Solver Links

MINOPT home	MINOPT Model Library	MINOPT Solver Links	MINOPT Purchasing & Licensing	MINOPT Software

MINOPT solver links

To solve the various problems and subproblems that MINOPT formulates, external solvers are required. To solve a particular type of problem, one of the listed solvers with the corresponding capability is required.

MINOPT currently has links with the following solvers:

MINOPT implements the following algorithms for the solution of MINLP problems:

• Generalized Benders Decomposition (GBD)
• Outer Approximation and its variants (OA, OA/ER, OA/ER/AP)
• Generalized Cross Decomposition (GCD)

For the solution of dynamic problems, MINOPT implements a control parameterization approach. This requires the integration of the system of differential and algebraic equations (DAEs) used to model the dynamic system coupled with the sensitivity analysis of the DAEs.

Please be aware of the following MINOPT requirements:

• To solve MINLP problems, both an NLP solve and an MILP solver are required.
• To solve NLP/DAE and OCP problems an NLP solver and an integrator are required.
• To solve MINLP/DAE and MIOCP problems, and NLP solver, an MILP solver, and an integrator are all required.

Special note for users who wish to use two or more of the solvers from the Systems Optimization Laboratory (MINOS, NPSOL, SNOPT).

CPLEX is a Division of ILOG which provides the CPLEX optimization package for the solution of linear, mixed-integer, and quadratic programming problems. MINOPT currently takes advantage of the linear and mixed-integer programming solvers as well as many of the CPLEX features and all of the CPLEX options.

MINOPT is currently linked to CPLEX version 6.0 and requires the CPLEX Linear Optimizer Base System and the CPLEX Callable Library. If the modeling requires mixed-integer capabilities, the Mixed-Integer Solver is also required.

For more information about CPLEX including how to purchase it, visit their website or contact them at

ILOG CPLEX Division
889 Alder Avenue, Suite 200
Incline Village, NV 89451
USA
Tel: 702-831-7744
Fax: 702-831-7755

LPSOLVE (lp_solve) is a freely available LP/MILP solver written by Michel Berkelaar at Eindhoven University of Technology in Eindhoven of the Netherlands. MINOPT currently uses version 2.2 of LPSOLVE which can be obtained via anonymous ftp at ftp://ftp.ics.ele.tue.nl/pub/lp_solve. The compiled versions of the library required by MINOPT are available at the MINOPT software archive.

MINOS is a solver for sparse LP and NLP problems. Linear models are solved using an efficient reduced gradient technique while nonlinear models are solved using a method that iteratively solves subproblems with linearized constraints and an augmented Lagrangian objective function.

MINOS has been developed at the Systems Optimization Laboratory (SOL) at Stanford University.

MINOPT currently links to version 5.5 of MINOS. Further information regarding MINOS and NPSOL can be found at http://www.stanford.edu/~saunders/brochure/brochure.html.
SOL optimization software is distributed by SBSI:

Special note for users who wish to use two or more of the solvers from the Systems Optimization Laboratory (MINOS, NPSOL, SNOPT).

NPSOL is a solver for dense LP and NLP problems. Linear models are solved using the package LSSOL which is part of NPSOL. Nonlinear models are solved using a dense Sequential Quadratic Programming (SQP) algorithm.

NPSOL has been developed at the Systems Optimization Laboratory (SOL) at Stanford University.

MINOPT currently links to version 4.06 of NPSOL. Further information regarding NPSOL and MINOS can be found at http://www.stanford.edu/~saunders/brochure/brochure.html.
SOL optimization software is distributed by SBSI:

Special note for users who wish to use two or more of the solvers from the Systems Optimization Laboratory (MINOS, NPSOL, SNOPT).

SNOPT is a solver for sparse LP and NLP problems. Nonlinear models are solved using a sparse Sequential Quadratic Programming (SQP) algorithm.

SNOPT has been developed at the University of California at San Diego by Philip E. Gill.

MINOPT currently links to version 5.3 of SNOPT. Further information regarding SNOPT can be found at http://www.stanford.edu/~saunders/brochure/brochure.html.
SOL optimization software is distributed by SBSI:

Special note for users who wish to use two or more of the solvers from the Systems Optimization Laboratory (MINOS, NPSOL, SNOPT).

DASOLV is an integrator for large, sparse systems of Differential and Algebraic Equations (DAEs). DASOLV implements a backward difference formula method for the efficient solution of the DAEs along with the sensitivity equations.

DAESSA is an integrator for systems of Differential and Algebraic Equations (DAEs) coupled with sensitivity analysis.

DAESSA is a variant of the DASSL software which has been modified to handle sensitivity evaluation as well as explicit and implicit discontinuities. DASSL was developed by Linda R. Petzold, Peter N. Brown, Alan C. Hindmarsh, and Clement W. Ulrich at the Center for Computational Sciences & Engineering at Lawrence Livermore National Laboratory. DAESSA was developed by Carl A. Schweiger at the Computer Aided Systems Laboratory at Princeton University It was developed particularly for use with MINOPT and can be obtained along with the MINOPT software distribution.

DAESSA requires LAPACK for the linear algebra calculations. This can be obtained from the Netlib Repository under the heading LAPACK--Linear Algebra PACKage. DAESSA also requires the MINPACK library which also can be obtained from the Netlib Repository under the heading minpack. For convenience an abbreviated version of this library containing the necessary functions and subroutines is also located at the MINOPT software archive.

MINOPT home	MINOPT Model Library	MINOPT Solver Links	MINOPT Purchasing & Licensing	MINOPT Software

Copyright © 1998 Princeton University
All Rights Reserved