The LiftOpt Package
LiftOpt is a toolkit for nonlinear optimization that aims to extend the advantages of multiple shooting to general NLP problems with nearly no additional computational and implementation effort compared to a classic approach. It contains an implementation of the lifting idea for general nonlinear optimization described in this SIOPT paper, including a lifted Newton, Gauss-Newton and SQP method, as well as the examples presented in the paper. Additional examples can be formulated by the user and solved using the lifting approach.
The source code of the latest release (0.70alpha) can be downloaded from here. It is released under the GNU GPL license. For instructions regarding installation (on LINUX platforms) and basic usage consult the pdf file contained in the release. Please note that the software is still in alpha stage. It is suitable for demonstration and illustration of the lifting idea, and for experiments with your own application problems, but cannot yet be considered as a stable and performance optimized "production" code. You are kindly invited to contribute to the software by sharing your application examples, submitting bugfixes, adding interface to other tools, etc... We are looking forward to your feedback.
- Version 0.70alpha ( Feb 11, 2010 )
Initial public release under GPL.