How to ensure that my ga code is good enough?
I guess I should test it on various test functions that are different in structure, that's one obvious thing but, what about the number of chromosomes one start with?
Is it a problem dependant? same question on the number of runs.
I would be very grateful if you could send me also some helpful papers on that issue.
Thanks!
Dear Ihsan,
what do you mean by genetic code:
1. the encoding representation (chromosome structure) or
2. the algorithm implementation variants (operator definition) or
3. the parameters (population size, crossover and mutation rates)?
There are many papers that devoted to GA in different test functions. So, the easiest way to define your starting parametres in certain test function is to based on that investigations and then you will see the effectivness of your modifications (and parametres values) compared to existing ones.
What about real problems - your guessing is absolutely right, cause every problem has his own solutions space and it's require individual approach for modifications and parameters setting up.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running