- Add support for optional set variables (
var opt set of int, including sets of enums). Avar opt setis either a set or absent (<>), and absent is different from the empty set. The set operations (in,subset,union,intersect,cardetc.) ignore absent operands, like the other operations on option types. Par optional sets support the same operations, including optional sets of floats. - Allow record literals in data files (
.dznfiles and data given with-D) to omit optional fields of the declared record type. The missing fields are absent, so an array of records can be given as[(y: 1), (x: 2, y: 3)]for typerecord(opt int: x, int: y). This was already the case for JSON input, which now also supports record types that are combined using++. - Normalize linear expressions (term order, sign, and common divisor), so that common subexpression elimination finds more equal expressions.
- Decide at the end of flattening which linear expressions get a variable of
their own. An expression that more than one item reads is defined once, and
an expression with a single reader is written out in that reader. Linear
expressions that differ only by a constant offset or a common factor share
one definition. The option
mzn_opt_eager_linear_aggregationrestores the old behaviour, and the linear library sets it. - Flatten the definitions of
var boolvariables (top-level and in lets), and of the elements ofvar boolarrays, after their uses, in the context of those uses. A variable that is only used positively can now be half-reified, together with the expressions in its definition, as when the definition is written in place. - Plus 30+ bugfixes!
MiniZinc
MiniZinc is a high-level constraint modelling language that allows you to easily express and solve discrete optimisation problems.
Latest release: 2.11.0 (changelog)
MiniZinc is developed at Monash University with support from OPTIMA.
Why choose MiniZinc?
High-level language
Model constraint satisfaction and optimisation problems using readable, intuitive logical constructs, with full type safety.
Solver independent
Run the same model using a variety of different solver technologies, without being locked into a particular solver.
Pre-defined constraints
MiniZinc features a large library of pre-defined constraints, making modelling complex relationships simple.
Fully featured development environment
Use the MiniZinc IDE to write and run models, visualise search and solutions, profile performance and more.
In-depth learning materials
MiniZinc has comprehensive documentation, including the tutorial, several courses, the user manual, and library documentation.
Integrate MiniZinc into your project
MiniZinc can be embedded into Python and JavaScript programs, and can even be used entirely from within the browser.
Need help with MiniZinc? deciZns provides specialist support to organisations tackling complex strategic and operational decisions using MiniZinc. From modelling and language features to solver performance, deciZns has the expertise to help you get to make real progress.
Latest Release
MiniZinc 2.11.0
Released 18 November 2026
Get your MiniZinc stickers, mugs, t-shirts etc. (sold at cost price)!