Minimization Methods for Ill-Conditioned Problems

Implementation of methods for minimizing ill-conditioned problems. Currently only includes regularized (quasi-)newton optimization (Kanzow and Steck et al. (2023), ).

Minimization for ill-conditioned problems

Regularized quasi-Newton optimisation

Currently the only function, rnewt implements general-purpose regularized quasi-Newton optimisation routines as presented in Kanzow and Steck (2023). The C++ code is written from scratch, and the More-Thuente linesearch script is an R-port specifically written for this implementation, but translated from the python implementation associated to the article.

References

Kanzow, C., & Steck, D. (2023). Regularization of limited memory quasi-Newton methods for large-scale nonconvex minimization. Mathematical Programming Computation, 15(3), 417-444.

Sugimoto, S., & Yamashita, N. (2014). A regularized limited-memory BFGS method for unconstrained minimization problems. inf. téc.