Q-learningで迷路を解く！(強化学習) - Qiita

# ハイパーパラメータ gamma = 0.7 alpha = 0.5 epsiron = 0.3 # 初期値 pos_now = 0 step = 0 # 報酬 reward = 10 steps = [] for episode in range(50): while True: # epsiron greedy法 random_value = random.random() if random_value>epsiron : # print("Q値更新") # 行動 act = meiro_env[pos_now:pos_now+1].max().sort_values(ascending=False) act_now = act.index[0] # maxQのときの位置 max_Q_pos = meiro_next_act[act_now][pos_now] # maxQ(s,t