estiveljr / mp

An open-source library for mathematical programming

AMPL/MP

An open-source library for mathematical programming.

Features

• Expression-based solver interface. For solvers with an expression API, expression trees can be efficiently mapped. For example, AMPL expression max(a, b) directly maps to IBM ILOG Concert's IloMax. The library has the following C++ interfaces of this kind, all of which support AMPL extensions for logic and constraint programming:
  • Ilogcp: IBM ILOG CPLEX and CPLEX CP Optimizer
  • Gecode
  • JaCoP
  • LocalSolver
• Conversion-based solver interface (WIP). For solvers with more traditional 'flat' APIs, a customizable conversion layer translates expressions into constraints. For example, max(a, b) is translated into <new var> = max(a, b), which is in turn redefined into a MIP construct or passed to the solver (Gurobi: GRBgenconstrMax.) Logical and CP constraints are supported as well. For the solver API, an easy-to-adapt C++ wrapper class is provided. Currently it's two experimental interfaces:
  • Gurobi ('gurobidirect')
  • IBM ILOG CPLEX ('cplexdirect')
• End-to-end solver testing script for testing of various solver features. Documentation.
• An efficient type-safe C++ adapter for the previous ASL library for connecting solvers to AMPL and other systems: source
• Reusable high-performance .nl file reader which is up to 6x faster than the one provided by ASL. Documentation: http://ampl.github.io
• Database support on Linux and MacOS X. See Database and spreadsheet connection guide.
• SMPSWriter, a converter from deterministic equivalent of a two-stage stochastic programming (SP) problem written in AMPL to an SP problem in SMPS format.

Examples

Reading an .nl file:

#include "mp/nl.h"
#include "mp/problem.h"

mp::Problem p;
ReadNLFile("diet.nl", p);

Usage

Binaries for the open-source AMPL solvers for major platforms can be downloaded from the AMPL's Open Source Solvers page. To use a solver with AMPL, extract the binaries from a downloaded archive into the AMPL installation directory.

Building from source

An included CMake build script can be used to build the MP library, solver interfaces and function libraries on a wide range of platforms. You can download CMake for free from http://www.cmake.org/download/.

CMake works by generating native makefiles or project files that can be used in the compiler environment of your choice. The typical workflow starts with:

git clone https://github.com/ampl/mp.git
cd mp
git submodule init
git submodule update
mkdir build  # Create a directory to hold the build output.
cd build
cmake -DBUILD=all ..  # Generate native build scripts.

Note: If the arith.h file used by default does not match the target architecture, or if the compiler is not sufficiently compatible with gcc or Microsoft C/C++, run cmake -DBUILD=all -DGENERATE_ARITH=true .. to generate an architecture-specific arith.h file with arithchk.

If you are on a *nix system, you should now see a Makefile in the current directory. Now you can build MP by running make.

Once MP has been built you can invoke make test to run the tests.

If you use Windows and have Vistual Studio installed, an MP.sln file and several .vcproj files will be created. You can then build them using Visual Studio or msbuild.

On Mac OS X with Xcode installed, an .xcodeproj file will be generated.

Modules

AMPL/MP allows building only parts of the project you are interested in, for example you can choose to build only a single solver interface. This is done with the help of modules which are optional components that can be built separately. Each solver interface and function library is a module.

By default all modules are disabled and only the main MP libraries are built. To enable modules, pass their names as a comma-separated list in the BUILD variable when running CMake:

cmake -DBUILD=ilogcp,gecode .

Use -DBUILD=all to build all modules.

If a module is enabled, its dependencies are automatically downloaded and built when necessary. For example, enabling the gecode module will download the source code of Gecode constraint programming solver, build the solver and its AMPL interface.

Dependencies of some modules cannot be handled automatically due to licensing restrictions. If you enable such module, you should have its dependencies installed on the systems or it will not be built. For example, if you enable the ilogcp module, you should have IBM ILOG CPLEX Optimization Studio installed.

Using Eclipse CDT

You can generate Eclipse CDT project files with CMake:

cmake -G "Eclipse CDT 4 -  Unix Makefiles"

To get rid of semantic errors reported by Indexer add preprocessor symbols _GLIBCPP_USE_NAMESPACES, __GXX_EXPERIMENTAL_CXX0X__ and STAND_ALONE in "Project Properties" / "C/C++ Include Files and Symbols" and rebuild the index.

Building the documentation

To build the documentation you need the following software installed on your system:

• Python with pip and virtualenv
• Doxygen

First generate makefiles or project files using CMake as described in the previous section. Then compile the doc target/project, for example:

make doc

This will generate the HTML documenation in doc/ampl.github.io.

Links

AMPL home | AMPL book | Discussion group | SolverStudio for Excel

AMPL models by Håkan Kjellerstrand