Player
src.ga.playerAI Player driven by a tiny neural network.
OBSERVATION_LABELS = ('Ball dist x', 'Ball dist y', 'Paddle pos x', 'Ball pos x', 'Ball pos y', 'Ball speed x', 'Ball speed y')
module-attribute
ACTION_LABELS = ('Left', 'Right')
module-attribute
IndividualPlayer
Class for an Individual (AI) in the Population.
age = 0
instance-attribute
current_hit_streak = 0
instance-attribute
current_win_streak = 0
instance-attribute
fit_scores = None
instance-attribute
max_hit_streak = 0
instance-attribute
max_win_streak = 0
instance-attribute
neural_net = None
instance-attribute
scores = None
instance-attribute
uid = None
instance-attribute
__init__()
add_hit_to_streak() -> None
Update the streak counter.
add_win_to_streak() -> None
Update the streak counter.
apply_move(zone, move_left: bool, move_right: bool) -> bool
Perform a move based on directly supplied boolean actions from the Batch Brain.
calculate_fitness() -> float
create_neural_net()
get_fitness() -> float
Get the fitness.
get_scores() -> str
increase_age() -> int
Increase age by one.
look(zone) -> np.ndarray
staticmethod
look_into(zone, inputs: np.ndarray) -> None
staticmethod
Fill a reusable seven-value observation buffer.
reset_defaults() -> None
Reset stats and assign new UID for brand-new offspring/pupils.
reset_hit_streak() -> None
Reset the streak counter.
reset_scores() -> None
Specifically, reset game-play stats without wiping UID or Age.
reset_winning_streak() -> None
Reset the streak counter.
think(inputs_list) -> np.ndarray
Feed-forward decision using the neural network.