예제 #1
0
 def test_halt_expected(self):
     """
     Ensure the function returns true if we're in a halting state.
     """
     halt = make_halt_function([6, 5])
     g1 = Genome([6, 5])
     g1.fitness = MAX_REWARD
     population = [g1, ]
     result = halt(population, 1)
     self.assertTrue(result)
예제 #2
0
 def test_halt_expected(self):
     """
     Ensure the function returns true if we're in a halting state.
     """
     halt = make_halt_function([6, 5])
     g1 = Genome([6, 5])
     g1.fitness = MAX_REWARD
     population = [
         g1,
     ]
     result = halt(population, 1)
     self.assertTrue(result)
예제 #3
0
 def test_halt_not(self):
     """
     Ensures if the fittest genome has fitness < MAX_REWARD then halt
     doesn't succeed.
     """
     halt = make_halt_function([3, 2, 1])
     g1 = Genome([1, 2, 3])
     g1.fitness = MAX_REWARD - 0.1
     g2 = Genome([1, 2, 3])
     g2.fitness = 3
     g3 = Genome([1, 2, 3])
     g3.fitness = 2
     # Any fittest solution with fitness < MAX_REWARD means no halt.
     population = [g1, g2, g3]
     result = halt(population, 1)
     self.assertFalse(result)
예제 #4
0
 def test_halt_not(self):
     """
     Ensures if the fittest genome has fitness < MAX_REWARD then halt
     doesn't succeed.
     """
     halt = make_halt_function([3, 2, 1])
     g1 = Genome([1, 2, 3])
     g1.fitness = MAX_REWARD - 0.1
     g2 = Genome([1, 2, 3])
     g2.fitness = 3
     g3 = Genome([1, 2, 3])
     g3.fitness = 2
     # Any fittest solution with fitness < MAX_REWARD means no halt.
     population = [g1, g2, g3]
     result = halt(population, 1)
     self.assertFalse(result)
예제 #5
0
 def test_halt_checks_suspension_count(self):
     """
     If the solution contains suspensions the halt function should ensure
     that the MAX_REWARD is incremented by the number of suspensions
     (rewarded because they're part of a valid step wise motion).
     """
     halt = make_halt_function([9, 8, 7, 6, 5])
     g1 = Genome([11, 10, 9, 8, 7])
     # only one our of two "correct" dissonances
     g1.fitness = MAX_REWARD + REWARD_SUSPENSION
     population = [g1, ]
     result = halt(population, 1)
     self.assertFalse(result)
     # Try again
     # two out of two "correct" dissonances
     g1.fitness = MAX_REWARD + (REWARD_SUSPENSION * 2)
     population = [g1, ]
     result = halt(population, 1)
     self.assertTrue(result)
예제 #6
0
 def test_halt_checks_suspension_count(self):
     """
     If the solution contains suspensions the halt function should ensure
     that the MAX_REWARD is incremented by the number of suspensions
     (rewarded because they're part of a valid step wise motion).
     """
     halt = make_halt_function([9, 8, 7, 6, 5])
     g1 = Genome([11, 10, 9, 8, 7])
     # only one our of two "correct" dissonances
     g1.fitness = MAX_REWARD + REWARD_SUSPENSION
     population = [
         g1,
     ]
     result = halt(population, 1)
     self.assertFalse(result)
     # Try again
     # two out of two "correct" dissonances
     g1.fitness = MAX_REWARD + (REWARD_SUSPENSION * 2)
     population = [
         g1,
     ]
     result = halt(population, 1)
     self.assertTrue(result)
예제 #7
0
        mutation_range = third.DEFAULT_MUTATION_RANGE
        mutation_rate = third.DEFAULT_MUTATION_RATE
        start_population = third.create_population(population_size, cf)
        fitness_function = third.make_fitness_function(cf)
        generate_function = third.make_generate_function(mutation_range,
            mutation_rate, cf)
        halt_function = third.make_halt_function(cf)
    elif species == 4:
        population_size = fourth.DEFAULT_POPULATION_SIZE
        mutation_range = fourth.DEFAULT_MUTATION_RANGE
        mutation_rate = fourth.DEFAULT_MUTATION_RATE
        start_population = fourth.create_population(population_size, cf)
        fitness_function = fourth.make_fitness_function(cf)
        generate_function = fourth.make_generate_function(mutation_range,
            mutation_rate, cf)
        halt_function = fourth.make_halt_function(cf)

    ga = ga.genetic_algorithm(start_population, fitness_function,
        generate_function, halt_function)
    fitness = 0.0

    counter = 0
    for generation in ga:
        counter += 1
        fitness = generation[0].fitness
        print "--- Generation %d ---" % counter
        print generation[0]
        print fitness
    with open('%s.ly' % output, 'w') as output:
        output.write(lilypond.render(species, cf, generation[0].chromosome))