genopt.control {BurStMisc}

Control parameters for genopt

Description

Returns a list suitable as the control argument of the genopt function.

Usage

  genopt.control(births = 100, random.n = 0, jitters.n = 3,
                 trace = TRUE, eps = 0.1, prob = 0.4, scale.min = 1e-12,
                 maxeval = Inf)

Arguments

births

the number of times two parents are combined to create a new solution.

random.n

the number of random solutions created (and evaluated) before the genetic phase starts.

jitters.n

the number of times a newly successful solution is changed (in the simulated annealing phase) in an attempt to find a nearby better one.

trace

logical value: should progress be printed?

eps

the default value for the scale of the jittering in simulated annealing.

prob

in the selection of parameters between two parents, the probability of each coming from the first parent.

scale.min

the minimum value allowed in the scale argument of genopt.

maxeval

the maximum number of function evaluations allowed. This takes previous function calls into account if funevals is a component of the population argument of genopt.

Value

a list with components:

icontrol

vector of the control parameters that are logically integer (or logical).

dcontrol

vector of the control parameters that are logically real-valued.

See Also

genopt.

Examples

  goc1 <- genopt.control(random.n=200, births=1000, trace=FALSE)

  go1 <- genopt(function(x, other) sum(x, other),
                matrix(rexp(10), nrow=2, ncol=5), lower=0, add.arg=list(other=3),
                control=goc1)

  goc2 <- genopt.control(random.n=0, births=1000, trace=FALSE)

  go2 <- genopt(function(x, other) sum(x, other), go1,
                lower=0, add.arg=list(other=3), control=goc2)