Princeton Library of Nonlinear Programming Models

PrincetonLib is a collection of nonlinear programming (NLP) models. The purpose of the collection is to provide algorithm developers of nonlinear optimization codes with a large and varied set of both theoretical and practical test models. It also aids in the software quality assurance process by providing a set of tools to facilitate benchmarking and performance analysis.

The original models were in AMPL format and collected by Robert Vanderbei and colleagues at Princeton University. They were translated by Andre Savitsky into GAMS.

The raw GAMS models were converted into GAMS scalar format using the CONVERT facility. For models with multiple solves, only the first solve instance is translated.

Reference

Nonlinear Optimization Models, collected by Robert Vanderbei and colleagues

QuickStart

See the readme.txt

Download

Download PrincetonLib.zip (302Mb)

Note: A * next to the directory name indicates the group is missing some models due to the AMPL source missing. See the group listing for details.

Update (02/21/2013)

Several mistakes in translating the models from AMPL source to GAMS raw and scalar models have been corrected.

Note: If the same instance is contained in several source libraries, then we also kept duplicate versions of this instance. For example, the following instances are known to be duplicates: