glpk10.tex

著名的大规模线性规划求解器源码GLPK.C语言版本,可以修剪.内有详细帮助文档.

%* glpk10.tex *%

\chapter{Stand-alone LP/MIP Solver}
\label{chaglpsol}

The GLPK package includes the program \verb|glpsol| which is a
stand-alone LP/MIP solver. This program can be invoked from the command
line of from the shell to read LP/MIP problem data in any format
supported by GLPK, solve the problem, and write the obtained problem
solution to a text file in plain format.

\subsubsection*{Usage}

\noindent
\verb|glpsol| [{\it options\dots}] [{\it filename}]

\subsubsection*{General options}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--mps|      &  read LP/MIP problem in fixed MPS format \\
\verb|--freemps|  &  read LP/MIP problem in free MPS format (default)\\
\verb|--cpxlp|    &  read LP/MIP problem in CPLEX LP format \\
\verb|--math|     &  read LP/MIP model written in GNU MathProg modeling
                     language \\
\multicolumn{2}{@{}l}{{\tt -m} {\it filename}, {\tt --model}
{\it filename}} \\
                  &  read model section and optional data section from
                     filename (the same as \verb|--math|) \\
\multicolumn{2}{@{}l}{{\tt -d} {\it filename}, {\tt --data}
{\it filename}} \\
                  &  read data section from filename (for \verb|--math|
                     only); if model file also has data section, that
                     section is ignored \\
\multicolumn{2}{@{}l}{{\tt -y} {\it filename}, {\tt --display}
{\it filename}} \\
                  &  send display output to filename (for \verb|--math|
                     only); by default the output is sent to
                     \verb|stdout| \\
\end{tabular}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--min|      &  minimization \\
\verb|--max|      &  maximization \\
\verb|--scale|    &  scale problem (default) \\
\verb|--noscale|  &  do not scale problem \\
\verb|--simplex|  &  use simplex method (default) \\
\verb|--interior| &  use interior point method (for pure LP only) \\
\multicolumn{2}{@{}l}{{\tt -o} {\it filename}, {\tt --output}
{\it filename}} \\
                  &  write solution to filename in plain text format \\
\multicolumn{2}{@{}l}{{\tt --bounds} {\it filename}} \\
                  &  write sensitivity bounds to filename in plain text
                     format (LP only) \\
\verb|--tmlim| {\it nnn}
                  &  limit solution time to {\it nnn} seconds
                     (\verb|--tmlim 0| allows obtaining solution at
                     initial point) \\
\verb|--memlim| {\it nnn}
                  &  limit available memory to {\it nnn} Megabytes \\
\verb|--check|    &  do not solve problem, check input data only \\
\verb|--name| {\it probname}
                  &  change problem name to {\it probname} \\
\verb|--plain|    &  use plain names of rows and columns (default) \\
\verb|--orig|     &  try using original names of rows and columns \\
\verb|--wmps| {\it filename}
                  &  write problem to {\it filename} in fixed MPS
                     format \\
\multicolumn{2}{@{}l}{{\tt --wfreemps} {\it filename}} \\
                  &  write problem to {\it filename} in free MPS
                     format \\
\multicolumn{2}{@{}l}{{\tt --wcpxlp} {\it filename}} \\
                  &  write problem to {\it filename} in CPLEX LP
                     format \\
\verb|--wtxt| {\it filename}
                  &  write problem to {\it filename} in plain text
                     format \\
\verb|-h|, \verb|--help|
                  &  display this help information and exit \\
\verb|-v|, \verb|--version|
                  &  display program version and exit \\
\end{tabular}

\subsection*{LP basis factorization option}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--luf|      &  LU + Forrest--Tomlin update \\
                  &  (faster, less stable; default) \\
\verb|--cbg|      &  LU + Schur complement + Bartels--Golub update \\
                  &  (slower, more stable) \\
\verb|--cbg|      &  LU + Schur complement + Givens rotation update \\
                  &  (slower, more stable) \\
\end{tabular}

\subsubsection*{Options specific to simplex method}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--std|      &  use standard initial basis of all slacks \\
\verb|--adv|      &  use advanced initial basis (default) \\
\verb|--bib|      &  use Bixby's initial basis\\
\end{tabular}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--bas| {\it filename}
                  &  read initial basis from {\it filename} in MPS
                     format \\
\verb|--steep|    &  use steepest edge technique (default) \\
\verb|--nosteep|  &  use standard ``textbook'' pricing \\
\verb|--relax|    &  use Harris' two-pass ratio test (default) \\
\verb|--norelax|  &  use standard ``textbook'' ratio test \\
\verb|--presol|   &  use LP presolver (default; assumes \verb|--scale|
                     and \verb|--adv|) \\
\verb|--nopresol| &  do not use LP presolver \\
\verb|--exact|    & use simplex method based on exact arithmetic \\
\end{tabular}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--xcheck|   & check final basis using exact arithmetic \\
\verb|--wbas| {\it filename}
                  & write final basis to {\it filename} in MPS format \\
\end{tabular}

\subsubsection*{Options specific to MIP}

\noindent
\begin{tabular}{@{}p{30mm}p{92.3mm}@{}}
\verb|--nomip|    &  consider all integer variables as continuous
                     (allows solving MIP as pure LP) \\
\verb|--first|    &  branch on first integer variable \\
\verb|--last|     &  branch on last integer variable \\
\verb|--drtom|    &  branch using heuristic by Driebeck and Tomlin
                     (default) \\
\verb|--mostf|    &  branch on most fractional varaible \\
\verb|--dfs|      &  backtrack using depth first search \\
\verb|--bfs|      &  backtrack using breadth first search \\
\verb|--bestp|    &  backtrack using the best projection heuristic
                     (default) \\
\verb|--bestb|    &  backtrack using node with best local bound \\
\verb|--intopt|   &  use advanced MIP solver (enables MIP presolving)\\
\verb|--binarize| &  replace general integer variables by binary ones
                     (assumes \verb|--intopt|)\\
\verb|--cover|    &  generate mixed cover cuts\\
\verb|--clique|   &  generate clique cuts\\
\verb|--gomory|   &  generate Gomory's mixed integer cuts\\
\verb|--mir|      &  generate MIR (mixed integer rounding) cuts\\
\verb|--cuts|     &  generate cuts of all classes above (assumes
                     \verb|--intopt|)\\
\end{tabular}

\bigskip

\noindent
For description of the MPS format see Appendix \ref{champs},
page \pageref{champs}.

\bigskip

\noindent
For description of the CPLEX LP format see Appendix \ref{chacplex},
page \pageref{chacplex}.

\bigskip

\noindent
For description of the modeling language see the document ``Modeling
Language GNU MathProg: Language Reference'' included in the GLPK
distribution.

%* eof *%