import random

# 正しく8×8×(list)を作る
Q = [[0 for _ in range(8)] for _ in range(8)]
K = [[list(range(16)) for _ in range(16)] for _ in range(8)]
N = [[[random.randint(0, 15)] for _ in range(8)] for _ in range(8)]

def reward(action, prediction):
    return action != prediction   # True = reward

for _ in range(2**20):
    a = random.randint(0, 7)
    b = random.randint(0, 7)
    c = random.randint(0, len(K[a][b]) - 1)

    action = N[a][b][0]
    prediction = K[a][b][c]

    if reward(action, prediction):
        K[a][b].pop(c)

print(K)
print(N)