def eval_genomes(genomes, config):
solution = [random.randint(1, 4) for i in range(4)]
solution_2 = [random.randint(1, 4) for i in range(4)]
for genome_id, genome in genomes:
genome.fitness = 4.0
net = neat.nn.FeedForwardNetwork.create(genome, config)
## master mind
mm = Mastermind(4, 8, solution)
mm_2 = Mastermind(4, 8, solution_2)
hints_table = mm.play_neat_bot(net, show_board=False)
hints_table_2 = mm_2.play_neat_bot(net, show_board=False)
genome.fitness += float(fitness_function(hints_table))
genome.fitness += float(fitness_function(hints_table_2))
# winner_net = p.run(eval_genomes, 2)
solution = [random.randint(1, 8) for i in range(4)]
solutions = [[1, 3, 4, 2], [1, 4, 1, 4], [2, 1, 1, 3], [4, 1, 3, 2]]
solutions = list(itertools.product([1, 2, 3, 4], repeat=4))
solutions = [list(sol) for sol in solutions]
how_many_win = 0
def game_won(hints_table):
winner = False
winning_round = None # 0, 1, ...
for i, row in enumerate(hints_table):
if all(hint == 2 for hint in list(row)):
winner = True
winning_round = i
break
return winner
for s in solutions:
mm = Mastermind(4, 8, s)
hints_table = mm.play_neat_bot(winner_net, win_show=True, show_board=False)
if game_won(hints_table):
how_many_win += 1
print("NN won {0} times per {1} possible.".format(how_many_win, 256))
# p = neat.Checkpointer.restore_checkpoint('neat-checkpoint-299')
# p.run(eval_genomes, 10)