Has anyone used the optimizer for strategy mining?
Let's say that I had a set of signals and I wanted to use the optimizer to test every possible subset of signal combinations. ie. sig1, sig2, sig1+sig2, sig1+sig3....
So the total number of test cases would be n^2-1, where n= number of features/signals (obviously not testing the case where no signals selected)
Where the optimizer would iterate 1 to n^2-1 in "flags"
This is how the optimizer could iterate over all possible subset combinations, but I'm curious to how to best implement it throughout the EA... any ideas?
Let's say that I had a set of signals and I wanted to use the optimizer to test every possible subset of signal combinations. ie. sig1, sig2, sig1+sig2, sig1+sig3....
So the total number of test cases would be n^2-1, where n= number of features/signals (obviously not testing the case where no signals selected)
Inserted Code
input int flags = 1;
Inserted Code
bool UseSignal(const int signal_num) { return bool(flags&(1<<signal_num)); }
This is how the optimizer could iterate over all possible subset combinations, but I'm curious to how to best implement it throughout the EA... any ideas?