mxComputeTryHard: Repeatedly attempt a compute plan until successful
Description
The provided compute plan is run until the status code indicates
success (0 or 1). It gives up after a small number of retries.
Usage
mxComputeTryHard(plan, ..., freeSet = NA_character_, verbose = 0L, location = 1, scale = 0.25, maxRetries = 3L)
Arguments
plan
compute plan to optimize the model
...
Not used. Forces remaining arguments to be specified by name.
freeSet
names of matrices containing free variables
verbose
level of debugging output
location
location of the perturbation distribution
scale
scale of the perturbation distribution
maxRetries
maximum number of plan evaluations per invocation (including the first evaluation)
Details
Upon failure, start values are randomly perturbed. Currently only
the uniform distribution is implemented. The distribution is
parametrized by arguments location and scale. The
location parameter is the distribution's median. For the uniform
distribution, scale is the absolute difference between its
median and extrema (i.e., half the width of the rectangle). Each
start value is multiplied by a random draw and then added to a
random draw from a distribution with the same scale but
with a median of zero.
References
Shanno, D. F. (1985). On Broyden-Fletcher-Goldfarb-Shanno method. Journal of
Optimization Theory and Applications, 46(1), 87-94.