ask for new features about fitness_func #163

Activity

1. ahmedfgad commented on Feb 24, 2023

   

   ahmedfgad

   on Feb 24, 2023

   Last edited by ahmedfgad

   Owner

   More actions

   Can you use the on_fitness() callback to do your calculations? There the ga_instance is passed.

   The fitness of the population will be passed to on_fitness().
2. whubaichuan commented on Feb 24, 2023

   

   whubaichuan

   on Feb 24, 2023

   Author

   More actions

   @ahmedfgad Hej, thanks for your quick reply. Actually, I want to change my fitness_func according to the ga_instance.generations_completed. For example, for the last generation, I want to change some hyperparameters in the fitness_func. Although this could be achieved by calculating the fitness of ga_instance.population manually after ga_instance.run() by adjusting a new fitness_function, how can I do it for example, the 2nd or 4th generation in general?
3. whubaichuan commented on Feb 25, 2023

   

   whubaichuan

   on Feb 25, 2023

   Last edited by whubaichuan

   Author

   More actions

   @ahmedfgad or could you slightly modify the capsulated Pygad-multiprocess function to make the change of generation_idx can be passed to the fitness_func()?

   import signal
   import time
   import sys
   import multiprocessing
   
   generation_idx = 0
   
   def change_idx():
       global generation_idx
       generation_idx =generation_idx+ 1
       print(generation_idx)
   
   def fitness_func():
      
       print("generation_idx = {0}\n".format(generation_idx))
       time.sleep(1)
   
   def main():
       for i in range(3):
           change_idx()
           worker = multiprocessing.Process(target=fitness_func, args=())
           worker.start()
           worker.join()
   
   
   if __name__ == "__main__":
       main()
   
   
   output:
   
   1
   generation_idx = 0
   
   2
   generation_idx = 0
   
   3
   generation_idx = 0
   
4. ahmedfgad

   added

   enhancementNew feature or requestNew feature or request

   on Feb 25, 2023
5. ahmedfgad commented on Feb 25, 2023

   

   ahmedfgad

   on Feb 25, 2023

   Owner

   More actions

   @whubaichuan,

   The ga_instance will be passed to the fitness function in the next release. Thank you!
6. whubaichuan commented on Feb 27, 2023

   

   whubaichuan

   on Feb 27, 2023

   Author

   More actions

   @ahmedfgad Thanks, looking forward to that.
7. ahmedfgad

   added a commit that references this issue on Apr 8, 2023

   PyGAD 3.0.0 Release

   ...

   6659d5d
8. ahmedfgad commented on Apr 9, 2023

   

   ahmedfgad

   on Apr 9, 2023

   Owner

   More actions

   @whubaichuan, PyGAD 3.0.0 is released and the GA instance is passed to the fitness function/method.
9. ahmedfgad

   closed this as completedon Apr 9, 2023
10. whubaichuan commented on Apr 10, 2023

    

    whubaichuan

    on Apr 10, 2023

    Author

    More actions

    @ahmedfgad Thanks a lot.