A new major release of the SCIP Optimization Suite (https://www.scipopt.org/) has been released.
This open-source software suite consists of the constraint integer programming solver and branch-cut-and-price framework SCIP, the generic decomposition framework GCG, the mixed-integer linear presolving library PaPILO, the linear programming solver SoPlex, the task parallelization framework UG, and the modeling language ZIMPL. It includes one of the fastest and most robust free solvers available for mixed-integer linear, mixed-integer nonlinear, and pseudoboolean optimization.
The software is developed at MODAL SynLab at Zuse Institute Berlin in a cooperation of several academic and industrial partners, including MODAL partners HTW Berlin, ZIB, FICO, GAMS, Gurobi, and Siemens.
With the brand-new version 10.0, a new solving mode for exactly solving rational mixed-integer linear programs is now available in SCIP. Further additions are a new presolver for detecting implied integral variables, a novel cut-based conflict analysis, a separator for flower inequalities, two new heuristics, a novel tool for explaining infeasibility, a new interface for a nonlinear solver, as well as improvements in symmetry handling, branching strategies, and Benders' decomposition. For GCG, new data structures and file formats for storing decompositions, newly developed solvers for pricing problems, parallelization of pricing, and support for handling branching and cutting decisions directly in the extended formulation have been added. For an in-depth overview, a release report has been assembled by the 30+ authors of this release: https://optimization-online.org/?p=32699
The new release is available in source code or precompiled form at https://www.scipopt.org/index.php#download.
