目前流行的強化學習算法包括 Q-learning、SARSA、DDPG、A2C、PPO、DQN 和 TRPO。 這些算法已被用于在遊戲、機器人和決策制定等各種應用中，并且這些流行的算法還在不斷發展和改進，本文我們将對其做一個簡單的介紹。
1、Q-learning
Q-learning：Q-learning 是一種無模型、非政策的強化學習算法。 它使用 Bellman 方程估計最佳動作值函數，該方程疊代地更新給定狀态動作對的估計值。 Q-learning 以其簡單性和處理大型連續狀态空間的能力而聞名。
下面是一個使用 Python 實作 Q-learning 的簡單示例：
import numpy as np
# Define the Q-table and the learning rate
Q = np.zeros((state_space_size, action_space_size))
alpha = 0.1
# Define the exploration rate and discount factor
epsilon = 0.1
gamma = 0.99
for episode in range(num_episodes):
current_state = initial_state
while not done:
# Choose an action using an epsilon-greedy policy
if np.random.uniform(0, 1) < epsilon:
action = np.random.randint(0, action_space_size)
else:
action = np.argmax(Q[current_state])
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Update the Q-table using the Bellman equation
Q[current_state, action] = Q[current_state, action] + alpha * (reward + gamma * np.max(Q[next_state]) - Q[current_state, action])
current_state = next_state 上面的示例中，state_space_size 和 action_space_size 分别是環境中的狀态數和動作數。 num_episodes 是要為運作算法的輪次數。 initial_state 是環境的起始狀态。 take_action(current_state, action) 是一個函數，它将目前狀态和一個動作作為輸入，并傳回下一個狀态、獎勵和一個訓示輪次是否完成的布爾值。
在 while 循環中，使用 epsilon-greedy 政策根據目前狀态選擇一個動作。 使用機率 epsilon選擇一個随機動作，使用機率 1-epsilon選擇對目前狀态具有最高 Q 值的動作。
采取行動後，觀察下一個狀态和獎勵，使用Bellman方程更新q。 并将目前狀态更新為下一個狀态。這隻是 Q-learning 的一個簡單示例，并未考慮 Q-table 的初始化和要解決的問題的具體細節。
2、SARSA
SARSA：SARSA 是一種無模型、基于政策的強化學習算法。 它也使用Bellman方程來估計動作價值函數，但它是基于下一個動作的期望值，而不是像 Q-learning 中的最優動作。 SARSA 以其處理随機動力學問題的能力而聞名。
import numpy as np
# Define the Q-table and the learning rate
Q = np.zeros((state_space_size, action_space_size))
alpha = 0.1
# Define the exploration rate and discount factor
epsilon = 0.1
gamma = 0.99
for episode in range(num_episodes):
current_state = initial_state
action = epsilon_greedy_policy(epsilon, Q, current_state)
while not done:
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Choose next action using epsilon-greedy policy
next_action = epsilon_greedy_policy(epsilon, Q, next_state)
# Update the Q-table using the Bellman equation
Q[current_state, action] = Q[current_state, action] + alpha * (reward + gamma * Q[next_state, next_action] - Q[current_state, action])
current_state = next_state
action = next_action state_space_size和action_space_size分别是環境中的狀态和操作的數量。num_episodes是您想要運作SARSA算法的輪次數。Initial_state是環境的初始狀态。take_action(current_state, action)是一個将目前狀态和作為操作輸入的函數，并傳回下一個狀态、獎勵和一個訓示情節是否完成的布爾值。
在while循環中，使用在單獨的函數epsilon_greedy_policy(epsilon, Q, current_state)中定義的epsilon-greedy政策來根據目前狀态選擇操作。使用機率 epsilon選擇一個随機動作，使用機率 1-epsilon對目前狀态具有最高 Q 值的動作。
上面與Q-learning相同，但是采取了一個行動後，在觀察下一個狀态和獎勵時它然後使用貪心政策選擇下一個行動。并使用Bellman方程更新q表。
3、DDPG
DDPG 是一種用于連續動作空間的無模型、非政策算法。 它是一種actor-critic算法，其中actor網絡用于選擇動作，而critic網絡用于評估動作。 DDPG 對于機器人控制和其他連續控制任務特别有用。
import numpy as np
from keras.models import Model, Sequential
from keras.layers import Dense, Input
from keras.optimizers import Adam
# Define the actor and critic models
actor = Sequential()
actor.add(Dense(32, input_dim=state_space_size, activation='relu'))
actor.add(Dense(32, activation='relu'))
actor.add(Dense(action_space_size, activation='tanh'))
actor.compile(loss='mse', optimizer=Adam(lr=0.001))
critic = Sequential()
