-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtrainer.py
126 lines (106 loc) · 5.07 KB
/
trainer.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
'''''''''
@file: trainer.py
@author: MRL Liu
@time: 2021/2/15 15:43
@env: Python,Numpy
@desc: Maze项目的训练器，负责切换不同的训练流程
@ref:
@blog: https://blog.csdn.net/qq_41959920
'''''''''
from matplotlib import pyplot as plt
plt.rcParams['font.sans-serif']=['SimHei'] #使用中文字符
plt.rcParams['axes.unicode_minus'] = False #显示负数的负号
class Trainer(object):
def __init__(self,env,agent):
self.env = env
self.agent = agent
# 每回合最大步数记录器
self.list_step_train = []
self.list_reward_train = []
def train_q_learning(self,max_episodes):
print('仿真训练任务启动...')
# 训练主循环
for episode in range(1,max_episodes+1):
step_counter = 0
reward_episode = 0
# 获取初始环境状态
observation = self.env.reset()
# 开始本回合的仿真
while True:
self.env.render()
# 获取动作和环境反馈
action = self.agent.choose_action(observation)# agent根据当前状态采取动作
observation_, reward, done = self.env.step(observation,action)# env根据动作做出反馈
# 学习本回合的经验(s, a, r, s)
#reward-=step_counter*0.01
self.agent.update(observation, action, reward, observation_,done)
# 更新
observation = observation_
step_counter += 1
reward_episode += reward
# 检测本回合是否需要停止
if done:
self.list_reward_train.append(reward_episode) # 记录最大回合奖励
self.list_step_train.append(step_counter) # 记录最大回合数
print('episode：{} ,step：{},reward_episode：{}'.format(episode, step_counter,reward_episode))
break
print('仿真训练任务结束')
def train_sarsa(self,max_episodes):
print('仿真训练任务启动...')
# 训练主循环
for episode in range(1,max_episodes+1):
step_counter = 0
reward_episode = 0
# 获取初始环境状态
observation = self.env.reset()
action = self.agent.choose_action(observation) # agent根据当前状态采取动作
# 开始本回合的仿真
while True:
self.env.render()
# 获取动作和环境反馈
observation_, reward, done = self.env.step(observation,action)# env根据动作做出反馈
action_ = self.agent.choose_action(observation_) # agent根据当前状态采取动作
# 学习本回合的经验(s, a, r, s)
#reward-=step_counter*0.01
self.agent.update(observation, action, reward, observation_,action_,done)
# 更新
observation = observation_
action = action_
step_counter += 1
reward_episode += reward
# 检测本回合是否需要停止
if done:
self.list_reward_train.append(reward_episode) # 记录最大回合奖励
self.list_step_train.append(step_counter) # 记录最大回合数
print('episode：{} ,step：{},reward_episode：{}'.format(episode, step_counter,reward_episode))
break
print('仿真训练任务结束')
def draw_plot(self):
# 创建画布
fig = plt.figure(figsize=(6, 3)) # 创建一个指定大小的画布
# 创建画布
print('绘制数据')
# 添加第1个窗口
ax1 = fig.add_subplot(121) # 添加一个1行1列的序号为1的窗口
# 添加标注
ax1.set_title('训练中的累计步数变化状况', fontsize=14) # 设置标题
ax1.set_xlabel('x-回合数', fontsize=14, fontfamily='sans-serif', fontstyle='italic')
ax1.set_ylabel('y', fontsize=14, fontstyle='oblique')
# 绘制函数
x_data_train = range(1,len(self.list_step_train)+1)
y_data_train = self.list_step_train
line1, = ax1.plot(x_data_train, y_data_train, color='blue', label="训练值")
ax1.legend(handles=[line1], loc=1) # 绘制图例说明
plt.grid(True) # 启用表格
# 添加第1个窗口
ax1 = fig.add_subplot(122) # 添加一个1行1列的序号为1的窗口
# 添加标注
ax1.set_title('训练中的累计奖励变化状况', fontsize=14) # 设置标题
ax1.set_xlabel('x-回合数', fontsize=14, fontfamily='sans-serif', fontstyle='italic')
ax1.set_ylabel('y', fontsize=14, fontstyle='oblique')
# 绘制函数
x_data_train = range(1, len(self.list_reward_train) + 1)
y_data_train = self.list_reward_train
line1, = ax1.plot(x_data_train, y_data_train, color='blue', label="训练值")
ax1.legend(handles=[line1], loc=1) # 绘制图例说明
plt.grid(True) # 启用表格