ADMB-13.2
Released March 15, 2024
The ADMB (Automatic Differentiation Model Builder) software suite is an environment for nonlinear statistical modeling enabling rapid model development, numerical stability, fast and efficient computation, and high accuracy parameter estimates. AD Model Builder is a high level language built around the AUTODIF Library, a C++ language extension which transparently implements reverse mode automatic differentiation. A closely related software package, ADMB-RE, implements random effects in nonlinear models.
ADMB is an open source project. Read the AD Model Builder LICENSE.
ADMB is freely available for download. If you would like to contribute or donate funds, please contact users@admb-project.org.
Fournier, D.A., Skaug, H.J., Ancheta, J., Ianelli, J., Magnusson, A., Maunder, M.N., Nielsen, A., and Sibert, J. 2012. AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optim. Methods Softw. 27:233-249.
The abstract and publication can be freely downloaded from the publisher.
Read release notes for changes and features in CHANGES.
Read the installation procedures in INSTALL for Unix and Windows computers.
User manuals for AD Model Builder, the AUTODIF library and the ADMB-RE nonlinear random effects module are on the ADMB website.
Additional documentation is also available on the ADMB website.
For help and support, email users@admb-project.org.
Mailing list archives
The ADMB software was originally developed by David Fournier of Otter Research Ltd.
Thanks to the following who have contributed to the ADMB Project by developing features, testing, providing fixes, improving documentation and giving suggestions:
Copyright (c) 2008-2024 ADMB Foundation and Regents of the University of California