SCIP-SDP

SCIP-SDP is a solver for mixed-integer semidefinite programs (SDPs). It implements a SDP-based branch-and-cut approach. The SDP-relaxations are solved using interior-point SDP-solvers like DSDP, Mosek or SDPA. The SCIP-SDP-Package is a plug-in for the branch-and-cut framework SCIP. The goal of SCIP-SDP is to provide a solver for general mixed-integer SPDs. Thus, it provides data handling procedures, presolve routines, primal heuristics, a dual fixing method, and an linear approximation procedure for MISDPs. SCIP-SDP can also be parallelized with ParaSCIP. This software is currently developed at TU Darmstadt.