DQNアルゴリズムの部分コード

if-elseの略語:

e_greedy_increment = None
epsilon_max = 0.9
epsilon = 0 if e_greedy_increment is not None else epsilon_max
print(epsilon)

結果:0.9e_greedy_incrementに値がない場合self.epsilonはselfに設定.epsilon_max=0.9
1次元配列:

import numpy as np
num_episodes = 10000
rewards = np.zeros(num_episodes)
print(rewards)
print(len(rewards))

結果:

[0. 0. 0. ... 0. 0. 0.]
10000

np.array:

import numpy as np

action = []
for i in range(5):
    action.append(i)
arr_actions = np.array(action)
print(arr_actions)

結果:

[0 1 2 3 4]

配列要素の操作:

import numpy as np
L1 = np.zeros(5, dtype=int)
for i in range(5):
    L1[i] = i + 1
print(L1)

結果:

[1 2 3 4 5]