.. _method-approximate_control_variate-pilot_samples:

"""""""""""""
pilot_samples
"""""""""""""


Initial set of samples for multilevel/multifidelity sampling methods.


.. toctree::
   :hidden:
   :maxdepth: 1



**Specification**

- *Alias:* initial_samples 

- *Arguments:* INTEGERLIST


**Description**


The pilot sample provides initial estimates of variance
and/or correlation within the first iteration of a
multilevel/multifidelity sampling approach.  These initial estimates
then guide the algorithm toward additional sample increments in order
to iteratively achieve a prescribed error at minimum total cost, or
alternatively, minimum error for a prescribed budget.

*Default Behavior*

100 samples per model fidelity and/or discretization level.



*Usage Tips*

The number of specified values can be none (default values used for
all fidelities and levels), one (all fidelities and levels use the
same specified value), the number of discretization levels for every
model (each model uses the same discretization level profile), or the
aggregate number of discretization levels for all models (samples for
each discretization level are distinct for each model).