Gurobi Optimization, LLC released Gurobi 13.0, which introduces performance improvements across core model types, new algorithm support for large-scale and non-linear problems, GPU acceleration, and expanded cloud-native functionality.
“Today’s organizations face domain-specific optimization challenges that are larger and more complex than ever,” says Oliver Bastert, CTO, Gurobi Optimization. “With Gurobi 13.0, we’re extending our solver technology to handle even larger models and more nonlinear problem types—delivering measurable speed-ups and greater flexibility across a wide range of applications.”
Key takeaways:
· Enhancements include a faster core solver, particularly for difficult mixed-integer programming (MIP) and mixed-integer nonlinear programming (MINLP) models; a new Primal-Dual Hybrid Gradient (PDHG) implementation that allows users to solve some large-scale linear programming problems faster; and a newly supported nonlinear barrier method, which offers faster locally optimal solutions for nonlinear problems.
· Improvements to the Gurobi Cluster Manager and Compute Server include autoscaling of Compute Server clusters, which allows users to automatically scale Gurobi Compute Server deployments; a redesigned web interface aligned with WCAG 2.1 AA standards for better usability; and adoption of a TLS 1.3–only cipher policy to ensure secure communications.
