Currently, fWeight can be any float, but deopt.m says:
% fWeight (I) DE-stepsize fWeight from interval [0, 2].
however the Differential Evolution book says
The scale factor, $F \in (0, 1+)$, is a positive real number that controls the rate
at which the population evolves. While there is no upper limit on $F$, effec-
tive values are seldom greater than 1.0.
and also
The stated range for F is (0,1), although 1.0 is an empirically de-
rived upper limit in the sense that no function that has been successfully
optimized has required F > 1.
suggesting that there is no upper limit but it should certainly be greater than zero, and rarely needs to be greater than 1. I can't tell whether Storn et al. limited fWeight to [0, 2] as a suggestion based on this, or as a necessity.
Should we limit the size of fWeight to the range [0, 2] in line with the code, the range [0, $\infty$) in line with the theory, or not limit at all?
Currently,
fWeightcan be any float, butdeopt.msays:% fWeight (I) DE-stepsize fWeight from interval [0, 2].however the Differential Evolution book says
and also
suggesting that there is no upper limit but it should certainly be greater than zero, and rarely needs to be greater than 1. I can't tell whether Storn et al. limited
fWeightto [0, 2] as a suggestion based on this, or as a necessity.Should we limit the size of$\infty$ ) in line with the theory, or not limit at all?
fWeightto the range [0, 2] in line with the code, the range [0,