Dependencies

Optimas relies on the following packages:

• jinja2 - Needed to generate simulation scripts from templates.

• libEnsemble - The backbone of optimas, orchestrates the concurrent evaluation of simulations, the resource detection and allocation, and the communication between simulations and manager.

• mpi4py - Python bindings for MPI. Required for launching parallel simulations.

• pandas - Enable output as pandas DataFrames.

• pydantic - Input validation and object serialization.

• (optional) Ax - Algorithms for Bayesian optimization.

The installed dependencies will determine which generators are available for use. See table below for a summary.

Available generators and their dependencies

Generator	pip install optimas	pip install 'optimas[all]'
LineSamplingGenerator	\(\checkmark\)	\(\checkmark\)
GridSamplingGenerator	\(\checkmark\)	\(\checkmark\)
RandomSamplingGenerator	\(\checkmark\)	\(\checkmark\)
AxSingleFidelityGenerator	\(\times\)	\(\checkmark\)
AxMultiFidelityGenerator	\(\times\)	\(\checkmark\)
AxMultitaskGenerator	\(\times\)	\(\checkmark\)
AxClientGenerator	\(\times\)	\(\checkmark\)