evoprob: Instance-evolving EA.
In jakobbossek/evoprob: Evolutionary Algorithm to Produce Diverse Problems

Description Usage Arguments Value

View source: R/evoprob.R

To be written ...

evoprob(
  n,
  P,
  maximize = FALSE,
  runner.fun,
  fitness.fun,
  mut.fun,
  diversity.fun = NULL,
  max.iters = 10L,
  max.time = Inf,
  logger = NULL,
  ...
)

`n`	[`integer(1L)`] Desired instance size.
`P`	[`list`] Initial population. This is problem dependent. Therefore, the used must provide it.
`maximize`	[`logical(1)`] Is the goal to maximize performance values? Defaults to `FALSE`.
`runner.fun`	[`function`] Function used to run algorithms on problem instance. Should return aggregated performance values.
`fitness.fun`	[`function`] Fitness function.
`mut.fun`	[`function`] Mutation operator.
`diversity.fun`	[`function`] Diversity function, i.e. function that calculates the diversity of a set of instances.
`max.iters`	[`integer(1)`] Stopping condition: maximum number of iterations.
`max.time`	[`integer(1)`] Maximum time in seconds. Default is `NULL`, i.e. the number of iterations serves as the single termination criterion.
`logger`	[`evoprob_logger`] Optional logger environemt (see `init_logger`). Possible values for parameter `what` are iter = "numeric" Mandatory iteration counter. time = "numeric" Time passed in seconds. P = "list" The population. fP = "list" The fitness values. div = "numeric" Scalar diversity measure. divtab = "list" Vector of absolute frequencies of algorithm rankings. Default is `NULL`, i.e., logging is not active.
`...`	[any] Not used at the moment.