The variable projection algorithm of Golub and Pereyra (1973) has proven to be quite valuable in the solution of nonlinear least squares problems in which a substantial number of the parameters are linear. Its advantages are efficiency and, more importantly, a better likelihood of finding a global minimizer rather than a local one. The purpose of our work is to provide a "modern" implementation of this algorithm so that it can be easily used. We consider how 35 years of changes in hardware and software influence our software design, and we provide a more robust algorithm that can also handle constraints on the parameters.
Citation: ACM Transactions on Mathematical Software
Pub Type: Journals
nonlinear least squares, variable projection algorithm, varpro