I have an optimization function that is doing a sort of data mining by filtering criteria to find an optimal solution.
stat_arb = function(data = data_subset_test,
kurt_lower,
kurt_upper,
skew_lower,
skew_upper,
mean_lower,
mean_upper,
sd_lower,
sd_upper,
acf_lower,
acf_upper,
min_row = 3000)
I have many upper and lower values for each, and want to front-load the constraint to make sure the GA is not calculating the function if lower > upper. I worry that using a penalization function within my function will take much more energy to find optimal because of the number of variables, most solutions will not satisfy my constraints. Is there a way to do this with ga() or nsga2() in R? Any help is appreciated.