Highlights of Gurobi Optimizer 3.0

Back to News

Gurobi Optimization is pleased to announce the release of Gurobi Optimizer 3.0, providing improved performance and many new features:

  • New parallel barrier optimizer, with crossover: Highly tuned for multi-core architectures, the new barrier solver can provide a significant performance boost when solving large, difficult LP models.
  • Significant improvements in MIP performance: Gurobi Optimizer 3.0 includes a number of MIP enhancements, including symmetry detection, improved cutting planes, new heuristics, and new presolve reductions. This release also includes two new types of cuts, network cuts and sub-MIP cuts, that can greatly improve MIP model formulations. Overall, these enhancements provide major performance improvements on a wide range of MIP models. The new version is 50% faster across our full test set, and 2 to 10 times faster on more difficult test sets.
  • New .NET interface: A new object-oriented interface based on the same object model as our C++, Java, and Python interfaces. Gurobi Optimizer 3.0 includes complete IntelliSense help for use in Visual Studio, as well as a full set of examples in C# and Visual Basic.
  • New MIPFocus parameter: Allows you to modify the MIP solution strategy to better suit the needs of specific models through a single parameter.
  • Improved MIP memory management: Memory usage is reduced for most MIP models, and node files make more effective use of available memory.
  • Additional heuristic control: New parameters provide fine grain control over the feasibility pump and RINS heuristics, allowing you to better customize the heuristic behavior to your models.
  • MPS and LP reader performance improvements: Models read as much as 3x faster.
  • New Mac installer: Simplifies installation on the Mac platform.

Gurobi Roadmap

The next major release - Gurobi Optimizer 4.0 - is targeted for November 2010. This release will include new simplex and barrier QP (quadratic programming) solvers, as well as an MIQP solver (mixed-integer QP). It will also include continued performance improvement in MIP. Future releases beyond Gurobi Optimizer 4.0 will include:

  • Barrier-based SOCP (second order cone programming) solver
  • MISOCP (mixed-integer SOCP) solver