Penneys Spielwahrscheinlichkeit: Wie optimiere ich meinen Code? ⇐ Python

[Anonymous]

Ich simuliere derzeit Penneys Spiel und berechnet die Gewinn-/Verlustwahrscheinlichkeit für 1.000.000 Decks. Mein Code läuft ~ 10 Minuten, aber ich möchte, dass er viel schneller läuft. Ich speichere 100.000 Decks Numpy -Dateien (die ich 10 von 1.000.000 Decks erstelle), durch die ich iteriere.

Code: Select all

sequences = ['000','001','010','011','100','101','110', '111']
n = len(sequences)

# create empty array
penney_prob_arr_wins = np.zeros((n,n))
penney_prob_arr_losses = np.zeros((n,n))

print("Calculating probabilities...generating heatmaps...")
# iterate twice over sequences; create progress bar
for i in tqdm(range(n)):
P1 = list(map(int, sequences[i]))
for j in range(n):
# create all possible sequences for players 1 and 2 so they may be
# used in the penneys_game() function as list objects
P2 = list(map(int, sequences[j]))
# save win rate to variable
win_rate = penneys_game(P1, P2)
# fill dataframe with combonation data
# for each sequence played against another
penney_prob_arr_wins[i,j] = win_rate[0]
penney_prob_arr_losses[i,j] = win_rate[1]
< /code>
Und hier ist die Funktion zur Berechnung von Penneys Spiel: < /p>
def penneys_game(P1: list, # store data
P2: list):
"""
Loop through each card in the deck,
save these cards to a list where the final three cards care always being checked
for a match to the players' sequences
if the player sequence matches the last three, this player gets a trick
"""
# numpy array to store data for win frequency calculation
win_frequency = np.zeros(3)

# list for trick data
trick_check = []

# making sure it can find the correct directory path to the decks
deck_directory = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..", "deck_storage")

# length for calculating win rate
N = 0

# loop through decks in the deck_storage folder and count the length of each one for storage in the N variable
deck_arrays = [file for file in os.listdir(deck_directory)if file.endswith(".npy")]
for deck_array in deck_arrays:
file_path = os.path.join(deck_directory, deck_array)
# load the data; pickled objects are allowed in order to load the array .npy files in the folder
decks = np.load(file_path, allow_pickle = True)
N += len(decks)

# store tricks counts for player 1 and 2 for future calculations
P1_tricks = 0
P2_tricks = 0

for deck in decks:
for card in deck:
trick_check.append(card)
if len(trick_check) >= 3:
if trick_check[-3:] == P1:
P1_tricks += 1
trick_check = []
elif trick_check[-3:] == P2:
P2_tricks += 1
trick_check = []
# calculate win frequency
if P1_tricks > P2_tricks:
win_frequency[0] += 1
elif P2_tricks > P1_tricks:
win_frequency[1] += 1
elif P1_tricks == P2_tricks:
win_frequency[2] += 1
# reset trick check and trick counts after each deck
trick_check = []
P1_tricks = 0
P2_tricks = 0

# calculate win rate
win_rate = win_frequency*1.0/N
return win_rate

Wenn Sie Tipps zum Beschleunigen haben, vielen Dank!