The flexibility of the constraint integer programming solver SCIP

SCIP is a constraint integer programming solver. Constraint integer programming (CIP) is a novel paradigm which integrates constraint programming and mixed integer programming (MIP) modeling and solving techniques. The solver SCIP can be used as black-box MIP solver and is currently one of the fastest noncommercial solvers. More remarkable, however, is the flexibility of the framework SCIP. Due to underlying design, SCIP can be easily customize to take advantage of certain structures of a given problem class. For example, by adding problem specific separators, presolvers, propagators, or constraint handlers. The latter one can be used to handle nonlinear constraints directly. In this talk, we introduce the design of SCIP and display the possibilities of this framework. Therefore, we present two examples. First, we show how to solve the Traveling Salesman Problem as CIP using SCIP as branch-and-bound framework. Second, we use the Steel Mill Slab Problem to sketch the use of SCIP as branch-and-price framework.