Peter, thanks for your comments. My thinking was - use ALL available Operators, Entry Modes and Normalization to build "benchmark". Then narrow it,
build 50.000 systems with 1 Operator only (and still full set of Entry Modes and Normalization) and compare it to "benchmark". Repeat this with 2nd,
3rd and 4th Operator and see, which Operator gives the best result..
Then I did the same with Entry Modes (choose 1 Entry mode and have full set of Operators and full set of Normalization) and Normalization. So you
think this is not working and right process?
If not, than the second option would be to test all possible combinations (meaning Operator 1, Entry mode 1, Normalization 1; then Operator 1, Entry
Mode 1, Normalization 2; etc.).
What is your (or community) approach to this, please? Thanks. |