CGN stands for "Constrained Gauss-Newton method". This package contains a Python implementation of the Gauss-Newton method for nonlinear least-squares problems with nonlinear constraints. That is, CGN provides a method for solving constrained optimization problems of the form
Such problems arise frequently in maximum likelihood estimation or in Bayesian maximum-a-posteriori estimation with Gaussian priors.
The Installation Guide explains how the cgn package can be
installed.
For some simple end-to-end use cases, see the Examples. For a
more detailed explanation of the interface to cgn, see the Tutorial.
If you want to apply cgn, you might find it useful to start with the pre-defined
Template, and adapt it according to your needs.
For full details on the individual classes
and functions provided by cgn, please refer to the Documentation.