google/or-tools

OR-Tools is Google's open-source suite for combinatorial optimization: you describe a problem such as scheduling, routing, allocation or packing, and the library returns a solution. It suits developers and researchers who can express a problem mathematically and want a choice of solver backends rather than a hosted scheduling service.

What it solves

The suite includes CP* and CP-SAT constraint programming solvers, Glop and PDLP linear programming solvers, wrappers around commercial and other open-source solvers including mixed-integer ones, bin packing and knapsack algorithms, Traveling Salesman and Vehicle Routing algorithms, and graph algorithms such as shortest paths and min/max cost flow About OR-Tools. The core is written in C++, with wrappers for Python, C# and Java About OR-Tools.

How a workflow looks

A typical use is a routing or assignment problem: model the variables and constraints, choose a solver, run it, and interpret the returned assignments. Ready-made samples are grouped by component, including routing/constraint solving, SAT, linear solving, and graph algorithms, plus full examples in C++, .NET, Java, Python and Jupyter notebooks Codemap. The project points new users to its developer-guide tutorials and samples rather than to a single command-line tool Quick Start.

Requirements and build paths

The README says the suite has been tested on Ubuntu 18.04 LTS and up (64-bit), macOS Mojave with Xcode 9.x (64-bit), and Windows with Visual Studio 2022 (64-bit) Installation. It builds with Make (described as legacy), CMake or Bazel; each path has its own build instructions, so a working build depends on following the matching one Installation. Binary and source installation options are documented on the installation page Installation instructions. There is also a dedicated installation page rather than one universal prebuilt binary assumption.

Limits to keep in mind

This is a solver library, not an application: the README does not promise that every problem is tractable, and the choice among CP-SAT, linear solvers, and third-party wrappers affects the model form and available results. License is Apache-2.0, which permits commercial use under its terms License.