-
Notifications
You must be signed in to change notification settings - Fork 0
/
ippo.py
182 lines (161 loc) · 7.17 KB
/
ippo.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
import torch
import torch.nn.functional as F
import numpy as np
import rl_utils
from tqdm import tqdm
import matplotlib.pyplot as plt
from ma_gym.envs.combat.combat import Combat
class PolicyNet(torch.nn.Module):
def __init__(self, state_dim, hidden_dim, action_dim):
super(PolicyNet, self).__init__()
self.fc1 = torch.nn.Linear(state_dim, hidden_dim)
self.fc2 = torch.nn.Linear(hidden_dim, hidden_dim)
self.fc3 = torch.nn.Linear(hidden_dim, action_dim)
def forward(self, x):
x = F.relu(self.fc2(F.relu(self.fc1(x))))
return F.softmax(self.fc3(x), dim=1)
class ValueNet(torch.nn.Module):
def __init__(self, state_dim, hidden_dim):
super(ValueNet, self).__init__()
self.fc1 = torch.nn.Linear(state_dim, hidden_dim)
self.fc2 = torch.nn.Linear(hidden_dim, hidden_dim)
self.fc3 = torch.nn.Linear(hidden_dim, 1)
def forward(self, x):
x = F.relu(self.fc2(F.relu(self.fc1(x))))
# x = torch.sigmoid(self.fc3(x))
return self.fc3(x)
class PPO:
def __init__(self, state_dim, hidden_dim, action_dim, actor_lr, critic_lr,
lmbda, 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.lmbda = lmbda
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 = rl_utils.compute_advantage(self.gamma, self.lmbda, td_delta.cpu()).to(self.device)
old_log_probs = torch.log(self.actor(states).gather(1, actions)).detach()
log_probs = torch.log(self.actor(states).gather(1, actions))
ratio = torch.exp(log_probs - old_log_probs)
surr1 = ratio * advantage
surr2 = torch.clamp(ratio, 1 - self.eps, 1 + self.eps) * advantage # 截断
actor_loss = torch.mean(-torch.min(surr1, surr2)) # PPO损失函数
critic_loss = torch.mean(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()
actor_lr = 3e-4
critic_lr = 1e-3
num_episodes = 50000
hidden_dim = 64
gamma = 0.99
lmbda = 0.97
eps = 0.2
# device = torch.device("cpu")
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
team_size = 3
grid_size = (15, 15)
# 创建Combat环境，格子世界的大小为15x15，己方智能体和敌方智能体数量都为2
env = Combat(grid_shape=grid_size, n_agents=team_size, n_opponents=team_size)
state_dim = env.observation_space[0].shape[0]
action_dim = env.action_space[0].n
# 两个智能体共享同一个策略
agent = PPO(state_dim, hidden_dim, action_dim, actor_lr, critic_lr, lmbda, eps, gamma, device)
win_list = []
for i in range(10):
with tqdm(total=int(num_episodes / 10), desc='Iteration %d' % i) as pbar:
for i_episode in range(int(num_episodes / 10)):
transition_dict_1 = {
'states': [],
'actions': [],
'next_states': [],
'rewards': [],
'dones': []
}
transition_dict_2 = {
'states': [],
'actions': [],
'next_states': [],
'rewards': [],
'dones': []
}
transition_dict_3 = {
'states': [],
'actions': [],
'next_states': [],
'rewards': [],
'dones': []
}
s = env.reset()
terminal = False
while not terminal:
a_1 = agent.take_action(s[0])
a_2 = agent.take_action(s[1])
a_3 = agent.take_action(s[2])
next_s, r, done, info = env.step([a_1, a_2, a_3])
transition_dict_1['states'].append(s[0])
transition_dict_1['actions'].append(a_1)
transition_dict_1['next_states'].append(next_s[0])
# transition_dict_1['rewards'].append(r[0])
transition_dict_1['rewards'].append(
r[0] + 100 if info['win'] else r[0] - 0.1)
transition_dict_1['dones'].append(False)
transition_dict_2['states'].append(s[1])
transition_dict_2['actions'].append(a_2)
transition_dict_2['next_states'].append(next_s[1])
# transition_dict_2['rewards'].append(r[1])
transition_dict_2['rewards'].append(
r[1] + 100 if info['win'] else r[1] - 0.1)
transition_dict_2['dones'].append(False)
transition_dict_3['states'].append(s[2])
transition_dict_3['actions'].append(a_3)
transition_dict_3['next_states'].append(next_s[2])
# transition_dict_2['rewards'].append(r[1])
transition_dict_3['rewards'].append(
r[2] + 100 if info['win'] else r[2] - 0.1)
transition_dict_3['dones'].append(False)
s = next_s
terminal = all(done)
win_list.append(1 if info["win"] else 0)
agent.update(transition_dict_1)
agent.update(transition_dict_2)
agent.update(transition_dict_3)
if (i_episode + 1) % 100 == 0:
pbar.set_postfix({
'episode':
'%d' % (num_episodes / 10 * i + i_episode + 1),
'return':
'%.3f' % np.mean(win_list[-100:])
})
pbar.update(1)
win_array = np.array(win_list)
# 每100条轨迹取一次平均
win_array = np.mean(win_array.reshape(-1, 100), axis=1)
episodes_list = np.arange(win_array.shape[0]) * 100
plt.plot(episodes_list, win_array)
plt.xlabel('Episodes')
plt.ylabel('Win rate')
plt.title('IPPO on Combat')
plt.show()