RTMB - R bindings for TMB

Description

The RTMB package provides a native R interface for a substantial subset of TMB so you can avoid coding in C++.

RTMB only affects the TMB function MakeADFun that builds the objective function. Once MakeADFun has been invoked, everything else is exactly the same and models run as fast as if coded in C++.

PROS:

• Fast to change and re-run models because no compilation is needed.
• Debugging can be performed using the normal R debugger rather than gdb.
• Most common TMB features are supported. See current list of working examples.
• Simplified interface can automatically simulate from the model and calculate OSA residuals.
• Easy to provide multiple models in one package.
• Easy to construct a model using functions from external packages and scripts.

More information, including vignettes (introduction / advanced) and documentation, can be found on the RTMB universe page.

Install

CRAN version

install.packages('RTMB')

Development version (Rtools/compiler not needed)

install.packages('RTMB', repos = c('https://kaskr.r-universe.dev', 'https://cloud.r-project.org'))

Source

NOTE: Requires at least TMB-1.9.7.

remotes::install_github("https://github.com/kaskr/RTMB", subdir="RTMB")

Test the package

tinytest::test_package("RTMB")

runs the tests.

Known issues

• If you get a segfault while installing the package, please re-install Rcpp from source and try again (see https://github.com/kaskr/RTMB/issues/5).

Other packages

Experimental parallel version available as a separate package 'RTMBp'

install.packages('RTMBp', repos = c('https://kaskr.r-universe.dev', 'https://cloud.r-project.org'))

Experimental package to solve ODEs

install.packages('RTMBode', repos = c('https://kaskr.r-universe.dev', 'https://cloud.r-project.org'))

documentation