critic.add(Dense(32, input_dim=state_space_size, activation='relu'))
critic.add(Dense(32, activation='relu'))
critic.add(Dense(1, activation='linear'))
critic.compile(loss='mse', optimizer=Adam(lr=0.001))
# Define the replay buffer
replay_buffer = []
# Define the exploration noise
exploration_noise = OrnsteinUhlenbeckProcess(size=action_space_size, theta=0.15, mu=0, sigma=0.2)
for episode in range(num_episodes):
current_state = initial_state
while not done:
# Select an action using the actor model and add exploration noise
action = actor.predict(current_state)[0] + exploration_noise.sample()
action = np.clip(action, -1, 1)
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Add the experience to the replay buffer
replay_buffer.append((current_state, action, reward, next_state, done))
# Sample a batch of experiences from the replay buffer
batch = sample(replay_buffer, batch_size)
# Update the critic model
states = np.array([x[0] for x in batch])
actions = np.array([x[1] for x in batch])
rewards = np.array([x[2] for x in batch])
next_states = np.array([x[3] for x in batch])
target_q_values = rewards + gamma * critic.predict(next_states)
critic.train_on_batch(states, target_q_values)
# Update the actor model
action_gradients = np.array(critic.get_gradients(states, actions))
actor.train_on_batch(states, action_gradients)
current_state = next_state 在本例中，state_space_size和action_space_size分别是環境中的狀态和操作的數量。num_episodes是輪次數。Initial_state是環境的初始狀态。Take_action (current_state, action)是一個函數，它接受目前狀态和操作作為輸入，并傳回下一個操作。
4、A2C
A2C（Advantage Actor-Critic）是一種有政策的actor-critic算法，它使用Advantage函數來更新政策。 該算法實作簡單，可以處理離散和連續的動作空間。
import numpy as np
from keras.models import Model, Sequential
from keras.layers import Dense, Input
from keras.optimizers import Adam
from keras.utils import to_categorical
# Define the actor and critic models
state_input = Input(shape=(state_space_size,))
actor = Dense(32, activation='relu')(state_input)
actor = Dense(32, activation='relu')(actor)
actor = Dense(action_space_size, activation='softmax')(actor)
actor_model = Model(inputs=state_input, outputs=actor)
actor_model.compile(loss='categorical_crossentropy', optimizer=Adam(lr=0.001))
state_input = Input(shape=(state_space_size,))
critic = Dense(32, activation='relu')(state_input)
critic = Dense(32, activation='relu')(critic)
critic = Dense(1, activation='linear')(critic)
critic_model = Model(inputs=state_input, outputs=critic)
critic_model.compile(loss='mse', optimizer=Adam(lr=0.001))
for episode in range(num_episodes):
current_state = initial_state
done = False
while not done:
# Select an action using the actor model and add exploration noise
action_probs = actor_model.predict(np.array([current_state]))[0]
action = np.random.choice(range(action_space_size), p=action_probs)
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Calculate the advantage
target_value = critic_model.predict(np.array([next_state]))[0][0]
advantage = reward + gamma * target_value - critic_model.predict(np.array([current_state]))[0][0]
# Update the actor model
action_one_hot = to_categorical(action, action_space_size)
actor_model.train_on_batch(np.array([current_state]), advantage * action_one_hot)
# Update the critic model
critic_model.train_on_batch(np.array([current_state]), reward + gamma * target_value)
