setupOptPathLoggingMonitor: Logging monitor.
In jakobbossek/ecr: Evolutionary Computing in R

Description Usage Arguments Value

This is the default logging function used by ecr. It makes use of the fantastic OptPath which allows optimizers to log evaluated points. The logger is registered to be called once the EA is initialized (after the creation of the initial population), right after each survival selection and once the EA finished. By default the logger stores the entire population, i.e., the individuals, the fitness values and moreover some additional statistics of the fitness value distribution.

A word of causion: In case of a custom representation, which can not be stored in the OptPath by default since the individuals are more complex than normal parameters, the serialized individuals are stored in the OptPath.

1	setupOptPathLoggingMonitor(step = 1L, log.extras.fun = NULL)

step

[integer(1)]
After how many generations should the logger be active? Default is 1, which means that logging takes place in each generation.

log.extras.fun

[function | NULL]
Function which expects the internal optimization state opt.state and additional parameters ... and returns a numeric vector of scalar qualities computed based e.g. on the individual genomes or the fitness values. The function is called every time the logger is called in each generation. The results are then stored additionally in the OptPath. Default is NULL, which means no extra logging.

[ecr_monitor]

jakobbossek/ecr documentation built on May 18, 2019, 9:09 a.m.