = -coeffn(poly,x,i-1) for i = 1 \ldots n, ${\cal C}$(i,i-1) = 1 for i = 2 \ldots n and the rest are 0. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt companion}(x^4+17*x^3-9*x^2+11,x) & = & \left( \begin{array}{cccc} 0 & 0 & 0 & -11 \\ 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 9 \\ 0 & 0 & 1 & -17 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt find\_companion}. \subsection{copy\_into} %{\bf How to use it:} \hspace*{0.175in} {\tt copy\_into(${\cal A,B}$,r,c);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A,B}$ &:-& matrices. \\ r,c &:-& positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt copy\_into} copies matrix ${\cal A}$ into ${\cal B}$ with ${\cal A}$(1,1) at ${\cal B}$(r,c). {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal G} = \left( \begin{array}{cccc} 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{array} \right) \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt copy\_into}({\cal A,G},1,2) & = & \left( \begin{array}{cccc} 0 & 1 & 2 & 3 \\ 0 & 4 & 5 & 6 \\ 0 & 7 & 8 & 9 \\ 0 & 0 & 0 & 0 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \begin{addtolength}{\leftskip}{0.22in} {\tt augment\_columns}, {\tt extend}, {\tt matrix\_augment}, {\tt matrix\_stack}, {\tt stack\_rows}, {\tt sub\_matrix}. \end{addtolength} \subsection{diagonal} %{\bf How to use it:} \hspace*{0.175in} {\tt diagonal(\{\matlist{}\});}\lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \matlist &:-& \parbox[t]{.58\linewidth}{each can be either a scalar expr or a square matrix. } \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt diagonal} creates a matrix that contains the input on the diagonal. {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal H} = \left( \begin{array}{cc} 66 & 77 \\ 88 & 99 \end{array} \right) \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt diagonal}(\{{\cal A},x,{\cal H}\}) & = & \left( \begin{array}{cccccc} 1 & 2 & 3 & 0 & 0 & 0 \\ 4 & 5 & 6 & 0 & 0 & 0 \\ 7 & 8 & 9 & 0 & 0 & 0 \\ 0 & 0 & 0 & x & 0 & 0 \\ 0 & 0 & 0 & 0 & 66 & 77 \\ 0 & 0 & 0 & 0 & 88 & 99 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt jordan\_block}. \subsection{extend} %{\bf How to use it:} \hspace*{0.175in} {\tt extend(${\cal A}$,r,c,expr);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ r,c &:-& positive integers. \\ expr &:-& algebraic expression or symbol. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt extend} returns a copy of ${\cal A}$ that has been extended by r rows and c columns. The new entries are made equal to expr. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt extend}({\cal A},1,2,x) & = & \left( \begin{array}{ccccc} 1 & 2 & 3 & x & x \\ 4 & 5 & 6 & x & x \\ 7 & 8 & 9 & x & x \\ x & x & x & x & x \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \begin{addtolength}{\leftskip}{0.22in} \parbox[t]{0.95\linewidth}{{\tt copy\_into}, {\tt matrix\_augment}, {\tt matrix\_stack}, {\tt remove\_columns}, {\tt remove\_rows}.} \end{addtolength} \subsection{find\_companion} \hspace*{0.175in} {\tt find\_companion(${\cal A}$,x);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ x &:-& the variable. \end{tabular} {\bf Synopsis:} \begin{addtolength}{\leftskip}{0.22in} Given a companion matrix, {\tt find\_companion} finds the polynomial from which it was made. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal C} = \left( \begin{array}{cccc} 0 & 0 & 0 & -11 \\ 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 9 \\ 0 & 0 & 1 & -17 \end{array} \right) \end{math} \end{flushleft} \vspace*{3mm} \begin{flushleft} \hspace*{0.175in} \begin{math} {\tt find\_companion}({\cal C},x) = x^4+17*x^3-9*x^2+11 \end{math} \end{flushleft} \vspace*{3mm} {\bf Related functions:} \hspace*{0.175in} {\tt companion}. \subsection{get\_columns, get\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt get\_columns(${\cal A}$,column\_list);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ c &:-& either a positive integer or a list of positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt get\_columns} removes the columns of ${\cal A}$ specified in column\_list