RcppTrust: Thread-Safe C++ Port of the 'trust' Trust Region Optimizer
Does local optimization using two derivatives and trust
regions. Guaranteed to converge to a local minimum of the objective
function. This is a thread-safe C++ port of the algorithm in the
'trust' package by Charles J. Geyer, exposed both as a drop-in R
function and as a C-callable, thread-safe entry point (with function
pointers registered for other packages to call without linking
against this package's shared library) so it can be used from
parallel C++ code such as in 'nlmixr2est'.
| Version: |
0.1.0 |
| Depends: |
R (≥ 4.2.0) |
| Imports: |
Rcpp |
| LinkingTo: |
Rcpp, RcppArmadillo, BH |
| Suggests: |
testthat (≥ 3.0.0), stats, knitr, rmarkdown, trust, microbenchmark |
| Published: |
2026-09-07 |
| DOI: |
10.32614/CRAN.package.RcppTrust (may not be active yet) |
| Author: |
Charles J. Geyer 
[aut],
Matthew Fidler
[aut, cre] |
| Maintainer: |
Matthew Fidler <matthew.fidler at gmail.com> |
| License: |
MIT + file LICENSE |
| URL: |
https://github.com/nlmixr2/RcppTrust/,
https://nlmixr2.github.io/RcppTrust/,
http://www.stat.umn.edu/geyer/trust/ |
| NeedsCompilation: |
yes |
| SystemRequirements: |
C++17 |
| Materials: |
README |
| CRAN checks: |
RcppTrust results |
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=RcppTrust
to link to this page.