seibert.breed: Combine two values of a parameter type using weighted...
In paleolimbot/easyoptim: Easy Muti-Parameter Optimisation

Description Usage Arguments Value References Examples

View source: R/gap_seibert.R

Combine two values of a parameter type using weighted probabilities

1
2
3

seibert.breed(param, value1, value2, prob = list(breed.use1 = 0.41, breed.use2
  = 0.41, breed.rinterpolate = 0.16, breed.mutate = 0.02), seed = NULL,
  cols = "result", validate = TRUE)

`param`	A 'param' object
`value1`	The first value
`value2`	The second value
`prob`	A named `list` in the form `functionname=probability`. The function described by `functionname` should take the parameter as its first argument and (at least) two vector parameters of arbitrary length and return a vector of identical length. Examples include breed.use1, breed.use2, breed.rinterpolate, and breed.mutate.
`seed`	A random number seed for replicability.
`cols`	A combination of 'result', 'func', 'value1', and/or 'value2'. Passing multiple column names will result in a data.frame; passing a single column will result in a vector. This is helpful for debugging.
`validate`	TRUE to validate arguments, FALSE otherwise. In most cases this function gets called hundreds to thousands of times using values that were generated by the function itself. Using validation may help track down bugs when unexpected output occurs.

A vector of new parameter values, or a data.frame if multiple cols are passed.

Seibert, J. 2000. Multi-criteria calibration of a conceptual runoff model using a generic algorithm. Hydrology and Earth System Sciences 4(2): 215<e2><80><93>224.

p <- param.discrete(c('heads', 'tails', 'edge of coin'), weights=c(0.45, 0.45, 0.1))
v1 <- random.value(p, n=10)
v2 <- random.value(p, n=10)
seibert.breed(p, v1, v2, cols=c('value1', 'value2', 'func', 'result'))