minizinc-solver

A javascript library to priovide an API to a constraints solver using the MiniZinc constraints modelling language.

Why

Constraints programming provides some powerful techniques to solve certain sets of problems. Many solvers exist, but not many seem to have javascript bindings. This library provides an API to solve problems specified in the MiniZinc modelling language so that this can be more easily integrated in larger programs.

Example

var solver = require('minizinc-solver');
 
var problem = buildMiniZincProblemString();
solver.solve(problem).then(doSomethingWithProblemOutput);

Usage

solve(minizincProblem: string, callback: (error?, result: string) => void): void

It takes a string containing a problem specified in the MiniZinc language and returns a Promise, which may return one of three ways:

• succeed with a string containing the output of the problem ** The format of the output is defined within the MiniZinc problem using the normal "output" command
• succeed with null if the problem cannot be solved
• fail with an error of the following form if there is a syntax error: { type: 'syntax_error', message: <string> }
• fail with an error of the following form if the problem is not solved within 30 seconds: { type: 'timeout_error', message: <string> }

Typescript definitions are also provided.

solveWithTimeout(minizincProblem: string, timeoutInMillis: number, callback: (error?, result: string) => void): void

This is the same as solve except that the timeout time can be specified.

Installation

This depends on having a minizinc installation already.

1. Install the binaries for your system here. This library uses the Gecode solver, so ensure the MiniZinc installation includes gecode binaries.
2. Add the binaries to your PATH: export PATH=$PATH:/path/to/MiniZincIDE-2.2.3-bundle-linux/bin
3. Install this library: npm install minizinc-solver

Testing

Tests are written using Mocha and Chai. To run them, use:

npm test

License

MIT

This project uses a MiniZinc compiler to translate constraints into a language that solvers can understand. Source code for the MiniZinc tool chain is available at http://www.minizinc.org/2.0/index.html.