mopt: Parallel optimizer for R

In tlamadon/mopt: Parallel Derivative-free Moment Estimation

Description

implements MCMC in parallel on OpenMP or MPI

Details

The main goal of this package is to provide a library for simulated moment estimation done in parallel on either MPI or openMP. The code is writtern specificaly for moment estiamtion in the sense that the objective function should return, not only an overall value, but also each moments. This is then used in the analysis part. For instance the researcher can analyze how each moment is affected by each parameter.

The main function is runMOpt, and the associated mopt_config. This gives full control to the researchers of the methods used to look for the minimum distance estimator. Several algorithms are implemented and we encourage other people to write more of them and leverage the parallelization structure built into this library.

A very useful function is compute.slices and the associated plot.slices which will generate plots of how the different moments vary as you vary each parameters separately.

Author(s)

Ran Gu [e-mail], Thibaut Lamadon [e-mail] , Florian Oswald [e-mail]

References

https://github.com/tlamadon/mopt

See Also

runMOpt

tlamadon/mopt documentation built on May 31, 2019, 3:48 p.m.