print x, population[0].fitness
# evaluate children
for individual in range(0, population_size):
children_pop[individual].evaluate()
children_pop.sort(key=lambda f:f.fitness, reverse=True)
# create a new population and fill it with the 20 best ones from parents and 20 best ones from children
new_population = []
for x in range(0, int(0.1 * population_size)):
new_population.append(population[x])
new_population.append(children_pop[x])
for x in range(0, int(0.4 * population_size)):
new_population.append(random.choice(children_pop))
new_population.append(random.choice(population))
#new_population.append(random.choice(children_pop))
population = new_population
# sort new population and keep track of the best player
from jumpmain import game_function
print best.fitness, '< fitness ', best.weights
ann = best.copyNeuron()
print game_function(ann, True)
best = population[0]
print x, population[0].fitness
# evaluate children
for individual in range(0, population_size):
children_pop[individual].evaluate()
children_pop.sort(key=lambda f: f.fitness, reverse=True)
# create a new population and fill it with the 20 best ones from parents and 20 best ones from children
new_population = []
for x in range(0, int(0.1 * population_size)):
new_population.append(population[x])
new_population.append(children_pop[x])
for x in range(0, int(0.4 * population_size)):
new_population.append(random.choice(children_pop))
new_population.append(random.choice(population))
#new_population.append(random.choice(children_pop))
population = new_population
# sort new population and keep track of the best player
from jumpmain import game_function
print best.fitness, '< fitness ', best.weights
ann = best.copyNeuron()
print game_function(ann, True)
def evaluate_neuron(self):
return jumpmain.game_function(self)
def evaluate_neuron(self):
return jumpmain.game_function(self)