and returns them as a list of column matrices. \end{addtolength} \hspace*{0.175in} {\tt get\_rows} performs the same task on the rows of ${\cal A}$. {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt get\_columns}({\cal A},\{1,3\}) & = & \left\{ \left( \begin{array}{c} 1 \\ 4 \\ 7 \end{array} \right), \left( \begin{array}{c} 3 \\ 6 \\ 9 \end{array} \right) \right\} \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt get\_rows}({\cal A},2) & = & \left\{ \left( \begin{array}{ccc} 4 & 5 & 6 \end{array} \right) \right\} \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt augment\_columns}, {\tt stack\_rows}, {\tt sub\_matrix}. \subsection{get\_rows} \hspace*{0.175in} see: {\tt get\_columns}. \subsection{gram\_schmidt} %{\bf How to use it:} \hspace*{0.175in} {\tt gram\_schmidt(\{\veclist{}\});} \lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \veclist &:-& \parbox[t]{.62\linewidth}{linearly independent vectors. Each vector must be written as a list, eg:\{1,0,0\}. } \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt gram\_schmidt} performs the gram\_schmidt orthonormalisation on the input vectors. It returns a list of orthogonal normalised vectors. \end{addtolength} {\bf Examples:} \hspace*{0.175in} {\tt gram\_schmidt(\{\{1,0,0\},\{1,1,0\},\{1,1,1\}\})} = \{\{1,0,0\},\{0,1,0\},\{0,0,1\}\} \hspace*{0.175in} {\tt gram\_schmidt(\{\{1,2\},\{3,4\}\})} $= \{\{ \frac{1}{{\sqrt{5}}} , \frac{2}{\sqrt{5}} \}, \{ \frac{2*\sqrt{5}}{5} , \frac{-\sqrt{5}}{5} \}\}$ \subsection{hermitian\_tp} %{\bf How to use it:} \hspace*{0.175in} {\tt hermitian\_tp(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt hermitian\_tp} computes the hermitian transpose of ${\cal A}$. This is a matrix in which the (i,$\,$j)'th entry is the conjugate of the (j,$\,$i)'th entry of ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal J} = \left( \begin{array}{ccc} i+1 & i+2 & i+3 \\ 4 & 5 & 2 \\ 1 & i & 0 \end{array} \right) \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt hermitian\_tp}({\cal J}) & = & \left( \begin{array}{ccc} -i+1 & 4 & 1 \\ -i+2 & 5 & -i \\-i+3 & 2 & 0 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt tp}\footnote{standard reduce call for the transpose of a matrix - see {\REDUCE} User's Manual[2].}. \subsection{hessian} %{\bf How to use it:} \hspace*{0.175in} {\tt hessian(expr,variable\_list);} \hspace*{0.1in} \begin{tabular}{l l l} expr &:-& a scalar expression. \\ variable\_list &:-& either a single variable or a list of variables. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt hessian} computes the hessian matrix of expr w.r.t. the varibles in variable\_list. This is an n by n matrix where n is the number of variables and the (i,$\,$j)'th entry is df(expr,variable\_list(i),variable\_list(j)). \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt hessian}(x*y*z+x^2,\{w,x,y,z\}) & = & \left( \begin{array}{cccc} 0 & 0 & 0 & 0 \\ 0 & 2 & z & y \\ 0 & z & 0 & x \\ 0 & y & x & 0 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt df}\footnote{standard reduce call for differentiation - see {\REDUCE} User's Manual[2]}. \subsection{hilbert} %{\bf How to use it:} \hspace*{0.175in} {\tt hilbert(square\_size,expr);} \hspace*{0.1in} \begin{tabular}{l l l} square\_size &:-& a positive integer. \\ expr &:-& an algebraic expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt hilbert} computes the square hilbert matrix of dimension square\_size. This is the symmetric matrix in which the (i,$\,$j)'th entry is 1/(i+j-expr). \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt hilbert}(3,y+x) & = & \left( \begin{array}{ccc} \frac{-1}{x+y-2} & \frac{-1}{x+y-3} & \frac{-1}{x+y-4} \\ \frac{-1}{x+y-3} & \frac{-1}{x+y-4} & \frac{-1}{x+y-5} \\ \frac{-1}{x+y-4} & \frac{-1}{x+y-5} & \frac{-1}{x+y-6} \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{jacobian} %{\bf How to use it:} \hspace*{0.175in} {\tt jacobian(expr\_list,variable\_list);} \hspace*{0.1in} \begin{tabular}{l l l} expr\_list \hspace*{0.175in} &:-& \parbox[t]{.72\linewidth}{either a single algebraic expression or a list of algebraic expressions.