current_state = next_state 在這個例子中，actor模型是一個神經網絡，它有2個隐藏層，每個隐藏層有32個神經元，具有relu激活函數，輸出層具有softmax激活函數。critic模型也是一個神經網絡，它有2個隐含層，每層32個神經元，具有relu激活函數，輸出層具有線性激活函數。
使用分類交叉熵損失函數訓練actor模型，使用均方誤差損失函數訓練critic模型。動作是根據actor模型預測選擇的，并添加了用于探索的噪聲。
5、PPO
PPO（Proximal Policy Optimization）是一種政策算法，它使用信任域優化的方法來更新政策。 它在具有高維觀察和連續動作空間的環境中特别有用。 PPO 以其穩定性和高樣品效率而著稱。
import numpy as np
from keras.models import Model, Sequential
from keras.layers import Dense, Input
from keras.optimizers import Adam
# Define the policy model
state_input = Input(shape=(state_space_size,))
policy = Dense(32, activation='relu')(state_input)
policy = Dense(32, activation='relu')(policy)
policy = Dense(action_space_size, activation='softmax')(policy)
policy_model = Model(inputs=state_input, outputs=policy)
# Define the value model
value_model = Model(inputs=state_input, outputs=Dense(1, activation='linear')(policy))
# Define the optimizer
optimizer = Adam(lr=0.001)
for episode in range(num_episodes):
current_state = initial_state
while not done:
# Select an action using the policy model
action_probs = policy_model.predict(np.array([current_state]))[0]
action = np.random.choice(range(action_space_size), p=action_probs)
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Calculate the advantage
target_value = value_model.predict(np.array([next_state]))[0][0]
advantage = reward + gamma * target_value - value_model.predict(np.array([current_state]))[0][0]
# Calculate the old and new policy probabilities
old_policy_prob = action_probs[action]
new_policy_prob = policy_model.predict(np.array([next_state]))[0][action]
# Calculate the ratio and the surrogate loss
ratio = new_policy_prob / old_policy_prob
surrogate_loss = np.minimum(ratio * advantage, np.clip(ratio, 1 - epsilon, 1 + epsilon) * advantage)
# Update the policy and value models
policy_model.trainable_weights = value_model.trainable_weights
policy_model.compile(optimizer=optimizer, loss=-surrogate_loss)
policy_model.train_on_batch(np.array([current_state]), np.array([action_one_hot]))
value_model.train_on_batch(np.array([current_state]), reward + gamma * target_value)
current_state = next_state 6、DQN
DQN（深度 Q 網絡）是一種無模型、非政策算法，它使用神經網絡來逼近 Q 函數。 DQN 特别适用于 Atari 遊戲和其他類似問題，其中狀态空間是高維的，并使用神經網絡近似 Q 函數。
import numpy as np
from keras.models import Sequential
from keras.layers import Dense, Input
from keras.optimizers import Adam
from collections import deque
# Define the Q-network model
model = Sequential()
model.add(Dense(32, input_dim=state_space_size, activation='relu'))
model.add(Dense(32, activation='relu'))
model.add(Dense(action_space_size, activation='linear'))
model.compile(loss='mse', optimizer=Adam(lr=0.001))
# Define the replay buffer
replay_buffer = deque(maxlen=replay_buffer_size)
for episode in range(num_episodes):
current_state = initial_state
while not done:
# Select an action using an epsilon-greedy policy
if np.random.rand() < epsilon:
action = np.random.randint(0, action_space_size)
else:
action = np.argmax(model.predict(np.array([current_state]))[0])
# Take the action and observe the next state and reward
next_state, reward, done = take_action(current_state, action)
# Add the experience to the replay buffer
replay_buffer.append((current_state, action, reward, next_state, done))
