Conference Agenda

The SCIP Optimization Suite is a set of packages for modeling and solving a large variety of optimization problems. At its center is SCIP, an open-source optimization solver for mixed-integer linear and nonlinear optimization problems and a constraint integer programming framework based on a branch-cut-and-price algorithm. This talk will present the developments introduced in SCIP 9.0, including significant improvements and restructuring of symmetry handling, new cutting planes for signomial expressions and a Lagromory cutting plane separator, new primal heuristics, cut selection strategies and branching rules. The presentation will also discuss new features in SCIP-SDP, and new interfaces.

In this talk, we present the recent developments in the Cardinal Optimizer (COPT). We discuss some key techniques that contributed to the performance improvements of our MIP solver and present performance numbers of the latest COPT release for all problem classes.

In this presentation, we will give an overview of the latest enhancements, the newest features, and the most recent performance improvements in the FICO® Xpress Solver for mixed-integer linear and nonlinear optimization problems. These include a new, first-order hybrid gradient algorithm for linear optimization problems, new heuristics, cutting and branching techniques, an augmented API, and updates to our global MINLP solver.

We overview recent enhancements, new features, and performance improvements in our Gurobi 11 release.

In particular, we present our new global MINLP solver in more detail. Previously, non-linear terms have been statically approximated by piecewise linear functions. Now, spatial branching with dynamically adapted outer approximation constraints ensures global optimality, subject to tolerances.

Additionally, our automatic parameter tuning tool, in combination with the Gurobi Cluster Manager, has been improved to dynamically and equally re-distribute Compute Server nodes to all running jobs.