ILOG CPLEX algorithms can be accessed from the ILOG CPLEX Component Libraries as well as the ILOG CPLEX Interactive Optimizer, an easy-to-use interactive program. ILOG CPLEX provides all the basic features and utilities for using these solvers: sophisticated problem preprocessing; file reading and writing utilities; reporting; messaging control; interactive revision capability; efficient restart from an advanced basis; sensitivity analysis; and an infeasibility finder.
ILOG CPLEX Simplex Optimizers ILOG CPLEX Simplex Optimizers provide the power to solve quadratic programs and linear programs with millions of constraints and continuous variables, at record-breaking speed. The optimizers include implementations of dual simplex and primal simplex, as well as a network simplex that can solve problems with side constraints. The ILOG CPLEX Presolve algorithms for problem size reduction are integrated into the ILOG CPLEX Simplex Optimizers.
ILOG CPLEX Barrier Optimizer ILOG CPLEX Barrier Optimizer provides an alternative to the simplex method for solving linear and quadratic problems, and an approach for solving quadratically constrained problems and second order cone programming (SOCP) problems. Based on a primal-dual predictor-corrector method, ILOG CPLEX Barrier Optimizer provides unmatched performance for solving large-scale versions of these model types.
For models with linear constraints, the ILOG CPLEX Crossover algorithm converts complex solutions created by the Barrier algorithm into basic solutions. These basic solutions are typically provided by the simplex method and used for fast restarts and sensitivity analysis.
ILOG CPLEX Mixed Integer Optimizer ILOG CPLEX Mixed Integer Optimizer employs state-of-the-art algorithms and techniques to solve difficult mixed integer programs, including problems with quadratic terms in the objective function and/or constraints. The optimizer incorporates the latest research results as well as ILOG's own innovations.
ILOG CPLEX Mixed Integer Optimizer includes sophisticated mixed integer preprocessing routines and implements default strategies that work well for many problems. Users may also customize the cutting-plane and heuristics strategies and include their own when problem-specific techniques are valuable.