# Sample a batch of experiences from the replay buffer
batch = random.sample(replay_buffer, batch_size)
# Prepare the inputs and targets for the Q-network
inputs = np.array([x[0] for x in batch])
targets = model.predict(inputs)
for i, (state, action, reward, next_state, done) in enumerate(batch):
if done:
targets[i, action] = reward
else:
targets[i, action] = reward + gamma * np.max(model.predict(np.array([next_state]))[0])
# Update the Q-network
model.train_on_batch(inputs, targets)
current_state = next_state 上面的代碼，Q-network有2個隐藏層，每個隐藏層有32個神經元，使用relu激活函數。該網絡使用均方誤差損失函數和Adam優化器進行訓練。
7、TRPO
TRPO （Trust Region Policy Optimization）是一種無模型的政策算法，它使用信任域優化方法來更新政策。 它在具有高維觀察和連續動作空間的環境中特别有用。
TRPO 是一個複雜的算法，需要多個步驟群組件來實作。TRPO不是用幾行代碼就能實作的簡單算法。
是以我們這裡使用實作了TRPO的現有庫，例如OpenAI Baselines，它提供了包括TRPO在内的各種預先實作的強化學習算法，。
要在OpenAI Baselines中使用TRPO，我們需要安裝:
pip install baselines 然後可以使用baselines庫中的trpo_mpi子產品在你的環境中訓練TRPO代理，這裡有一個簡單的例子:
import gym
from baselines.common.vec_env.dummy_vec_env import DummyVecEnv
from baselines.trpo_mpi import trpo_mpi
#Initialize the environment
env = gym.make("CartPole-v1")
env = DummyVecEnv([lambda: env])
# Define the policy network
policy_fn = mlp_policy
#Train the TRPO model
model = trpo_mpi.learn(env, policy_fn, max_iters=1000) 我們使用Gym庫初始化環境。然後定義政策網絡，并調用TRPO子產品中的learn()函數來訓練模型。
還有許多其他庫也提供了TRPO的實作，例如TensorFlow、PyTorch和RLLib。下面時一個使用TF 2.0實作的樣例
import tensorflow as tf
import gym
# Define the policy network
class PolicyNetwork(tf.keras.Model):
def __init__(self):
super(PolicyNetwork, self).__init__()
self.dense1 = tf.keras.layers.Dense(16, activation='relu')
self.dense2 = tf.keras.layers.Dense(16, activation='relu')
self.dense3 = tf.keras.layers.Dense(1, activation='sigmoid')
def call(self, inputs):
x = self.dense1(inputs)
x = self.dense2(x)
x = self.dense3(x)
return x
# Initialize the environment
env = gym.make("CartPole-v1")
# Initialize the policy network
policy_network = PolicyNetwork()
# Define the optimizer
optimizer = tf.optimizers.Adam()
# Define the loss function
loss_fn = tf.losses.BinaryCrossentropy()
# Set the maximum number of iterations
max_iters = 1000
# Start the training loop
for i in range(max_iters):
# Sample an action from the policy network
action = tf.squeeze(tf.random.categorical(policy_network(observation), 1))
# Take a step in the environment
observation, reward, done, _ = env.step(action)
with tf.GradientTape() as tape:
# Compute the loss
loss = loss_fn(reward, policy_network(observation))
# Compute the gradients
grads = tape.gradient(loss, policy_network.trainable_variables)
# Perform the update step
optimizer.apply_gradients(zip(grads, policy_network.trainable_variables))
if done:
# Reset the environment
observation = env.reset() 在這個例子中，我們首先使用TensorFlow的Keras API定義一個政策網絡。然後使用Gym庫和政策網絡初始化環境。然後定義用于訓練政策網絡的優化器和損失函數。
在訓練循環中，從政策網絡中采樣一個動作，在環境中前進一步，然後使用TensorFlow的GradientTape計算損失和梯度。然後我們使用優化器執行更新步驟。
這是一個簡單的例子，隻展示了如何在TensorFlow 2.0中實作TRPO。TRPO是一個非常複雜的算法，這個例子沒有涵蓋所有的細節，但它是試驗TRPO的一個很好的起點。
總結
以上就是我們總結的7個常用的強化學習算法，這些算法并不互相排斥，通常與其他技術(如值函數逼近、基于模型的方法和內建方法)結合使用，可以獲得更好的結果。
作者:Siddhartha Pramanik
![THE BODY IS NOT A GIVEN: JOINT AGENT POLICY LEARNING AND MORPHOLOGY EVOLUTION[圖]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)