def evaluate_ann(ann):
score = blocky_bird.game_function(ann, False)
return score
# Evaluate fitness of each individual - make each individual object play the game
for individual in population:
individual.evaluate()
print "Scores for all individuals: " + str(list(i.fitness for i in population))
# Find the best individual
best = max(population, key = lambda x:x.fitness)
print "Best: " + str(best.fitness)
# Create new population, first adding the most fit individual from the old one
new_population = [best]
for x in range(1, pop_size):
if random.random() < 0.5:
# Mutate a random individual from the old population
child = random.choice(population).mutate()
new_population.append(child)
else:
# Breed and crossover two individuals from the old population
p1 = random.choice(population)
p2 = random.choice(population)
child = p1.crossover(p2)
new_population.append(child)
population = new_population
print ""
ann = best.create_neuron()
print "Score: " + str(game_function(ann, True))