} \end{tabular} \vspace*{0.04in} \hspace*{0.1in} \begin{tabular}{l l l} variable\_list &:-& either a single variable or a list of variables. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt jacobian} computes the jacobian matrix of expr\_list w.r.t. variable\_list. This is a matrix whose (i,$\,$j)'th entry is df(expr\_list(i), variable\_list(j)). The matrix is n by m where n is the number of variables and m the number of expressions. \end{addtolength} {\bf Examples:} \hspace*{0.175in} {\tt jacobian(\{$x^4,x*y^2,x*y*z^3$\},\{$w,x,y,z$\})} = \vspace*{0.1in} \begin{flushleft} \hspace*{0.175in} \begin{math} \left( \begin{array}{cccc} 0 & 4*x^3 & 0 & 0 \\ 0 & y^2 & 2*x*y & 0 \\ 0 & y*z^3 & x*z^3 & 3*x*y*z^2 \end{array} \right) \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt hessian}, {\tt df}\footnote{standard reduce call for differentiation - see {\REDUCE} User's Manual[2].}. \subsection{jordan\_block} %{\bf How to use it:} \hspace*{0.175in} {\tt jordan\_block(expr,square\_size);} \hspace*{0.1in} \begin{tabular}{l l l} expr &:-& an algebraic expression or symbol. \\ square\_size &:-& a positive integer. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt jordan\_block} computes the square jordan block matrix ${\cal J}$ of dimension square\_size. The entries of ${\cal J}$ are: ${\cal J}$(i,i) = expr for i=1 \ldots n, ${\cal J}$(i,i+1) = 1 for i=1 \ldots n-1, and all other entries are 0. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt jordan\_block(x,5)} & = & \left( \begin{array}{ccccc} x & 1 & 0 & 0 & 0 \\ 0 & x & 1 & 0 & 0 \\ 0 & 0 & x & 1 & 0 \\ 0 & 0 & 0 & x & 1 \\ 0 & 0 & 0 & 0 & x \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt diagonal}, {\tt companion}. \subsection{lu\_decom} %{\bf How to use it:} \hspace*{0.175in} {\tt lu\_decom(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& \parbox[t]{.848\linewidth}{a matrix containing either numeric entries or imaginary entries with numeric coefficients.} \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt lu\_decom} performs LU decomposition on ${\cal A}$, ie: it returns \{${\cal L,U}$\} where ${\cal L}$ is a lower diagonal matrix, ${\cal U}$ an upper diagonal matrix and ${\cal A} = {\cal LU}$. \end{addtolength} {\bf caution:} \begin{addtolength}{\leftskip}{0.22in} The algorithm used can swap the rows of ${\cal A}$ during the calculation. This means that ${\cal LU}$ does not equal ${\cal A}$ but a row equivalent of it. Due to this, {\tt lu\_decom} returns \{${\cal L,U}$,vec\}. The call {\tt convert(${\cal A}$,vec)} will return the matrix that has been decomposed, ie: ${\cal LU} = $ {\tt convert(${\cal A}$,vec)}. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal K} = \left( \begin{array}{ccc} 1 & 3 & 5 \\ -4 & 3 & 7 \\ 8 & 6 & 4 \end{array} \right) \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{cccc} ${\tt lu} := {\tt lu\_decom}$({\cal K}) & = & \left\{ \left( \begin{array}{ccc} 8 & 0 & 0 \\ -4 & 6 & 0 \\ 1 & 2.25 & 1.125 1 \end{array} \right), \left( \begin{array}{ccc} 1 & 0.75 & 0.5 \\ 0 & 1 & 1.5 \\ 0 & 0 & 1 \end{array} \right), [\; 3 \; 2 \; 3 \; ] \right\} \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} ${\tt first lu * second lu}$ & = & \left( \begin{array}{ccc} 8 & 6 & 4 \\ -4 & 3 & 7 \\ 1 & 3 & 5 \end{array} \right) \end{array} \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} ${\tt convert(${\cal K}$,third lu}$) \hspace*{0.055in} & = & \left( \begin{array}{ccc} 8 & 6 & 4 \\ -4 & 3 & 7 \\ 1 & 3 & 5 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.5in} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal P} = \left( \begin{array}{ccc} i+1 & i+2 & i+3 \\ 4 & 5 & 2 \\ 1 & i & 0 \end{array} \right) \end{math} \end{flushleft} \begin{eqnarray} \hspace*{0.22in} {\tt lu} := {\tt lu\_decom}({\cal P}) & = & \left\{ \left( \begin{array}{ccc} 1 & 0 & 0 \\ 4 & -4*i+5 & 0 \\ i+1 & 3 & 0.41463*i+2.26829 \end{array} \right), \right. \nonumber \\ & & \left. \: \; \, \left( \begin{array}{ccc} 1 & i & 0 \\ 0 & 1 & 0.19512*i+0.24390 \\ 0 & 0 & 1 \end{array} \right), \hspace*{0.05in} [\; 3 \; 2 \; 3 \;] \hspace*{0.05in} \right\} \nonumber \end{eqnarray} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} ${\tt first lu * second lu}$ & = & \left( \begin{array}{ccc} 1 & i & 0 \\ 4 & 5 & 2 \\ i+1 & i+2 & i+3 \end{array} \right) \end{array} \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} ${\tt convert({\cal P},third lu}$) \hspace*{0.1in} & = & \left( \begin{array}{c c c} 1 & i & 0 \\ 4 & 5 & 2 \\ i+1 & i+2 & i+3 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt cholesky}. \subsection{make\_identity} %{\bf How to use it:} \hspace*{0.175in} {\tt make\_identity(square\_size);} \hspace*{0.1in} \begin{tabular}{l l l} square\_size &:-& a positive integer. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt make\_identity} creates the identity matrix of dimension square\_size. {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt make\_identity}(4) & = & \left( \begin{array}{cccc} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt diagonal}. \subsection{matrix\_augment, matrix\_stack} %{\bf How to use it:} \hspace*{0.175in} {\tt matrix\_augment(\{\matlist\});}\lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \matlist &:-& matrices. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt matrix\_augment} sticks the matrices in matrix\_list together horizontally. \hspace*{0.175in} {\tt matrix\_stack} sticks the matrices in matrix\_list together vertically. {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt matrix\_augment}(\{{\cal A,A}\}) & = & \left( \begin{array}{cccccc} 1 & 2 & 3 & 1 & 2 & 3 \\ 4 & 4 & 6 & 4 & 5 & 6 \\ 7 & 8 & 9 & 7 & 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt matrix\_stack}(\{{\cal A,A}\}) & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 9 \\ 1 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt augment\_columns}, {\tt stack\_rows}, {\tt sub\_matrix}. \subsection{matrixp} %{\bf How to use it:} \hspace*{0.175in} {\tt matrixp(test\_input);} \hspace*{0.1in} \begin{tabular}{l l l} test\_input &:-& anything you like. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt matrixp} is a boolean function that returns t if the input is a matrix and nil otherwise. \end{addtolength} {\bf Examples:} \hspace*{0.175in} {\tt matrixp}(${\cal A}$) = t \hspace*{0.175in} {\tt matrixp}(doodlesackbanana) = nil {\bf Related functions:} \hspace*{0.175in} {\tt squarep}, {\tt symmetricp}. \subsection{matrix\_stack} \hspace*{0.175in} see: {\tt matrix\_augment}. \subsection{minor} %{\bf How to use it:} \hspace*{0.175in} {\tt minor(${\cal A}$,r,c);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ r,c &:-& positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt minor} computes the (r,c)'th minor of ${\cal A}$. This is created by removing the r'th row and the c'th column from ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt minor}({\cal A},1,3) & = & \left( \begin{array}{cc} 4 & 5 \\ 7 & 8 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt remove\_columns}, {\tt remove\_rows}. \subsection{mult\_columns, mult\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt mult\_columns(${\cal A}$,column\_list,expr);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ column\_list &:-& a positive integer or a list of positive integers. \\ expr &:-& an algebraic expression. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt mult\_columns} returns a copy of ${\cal A}$ in which the columns specified in column\_list have been multiplied by expr. {\tt mult\_rows} performs the same task on the rows of ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt mult\_columns}({\cal A},\{1,3\},x) & = & \left( \begin{array}{ccc} x & 2 & 3*x \\ 4*x & 5 & 6*x \\ 7*x & 8 & 9*x \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt mult\_rows}({\cal A},2,10) & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ 40 & 50 & 60 \\ 7 & 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt add\_to\_columns}, {\tt add\_to\_rows}. \subsection{\tt mult\_rows} \hspace*{0.175in} see: {\tt mult\_columns}. \subsection{pivot} %{\bf How to use it:} \hspace*{0.175in} {\tt pivot(${\cal A}$,r,c);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ r,c &:-& positive integers such that ${\cal A}$(r,c) neq 0. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt pivot} pivots ${\cal A}$ about its (r,c)'th entry. To do this, multiples of the r'th row are added to every other row in the matrix. This means that the c'th column will be 0 except for the (r,c)'th entry. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt pivot}({\cal A},2,3) & = & \left( \begin{array}{ccc} -1 & -0.5 & 0 \\ 4 & 5 & 6 \\ 1 & 0.5 & 0 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt rows\_pivot}. \subsection{pseudo\_inverse} %{\bf How to use it:} \hspace*{0.175in} {\tt pseudo\_inverse(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt pseudo\_inverse}, also known as the Moore-Penrose inverse, computes the pseudo inverse of ${\cal A}$. Given the singular value decomposition of ${\cal A}$, i.e: ${\cal A} = {\cal U} \sum {\cal V}^T$, then the pseudo inverse ${\cal A}^{-1}$ is defined by ${\cal A}^{-1} = {\cal V}^T \sum^{-1} {\cal U}$. Thus ${\cal A}$ $ * $ {\tt pseudo\_inverse}$({\cal A}) = {\cal I}$. \end{addtolength} {\bf Examples:} % \begin{flushleft} % \hspace*{0.175in} % \begin{math} % {\cal R} = \left( \begin{array}{cccc} 1 & 2 & 3 & 4 \\ 9 & 8 & 7 & 6 % \end{array} \right) % \end{math} % \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt pseudo\_inverse}({\cal A}) & = & \left( \begin{array}{cc} -0.2 & 0.1 \\ -0.05 & 0.05 \\ 0.1 & 0 \\ 0.25 & -0.05 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt svd}. \subsection{random\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt random\_matrix(r,c,limit);} \hspace*{0.1in} \begin{tabular}{l l l} r,c,$\,$limit &:-& positive integers. \\ \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt random\_matrix} creates an r by c matrix with random entries in the range $-$limit $<$ entry $<$ limit. \end{addtolength} {\bf switches:} \hspace*{0.1in} \begin{tabular}{l l l} {\tt imaginary} \hspace*{0.175in} &:-& \parbox[t]{0.685\linewidth}{if on then matrix entries are x+i$*$y where $-$limit $<$ x,y $<$ limit.} \end{tabular} \vspace*{0.04in} \hspace*{0.1in} \begin{tabular}{l l l} {\tt not\_negative} &:-& \parbox[t]{0.685\linewidth}{if on then 0 $<$ entry $<$ limit. In the imaginary case we have 0 $<$ x,y $<$ limit.} \end{tabular} \vspace*{0.04in} \hspace*{0.1in} \begin{tabular}{l l l} {\tt only\_integer} &:-& \parbox[t]{0.685\linewidth}{if on then each entry is an integer. In the imaginary case x and y are integers.} \end{tabular} \vspace*{0.04in} \hspace*{0.1in} \begin{tabular}{l l l} {\tt symmetric} &:-& if on then the matrix is symmetric. \\ {\tt upper\_matrix} &:-& \parbox[t]{0.685\linewidth}{if on then the matrix is upper triangular.} \\ {\tt lower\_matrix} &:-& if on then the matrix is lower triangular. \end{tabular} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt random\_matrix}(3,3,10) & = & \left( \begin{array}{ccc} -4.729721 & 6.987047 & 7.521383 \\ - 5.224177 & 5.797709 & - 4.321952 \\ - 9.418455 & - 9.94318 & - 0.730980 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.2in} \hspace*{0.165in} {\tt on only\_integer, not\_negative, upper\_matrix, imaginary;} \begin{flushleft} \hspace*{0.12in} \begin{math} \begin{array}{ccc} {\tt random\_matrix}(4,4,10) & = & \left( \begin{array}{cccc} 2*i+5 & 3*i+7 & 7*i+3 & 6 \\ 0 & 2*i+5 & 5*i+1 & 2*i+1 \\ 0 & 0 & 8 & i \\ 0 & 0 & 0& 5*i+9 \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{remove\_columns, remove\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt remove\_columns(${\cal A}$,column\_list);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ column\_list &:-& either a positive integer or a list of positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt remove\_columns} removes the columns specified in column\_list from ${\cal A}$. \hspace*{0.175in} {\tt remove\_rows} performs the same task on the rows of ${\cal A}$. {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt remove\_columns}({\cal A},2) & = & \left( \begin{array}{cc} 1 & 3 \\ 4 & 6 \\ 7 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt remove\_rows}({\cal A},\{1,3\}) & = & \left( \begin{array}{ccc} 4 & 5 & 6 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt minor}. \subsection{remove\_rows} \hspace*{0.175in} see: {\tt remove\_columns}. \subsection{row\_dim} \hspace{0.175in} see: {\tt column\_dim}. \subsection{rows\_pivot} %{\bf How to use it:} \hspace*{0.175in} {\tt rows\_pivot(${\cal A}$,r,c,\{row\_list\});} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ r,c &:-& positive integers such that ${\cal A}$(r,c) neq 0.\\ row\_list &:-& positive integer or a list of positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt rows\_pivot} performs the same task as {\tt pivot} but applies the pivot only to the rows specified in row\_list. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal N} = \left( \begin{array}{ccc} 1 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 9 \\1 & 2 & 3 \\ 4 & 5 & 6 \end{array} \right) \end{math} \end{flushleft} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt rows\_pivot}({\cal N},2,3,\{4,5\}) & = & \left( \begin{array} {c c c}1 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 9 \\ -0.75 & 0 & 0.75 \\ -0.375 & 0 & 0.375 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt pivot}. \subsection{simplex} %{\bf How to use it:} \hspace*{0.175in} {\tt simplex(max/min,objective function,\{linear inequalities\});} \hspace*{0.1in} \begin{tabular}{l l l} max/min & :- & \parbox[t]{.63\linewidth}{either max or min (signifying maximise and minimise).} \\ objective function & :- & the function you are maximising or minimising. \\ linear inequalities & :- & \parbox[t]{.63\linewidth}{the constraint inequalities. Each one must be of the form {\it sum of variables ($<=,=,>=$) number}.} \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt simplex} applies the revised simplex algorithm to find the optimal(either maximum or minimum) value of the objective function under the linear inequality constraints. It returns \{optimal value,\{ values of variables at this optimal\}\}. The algorithm implies that all the variables are non-negative. \end{addtolength} {\bf Examples:} \begin{addtolength}{\leftskip}{0.22in} %\begin{math} {\tt simplex($max,x+y,\{x>=10,y>=20,x+y<=25\}$);} %\end{math} {\tt ***** Error in simplex: Problem has no feasible solution.} \vspace*{0.2in} \parbox[t]{0.96\linewidth}{\tt simplex($max,10x+5y+5.5z,\{5x+3z<=200, x+0.1y+0.5z<=12$,\\ \hspace*{0.55in} $0.1x+0.2y+0.3z<=9, 30x+10y+50z<=1500\}$);} \vspace*{0.1in} {\tt $\{525.0,\{x=40.0,y=25.0,z=0\}$\}} \end{addtolength} \subsection{squarep} %{\bf How to use it:} \hspace*{0.175in} {\tt squarep(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt squarep} is a boolean function that returns t if the matrix is square and nil otherwise. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal L} = \left( \begin{array}{ccc} 1 & 3 & 5 \end{array} \right) \end{math} \end{flushleft} \vspace*{0.1in} \hspace*{0.175in} {\tt squarep}(${\cal A}$) = t \hspace*{0.175in} {\tt squarep}(${\cal L}$) = nil {\bf Related functions:} \hspace*{0.175in} {\tt matrixp}, {\tt symmetricp}. \subsection{stack\_rows} \hspace*{0.175in} see: {\tt augment\_columns}. \subsection{sub\_matrix} %{\bf How to use it:} \hspace*{0.175in} {\tt sub\_matrix(${\cal A}$,row\_list,column\_list);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ row\_list, column\_list &:-& \parbox[t]{.605\linewidth}{either a positive integer or a list of positive integers.} \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt sub\_matrix} produces the matrix consisting of the intersection of the rows specified in row\_list and the columns specified in column\_list. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt sub\_matrix}({\cal A},\{1,3\},\{2,3\}) & = & \left( \begin{array}{cc} 2 & 3 \\ 8 & 9 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt augment\_columns}, {\tt stack\_rows}. \subsection{svd (singular value decomposition)} %{\bf How to use it:} \hspace*{0.175in} {\tt svd(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix containing only numeric entries. