IBM ILOG CPLEX is the best known and most widely used large-scale solver. Its efficiency and robustness have been demonstrated over two decades in thousands of commercial installations worldwide.
Developer: IBM ILOG
Current version: 12.6
Problem types supported: Linear and quadratic optimization in continuous and integer variables. Support is provided for both convex and nonconvex quadratic objectives, and for convex quadratic constraints.
Special forms detected: Logical implications in the form of “indicator” constraints. Convex quadratic expressions in objectives, and convex quadratic constraints of elliptic and conic types.
Algorithms available: For continuous problems, primal and dual simplex, interior-point (barrier); for integer problems, advanced branch-and-bound with presolve, feasibility heuristics, and cut generators. For continuous problems comprised mostly or entirely of linear network flow constraints, network simplex.
Special features: Shared-memory parallel processing for barrier, branch-and-bound. Concurrent optimization by several methods to determine best choice. Special facilities for parameter tuning and infeasibility diagnosis.
See also: ILOG CP for AMPL, covered by the CPLEX for AMPL license, for access to constraint programming and more general CPLEX handling of logical constraints.
Instructions for joining the IBM Academic Initiative