Changeset 1272

Timestamp:

08/05/23 00:58:45 (9 months ago)

Author:

Maciej Komosinski

Message:

Added comments, formatting

Location:

framspy/evolalg

Files:

• base/experiment_niching_abc.py (modified) (3 diffs)
• structures/individual.py (modified) (1 diff)

framspy/evolalg/base/experiment_niching_abc.py

@@ -141 +141 @@
             newpop.append(Individual().copyFrom(ind))
 
-        pop_offspring = population+newpop
+        pop_offspring = population + newpop # this is OK for NSGA2, but TODO verify if this should also be used for NSLC? 
         print(len(pop_offspring))
-        if self.fit == "nslc":
+        if self.fit == "nslc": # TODO should NSLC be also equipped with a novelty archive? (with an admittance threshold?)
             self.do_nslc_dissim(pop_offspring)
         elif self.fit == "nsga2":
@@ -156 +156 @@
             # saved generation has been completed, start with the next one
             self.current_generation += 1
-            print("...Resuming from saved state: population size = %d, hof size = %d, stats size = %d, archive size = %d, generation = %d/%d" % (len(self.population_structures.population), len(self.hof),
-                                                                                                                                                 len(self.stats),  (len(self.population_structures.archive)), self.current_generation, generations))  # self.current_generation (and g) are 0-based, parsed_args.generations is 1-based
+            print("...Resuming from saved state: population size = %d, hof size = %d, stats size = %d, archive size = %d, generation = %d/%d" % (len(self.population_structures.population), len(self.hof), len(self.stats), (len(self.population_structures.archive)), self.current_generation, generations))  # self.current_generation (and g) are 0-based, parsed_args.generations is 1-based
         else:
             self.initialize_evolution(self.genformat, initialgenotype)
@@ -190 +189 @@
         parser.add_argument("-fit",type= str, default="raw",
                         help="Fitness type, availible types: niching, novelty, knn_niching (local), knn_novelty (local), nsga2, nslc and raw (default)")
-        parser.add_argument("-archive",type= int, default=50,
-                            help="Maximum archive size")
+        parser.add_argument("-archive",type= int, default=50, help="Maximum archive size")
         parser.add_argument("-normalize",type= str, default= "max",
                             help="What normalization to use for the dissimilarity matrix: max (default}, sum, or none")

framspy/evolalg/structures/individual.py

@@ -3 +3 @@
 
 class Individual:
-    only_positive_fitness = True
+    only_positive_fitness = True # Note: when using diversification techniques (e.g. niching), setting this to False and allowing negative fitness values requires verifying/improving diversification formulas. Dividing fitness by similarity (or multiplying by diversity) may have undesired consequences when fitness can be both positive and negative (e.g. low similarity may make positive fitness values higher and negative fitness values lower, while the intention would be to improve fitness for low similarity).
 
     def __init__(self):