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt svd} computes the singular value decomposition of ${\cal A}$. It returns \{${\cal U},\sum,{\cal V}$\} where ${\cal A} = {\cal U} \sum {\cal V}^T$ and $\sum = diag(\sigma_{1}, \ldots ,\sigma_{n}). \; \sigma_{i}$ for $i= (1 \ldots n)$ are the singular values of ${\cal A}$. n is the column dimension of ${\cal A}$. The singular values of ${\cal A}$ are the non-negative square roots of the eigenvalues of ${\cal A}^T {\cal A}$. ${\cal U}$ and ${\cal V}$ are such that ${\cal UU}^T = {\cal VV}^T = {\cal V}^T {\cal V} = {\cal I}_n$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal Q} = \left( \begin{array}{cc} 1 & 3 \\ -4 & 3 \end{array} \right) \end{math} \end{flushleft} \begin{eqnarray} \hspace*{0.1in} {\tt svd({\cal Q})} & = & \left\{ \left( \begin{array}{cc} 0.289784 & 0.957092 \\ -0.957092 & 0.289784 \end{array} \right), \left( \begin{array}{cc} 5.149162 & 0 \\ 0 & 2.913094 \end{array} \right), \right. \nonumber \\ & & \left. \: \; \, \left( \begin{array}{cc} -0.687215 & 0.726453 \\ -0.726453 & -0.687215 \end{array} \right) \right\} \nonumber \end{eqnarray} \subsection{swap\_columns, swap\_rows} %{\bf How to use it:} \hspace*{0.175in} {\tt swap\_columns(${\cal A}$,c1,c2);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ c1,c1 &:-& positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt swap\_columns} swaps column c1 of ${\cal A}$ with column c2. \hspace*{0.175in} {\tt swap\_rows} performs the same task on 2 rows of ${\cal A}$. {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt swap\_columns}({\cal A},2,3) & = & \left( \begin{array}{ccc} 1 & 3 & 2 \\ 4 & 6 & 5 \\ 7 & 9 & 8 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt swap\_entries}. \subsection{swap\_entries} %{\bf How to use it:} \hspace*{0.175in} {\tt swap\_entries(${\cal A}$,\{r1,c1\},\{r2,c2\});} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \\ r1,c1,r2,c2 &:-& positive integers. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \hspace*{0.175in} {\tt swap\_entries} swaps ${\cal A}$(r1,c1) with ${\cal A}$(r2,c2). {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt swap\_entries}({\cal A},\{1,1\},\{3,3\}) & = & \left( \begin{array}{ccc} 9 & 2 & 3 \\ 4 & 5 & 6 \\ 7 & 8 & 1 \end{array} \right) \end{array} \end{math} \end{flushleft} {\bf Related functions:} \hspace*{0.175in} {\tt swap\_columns}, {\tt swap\_rows}. \subsection{swap\_rows} \hspace*{0.175in} see: {\tt swap\_columns}. \subsection{symmetricp} %{\bf How to use it:} \hspace*{0.175in} {\tt symmetricp(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt symmetricp} is a boolean function that returns t if the matrix is symmetric and nil otherwise. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.175in} \begin{math} {\cal M} = \left( \begin{array}{cc} 1 & 2 \\ 2 & 1 \end{array} \right) \end{math} \end{flushleft} \vspace*{0.1in} \hspace*{0.175in} {\tt symmetricp}(${\cal A}$) = nil \hspace*{0.175in} {\tt symmetricp}(${\cal M}$) = t {\bf Related functions:} \hspace*{0.175in} {\tt matrixp}, {\tt squarep}. \subsection{toeplitz} %{\bf How to use it:} \hspace*{0.175in} {\tt toeplitz(\{\exprlist{}\});} \lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \exprlist{} &:-& algebraic expressions. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt toeplitz} creates the toeplitz matrix from the expression list. This is a square symmetric matrix in which the first expression is placed on the diagonal and the i'th expression is placed on the (i-1)'th sub and super diagonals. It has dimension n where n is the number of expressions. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt toeplitz}(\{w,x,y,z\}) & = & \left( \begin{array}{cccc} w & x & y & z \\ x & w & x & y \\ y & x & w & x \\ z & y & x & w \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{triang\_adjoint} %{\bf How to use it:} \hspace*{0.175in} {\tt triang\_adjoint(${\cal A}$);} \hspace*{0.1in} \begin{tabular}{l l l} ${\cal A}$ &:-& a matrix. