Sets suitable defaults for the SACOBRA part of SACOBRA.
With the call setOpts(mySAC,defaultSAC()) it is possible to extend a partial list
mySAC to a list containing all sac-elements (the missing ones are taken from
defaultSAC()).
defaultSAC(DOSAC = 1)[0|1|2] with default 1. 0: COBRA-R settings (turn off SACOBRA), 1: SACOBRA settings, 2: SACOBRA settings with fewer parameters and more online adjustements (aFF and aCF are done parameter free).
a list with the following elements (the values in parantheses [|] are the values
for DOSAC=[0|1|2]):
flag for random start algorithm [FALSE|TRUE|TRUE]
type of the function to calculate probability to start the internal optimizer
with a random starting point[NA|"SIGMOID"|"CONSTANT"] (see function
RandomStart in SACOBRA.R)
maximum probability of a random start when RStype=="SIGMOID"
(see RandomStart in SACOBRA.R). If RStype=="CONSTANT" then
random start is done with a constant probability determined from mean(c(RSmax,RSmin))
[NA|0.3|0.3]
minimum probability of a random start when RStype=="SIGMOID"
(see RandomStart in SACOBRA.R) [NA|0.05|0.05]
If TRUE then in every iteration where the fraction of feasible points in the population
is smaller than 0.05, the RS probability is set to 0.3. [FALSE|FALSE|TRUE]
flag for automatic DRC adjustment [FALSE|TRUE|TRUE]
flag for automatic objective function transformation [FALSE|TRUE|TRUE]
flag for automatic constraint function transformation [FALSE|TRUE|TRUE]
threshold, if FRange is larger than TFRange, then apply automatic objective
function transformation (see plog). [Inf|1e+05|-1]
threshold, if GRatio is larger than TGR, then apply automatic
constraint function transformation. GRatio is the ratio "largest GRange /
smallest GRange" where GRange is the min-max range of a specific constraint.
If TGR < 1, then the transformation is always performed. [Inf|1e+03|-1].
If Cs iterations in a row do not improve the ever-best feasible solution,
then perform a random restart. [10|10|10]
(experimental) flag for objective function transformation with plog,
where the parameter pShift is adapted during iterations.
[FALSE|FALSE|FALSE]
flag for online decision making wether use plog or not according to p-effect plog.
[FALSE|FALSE|TRUE]
Initial pEffect value when using onlinePLOG. If pEffectInit >= 2 then the initial model is built after plog transformation.
[NA|NA|2]
For backward compatibility, a logical DOSAC (deprecated) is mapped from FALSE to 0 and from TRUE to 1.