Quick Start

Run Local-MIP

Install the Python bindings in a local virtual environment, or build the CLI solver from source when you need a local binary, C++ APIs, examples, or callbacks.

PyPI Package v2.0.10 Release Solver README Python README

Fastest Path

Install with pip

For CPython 3.8-3.12 on Linux x86_64, the published 2.0.10 wheels are the quickest way to try Local-MIP. The command below creates a local virtual environment first, so it works on Debian/Ubuntu systems that block global pip installs.

Install

python3 -m venv .venv && .venv/bin/python -m pip install localmip

Import check

.venv/bin/python -c "import localmip_py as lm; print(lm.LocalMIP)"

Use the source build below when you need the standalone Local-MIP CLI, the static C++ library, or the repository examples.

Before Source Builds

Requirements

The source CLI path needs the build tools below. The wheel path above targets CPython 3.8-3.12 on Linux x86_64 and needs the standard venv module. A legacy local bindings build also requires pybind11>=2.10,<3.

CMake3.15+

CompilerC++20 GCC/Clang

Toolsbash, make, POSIX utilities

Source CLI Path

Build and Run in 3 Steps

These commands were verified against the current solver tree and keep the first run to 10 seconds.

Step 1

Download

Use a tagged source archive for a stable snapshot, or clone the repository for the latest development version.

v2.0.10 Release Repository

Step 2

Build

From the repository root, start with the release build. Use ./build.sh all when you also need examples and Python bindings; the local bindings step expects CPython 3.8-3.12 and pybind11>=2.10,<3.

./build.sh release

Step 3

Run

Run from build/ so bundled test-set paths resolve correctly.

cd build
./Local-MIP -i ../test-set/2club200v15p5scn.mps -t 10

The build writes the solver binary and static library under build/. Use ./Local-MIP --help for the exact CLI flags in your version.

Basic Usage

Command-Line Parameters

A representative CLI invocation looks like this:

./Local-MIP --model_file <model_file> [options]

CLI flags support both short and long forms, for example -i and --model_file.

Representative options from the current CLI:

Flag	Parameter	Description	Default
`-i`	`model_file`	Path to input model file (.mps/.lp)	Required
`-t`	`time_limit`	Positive time limit in seconds	10
`-s`	`sol_path`	Path to output solution file (.sol)	`""`
`-k`	`start_sol_path`	Path to a warm-start solution file (.sol)	`""`
`-m`	`start`	Built-in start method: `zero`, `random`, `objective`, or `locks`	`zero`
`-l`	`log_obj`	Log a bounded incumbent trace (at most 3000 points)	1
`-S`	`random_seed`	Random seed (0 uses default)	0

Example

# Solve the bundled sample for 10 seconds and save a solution
cd build
./Local-MIP --model_file ../test-set/2club200v15p5scn.mps --time_limit 10 --sol_path quickstart.sol --log_obj 1

Warm Start and Built-in Initialization

Use a previous .sol file with --start_sol_path, or select a built-in start method with --start. A warm-start file may omit variables; supplied values are validated against finiteness, bounds, and integrality.

# Reuse a prior solution
./Local-MIP --model_file ../test-set/2club200v15p5scn.mps --start_sol_path quickstart.sol --time_limit 10

# Start from objective-guided finite bounds
./Local-MIP --model_file ../test-set/2club200v15p5scn.mps --start objective --time_limit 10

Using a Parameter Configuration File

Instead of passing all parameters via command line, you can use a configuration file:

cd build
./Local-MIP --param_set_file ../default.set --model_file ../test-set/2club200v15p5scn.mps

The repository includes default.set as a template with all available parameters and their default values. Its default time limit is 10 seconds.

Configuration file format

One parameter per line: parameter_name = value
Lines starting with # or ; are comments
Command-line arguments override values from the configuration file

After the First Run

Next Steps

Explore More

Docs and examples

Use the website pages for releases, tutorials, and runnable examples.

Software Tutorials Examples

Use as a Library

C++ and Python

Use the PyPI package for Python, or integrate Local-MIP through the C++ API and repository examples.

PyPI Package Solver README Python README

Run Tests

CTest workflow

Start with the fast unit-test subset from build/; reserve full CTest runs for longer development sweeps.

ctest --output-on-failure -R "^(api|callbacks|start_strategies|constraint_recognition|scoring|model_manager|reader|move_operations|neighbor_config|shared_model|activity_arithmetic)$"

Test instructions

References

Where to Look Next

These files contain the full setup details:

PyPI 2.0.10CPython 3.8-3.12 manylinux x86_64 wheels Solver READMECLI build, run, parameter-file usage, and tests Python bindings READMEPython installation, demos, and development builds Model API example READMEExample-specific build and run notes Example directoryPer-demo README files and source code v2.0.10 ReleaseDownload the current tagged source archive Repository homeBrowse the project online

Checks

Troubleshooting

Build issue

CMake version too old

Install CMake 3.15 or later.

Build issue

C++20 compiler not found

Install a C++20-capable compiler (GCC or Clang) and ensure it is on your PATH.

Runtime issue

Cannot find input file

Run from the build/ directory or use absolute paths.

Runtime issue

Slow performance

Prefer a release build when you want the best runtime performance.

Support

Getting Help

Tutorials Examples GitHub repository Open an issue

Back to Home Software