CONOPT Documentation Page

CONOPT is a general-purpose optimization system for large-scale nonlinear models. CONOPT employs a feasible path algorithm that is based on active set methods, particularly suited to large, sparse models. CONOPT can be used as an dedicated non-linear solver, either stand-alone or integrated with in a larger system, by interfacing directly through the API. When using the CONOPT API, the user must define routines to evaluate non-linear expressions. This can achieved through hand-coded evaluation routines, or the integration with automatic differentiation packages, such as ADOL-C and CppAD. As an alternative, CONOPT can be used through modeling systems such as AIMMS, AMPL, GAMS and LINGO.

This documentation provides numerous resources for getting started with CONOPT. The Introduction provides an overview of the CONOPT library and some details on how to build your own project. Following this, it is recommended to work through the Tutorial and look at the Examples using CONOPT. Finally, the CONOPT API (Fortran, C, C++, Python and Java) is detailed in Public API.

Introduction

• The CONOPT algorithm
• Library Organization
• The Mathematical Model
• Overview over CONOPT routines – Fortran users
• Overview over CONOPT routines – C users
• ChangeLog

Tutorial and Examples

• Tutorial
• Examples using CONOPT

Using the CONOPT API

• Overview of the CONOPT API and Callback functions
• The public API of CONOPT
• Additional information