S+NuOPT Feature List

• Complete set of powerful and efficient solvers included
• Very fast and robust optimization
• Optimize problems with thousands of variables and constraints
• Complete flexibility in specifying objective function and constraints
• Specify problems in a natural way using SIMPLE, a class of objects in the S language
• Automated symbolic differentiation of objective function
• Complete control over solver tolerance and options

Optimization Problems Solved

• Linear Programming
• Mixed-Integer Linear Programming
• General Convex Programming
• Convex Quadratic Programming
• General Non-Linear Programming
• Mixed Integer Quadratic Programming

Optimization Methods

• Primal-dual interior point method with higher-order correction for linear programming
• Simplex method for linear programming
• Primal-dual interior point method based on line search for general convex Programming models including convex quadratic programming models
• Primal-dual interior point method based on trust region for general non-linear programming models
• Primal-dual interior point method based on quasi-Newton method for general non-linear programming models
• Active set method for convex quadratic programming models and mixed integer quadratic programming models

SIMPLE: A set of classes for defining optimization problems

• "Set" class: element, union, intersection, difference, direct product
• Defining objective functions: expressions, variables, integer variables, parameters.
• Defining contraints: relational, equality, inequality, inclusion in set, conditional expressions.

System Requirements

• S+NuOPT module
• S-PLUS 7 for Windows or Solaris or Linux