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt triang\_adjoint} computes the triangularizing adjoint ${\cal F}$ of matrix ${\cal A}$ due to the algorithm of Arne Storjohann. ${\cal F}$ is lower triangular matrix and the resulting matrix ${\cal T}$ of ${\cal F * A = T}$ is upper triangular with the property that the $i$-th entry in the diagonal of ${\cal T}$ is the determinant of the principal $i$-th submatrix of the matrix ${\cal A}$. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt triang\_adjoint}({\cal A}) & = & \left( \begin{array}{ccc} 1 & 0 & 0 \\ -4 & 1 & 0 \\ -3 & 6 & -3 \end{array} \right) \end{array} \end{math} \end{flushleft} \vspace*{0.1in} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\cal F} * {\cal A} & = & \left( \begin{array}{ccc} 1 & 2 & 3 \\ 0 & -3 & -6 \\ 0 & 0 & 0 \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{Vandermonde} %{\bf How to use it:} \hspace*{0.175in} {\tt vandermonde}(\{\exprlist{}\}); \addtocounter {footnote}{-1}\footnotemark %\lazyfootnote{} \hspace*{0.1in} \begin{tabular}{l l l} \exprlist{} &:-& algebraic expressions. \end{tabular} {\bf Synopsis:} %{\bf What it does:} \begin{addtolength}{\leftskip}{0.22in} {\tt Vandermonde} creates the Vandermonde matrix from the expression list. This is the square matrix in which the (i,$\,$j)'th entry is expr\_list(i) $^{(j-1)}$. It has dimension n where n is the number of expressions. \end{addtolength} {\bf Examples:} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} {\tt vandermonde}(\{x,2*y,3*z\}) & = & \left( \begin{array}{ccc} 1 & x & x^2 \\ 1 & 2*y & 4*y^2 \\ 1 & 3*z & 9*z^2 \end{array} \right) \end{array} \end{math} \end{flushleft} \subsection{kronecker\_product} \hspace*{0.175in} {\tt kronecker\_product}($Mat_1,Mat_2$) \hspace*{0.1in} \begin{tabular}{l l l} $Mat_1,Mat_2$ &:-& Matrices \end{tabular} {\bf Synopsis:} \begin{addtolength}{\leftskip}{0.22in} {\tt kronecker\_product} creates a matrix containing the Kronecker product (also called {\tt direct product} or {\tt tensor product}) of its arguments. \end{addtolength} {\bf Examples:} \begin{verbatim} a1 := mat((1,2),(3,4),(5,6))$ a2 := mat((1,1,1),(2,z,2),(3,3,3))$ kronecker_product(a1,a2); \end{verbatim} \begin{flushleft} \hspace*{0.1in} \begin{math} \begin{array}{ccc} \left( \begin{array}{cccccc} 1 & 1 & 1 & 2 & 2 & 2 \\ 2 & z & 2 & 4 &2*z &4 \\ 3 & 3 & 3 & 6 & 6 &6 \\ 3 & 3 & 3 & 4 & 4 &4 \\ 6 & 3*z& 6 & 8 &4*z &8 \\ 9 & 9 & 9 & 12 &12 &12\\ 5 & 5 & 5 & 6 & 6 &6 \\ 10 &5*z& 10& 12 &6*z &12 \\ 15 &15 & 15& 18 &18 &18 \end{array} \right) \end{array} \end{math} \end{flushleft} \section{Fast Linear Algebra} By turning the {\tt fast\_la} switch on, the speed of the following functions will be increased: \begin{tabular}{l l l l} add\_columns & add\_rows & augment\_columns & column\_dim \\ copy\_into & make\_identity & matrix\_augment & matrix\_stack\\ minor & mult\_column & mult\_row & pivot \\ remove\_columns & remove\_rows & rows\_pivot & squarep \\ stack\_rows & sub\_matrix & swap\_columns & swap\_entries\\ swap\_rows & symmetricp \end{tabular} The increase in speed will be insignificant unless you are making a significant number(i.e: thousands) of calls. When using this switch, error checking is minimised. This means that illegal input may give strange error messages. Beware. \newpage \section{Acknowledgments} Many of the ideas for this package came from the Maple[3] Linalg package [4]. The algorithms for {\tt cholesky}, {\tt lu\_decom}, and {\tt svd} are taken from the book Linear Algebra - J.H. Wilkinson \& C. Reinsch[5]. The {\tt gram\_schmidt} code comes from Karin Gatermann's Symmetry package[6] for {\REDUCE}. \begin{thebibliography}{} \bibitem{matt} Matt Rebbeck: NORMFORM: A {\REDUCE} package for the computation of various matrix normal forms. ZIB, Berlin. (1993) \bibitem{Reduce} Anthony C. Hearn: {\REDUCE} User's Manual 3.6. RAND (1995) \bibitem{Maple} Bruce W. Char\ldots [et al.]: Maple (Computer Program). Springer-Verlag (1991) \bibitem{linalg} Linalg - a linear algebra package for Maple[3]. \bibitem{WiRe} J. H. Wilkinson \& C. Reinsch: Linear Algebra (volume II). Springer-Verlag (1971) \bibitem{gat} Karin Gatermann: Symmetry: A {\REDUCE} package for the computation of linear representations of groups. ZIB, Berlin. (1992) \end{thebibliography} \end{document}