fixed_factor
For under-relaxation of shared sample increments, apply a fixed factor that is invariant with iteration
Specification
Alias: None
Arguments: REAL
Description
When initial correlation / covariance approximations are inaccurate, iterated sample updates can be overly aggressive, such that it becomes advisable to only accept a portion of this suggested increment through use of an under-relaxation factor.
In the fixed_factor
option, a single relaxation factor is
specified for use on all online iterations.
Usage Tips
Even through the under-relaxation factor does not decay to a value of 1, the sample increments will generally decay toward an integer truncation of zero, terminating the online iteration.
As for other relaxation specification options, the one case where the
relaxation factor will be automatically advanced to 1 is the case
where a max_iterations
constraint will force early termination of
the iteration. This advancement avoids polluting the final results
with an anticipated variance reduction that was not fully realized due
to increment throttling.
This option is equivalent to a factor_sequence
specification with
a single value.
Examples
method,
multilevel_blue
solution_mode online_pilot
relaxation fixed_factor = .8
pilot_samples = 25
seed = 8674132
max_function_evaluations = 500