PPO.py

import random

import gymnasium as gym
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F

import utils


class PolicyNet(nn.Module):
    def __init__(self, state_dim, hidden_dim, action_dim):
        super(PolicyNet, self).__init__()
        self.fc1 = nn.Linear(state_dim, hidden_dim)
        self.fc2 = nn.Linear(hidden_dim, action_dim)

    def forward(self, x):
        x = F.relu(self.fc1(x))
        return F.softmax(self.fc2(x), dim=1)


class ValueNet(nn.Module):
    def __init__(self, state_dim, hidden_dim):
        super(ValueNet, self).__init__()
        self.fc1 = nn.Linear(state_dim, hidden_dim)
        self.fc2 = nn.Linear(hidden_dim, 1)

    def forward(self, x):
        x = F.relu(self.fc1(x))
        return self.fc2(x)


def compute_advantage(gamma, lmbda, td_delta):
    td_delta = td_delta.detach().numpy()
    advantage_list = []
    advantage = 0.0
    for delta in td_delta[::-1]:
        advantage = gamma * lmbda * advantage + delta
        advantage_list.append(advantage)
    advantage_list.reverse()
    return torch.tensor(advantage_list, dtype=torch.float)


class PPO:
    ''' PPO算法,采用截断方式 '''

    def __init__(self, state_dim, hidden_dim, action_dim, actor_lr, critic_lr,
                 gae_lambda, epochs, eps, gamma, device):
        self.actor = PolicyNet(state_dim, hidden_dim, action_dim).to(device)
        self.critic = ValueNet(state_dim, hidden_dim).to(device)
        self.actor_optimizer = torch.optim.Adam(self.actor.parameters(),
                                                lr=actor_lr)
        self.critic_optimizer = torch.optim.Adam(self.critic.parameters(),
                                                 lr=critic_lr)
        self.gamma = gamma
        self.gae_lambda = gae_lambda
        self.epochs = epochs  # 一条序列的数据用来训练轮数
        self.eps = eps  # PPO中截断范围的参数
        self.device = device

    def take_action(self, state):
        state = torch.tensor([state], dtype=torch.float).to(self.device)
        probs = self.actor(state)
        action_dist = torch.distributions.Categorical(probs)
        action = action_dist.sample()
        return action.item()

    def update(self, transition_dict):
        states = torch.tensor(transition_dict['states'],
                              dtype=torch.float).to(self.device)
        actions = torch.tensor(transition_dict['actions']
                               ).view(-1, 1).to(self.device)
        rewards = torch.tensor(transition_dict['rewards'],
                               dtype=torch.float).view(-1, 1).to(self.device)
        next_states = torch.tensor(transition_dict['next_states'],
                                   dtype=torch.float).to(self.device)
        dones = torch.tensor(transition_dict['dones'],
                             dtype=torch.float).view(-1, 1).to(self.device)

        td_target = rewards + self.gamma * self.critic(next_states) * (1 - dones)
        td_delta = td_target - self.critic(states)
        advantage = compute_advantage(self.gamma, self.gae_lambda, td_delta.cpu()).to(self.device)
        old_log_probs = torch.log(self.actor(states).gather(1, actions)).detach()

        for _ in range(self.epochs):
            log_probs = torch.log(self.actor(states).gather(1, actions))
            ratio = torch.exp(log_probs - old_log_probs)

            surr_loss = ratio * advantage
            clipped_surr_loss = torch.clamp(ratio, 1 - self.eps, 1 + self.eps) * advantage  # 截断

            actor_loss = torch.mean(-torch.min(surr_loss, clipped_surr_loss))  # PPO损失函数
            critic_loss = F.mse_loss(self.critic(states), td_target.detach())

            self.actor_optimizer.zero_grad()
            self.critic_optimizer.zero_grad()
            actor_loss.backward()
            critic_loss.backward()
            self.actor_optimizer.step()
            self.critic_optimizer.step()


random.seed(0)
np.random.seed(0)
torch.manual_seed(0)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

alg_name = 'PPO'
actor_lr = 3e-4
critic_lr = 3e-4
num_episodes = 1000
hidden_dim = 128
gamma = 0.99
gae_lambda = 0.95
epochs = 10
eps = 0.2

env_name = 'CartPole-v1'
env = gym.make(env_name)
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.n
agent = PPO(state_dim, hidden_dim, action_dim, actor_lr, critic_lr, gae_lambda,
            epochs, eps, gamma, device)

if __name__ == '__main__':
    print(env_name)
    return_list = utils.train_on_policy_agent(env, agent, num_episodes)
    utils.dump(f'./results/{alg_name}.pkl', return_list)
    utils.show(f'./results/{alg_name}.pkl', alg_name)