MINOS
From DDL Wiki
MINOS is a linear programming (LP) and nonlinear programming (NLP) solver package for solving large-scale optimization problems. MINOS is designed to find locally optimal solutions. The nonlinear functions need to be smooth where the first derivative must exist. It implements different algorithms to handle different types of problems:
- Linear programming problem: When objective function and constraints are all linear, MINOS applies the simplex method to find optimal solution.
- Nonlinear objective function with linear constraints: MINOS uses reduced-gradient method to generate search direction.
- Problems with nonlinear constraints: MINOS uses projected Lagrangian method to linearlize the nonlinear constraints and search optimal solution in an augmented Lagrangian form. For highly nonlinear constraints, MINOS may be less efficient than the solvers using SQP algorithm.
External Links
- MINOS homepage
- GAMS/MINOS MINOS NLP solver in GAMS interface
Reference
- Bruce A. Murtagh, Michael A. Saunders, Walter Murray, Philip E. Gill, Ramesh Raman, Erwin Kalvelagen (2002) "GAMS/MINOS: A Solver for Large-Scale Nonlinear Optimization Problems" PDF retreived on March 18, 2007.
