Phil 10.2.19

ASRC GOES 7:00 – 5:00

  • Dissertation. Working on the differences between designed and evolved systems
  • Status report
  • Add statistical tests to the evolver.
    • based on this post, starting with the scikit-learn resample(). Here are the important bits:
      def calc_fitness_stats(self, resample_size:int = 100):
          boot = resample(self.population, replace=True, n_samples=resample_size, random_state=1)
          s = pd.Series(boot)
          conf = st.t.interval(0.95, len(boot)-1, loc=s.mean(), scale= st.sem(boot))
          self.meta_info = {'mean':s.mean(), '5_conf':conf[0], '95_conf':conf[1], 'max':s.max(), 'min':s.min()}
 = s.mean()
    • And the convergence on the test landscape looks good:


  • Added check that the same genome doesn’t get re-run, since it will be run for n times to produce a distribution:
    # randomly breed new genomes with a chance of mutation
    while len(self.current_genome_list) < self.num_genomes:
        g1i = random.randrange(len(self.best_genome_list))
        g2i = random.randrange(len(self.best_genome_list))
        g1 = self.best_genome_list[g1i]
        g2 = self.best_genome_list[g2i]
        g = self.breed_genomes(g1, g2, crossover_rate, mutation_rate)
        match = False
        for gtest in self.all_genomes_list:
            if g.chromosome_dict == gtest.chromosome_dict:
                match = True
        if not match: