gym强化学习入门demo——随机选取动作 其实有了这些动作和反馈值以后就可以用来训练DNN网络了...

 

复制代码# -*- coding: utf-8 -*- import gym import time env = gym.make('CartPole-v0') observation = env.reset() print(observation) print("env actionspace：") print(env.action_space) print("env observationspace：") print(env.observation_space) print(env.observation_space.high) print(env.observation_space.low) count = 0 for t in range(100): #随机选择一个动作 action = env.action_space.sample() #执行动作 获取环境反馈 observation, reward, done, info = env.step(action) #如果玩死了就退出 if done: break env.render() count+=1 time.sleep(0.2) print(count)

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
# -*- coding: utf-8 -*-
import gym
import time
env = gym.make('CartPole-v0')
observation = env.reset()
print(observation)

print("env actionspace：")
print(env.action_space)

print("env observationspace：")
print(env.observation_space)
print(env.observation_space.high)
print(env.observation_space.low)

count = 0
for t in range(100):
    #随机选择一个动作
    action = env.action_space.sample()
    #执行动作 获取环境反馈
    observation, reward, done, info = env.step(action)
    #如果玩死了就退出
    if done:
        break
    env.render()
    count+=1
    time.sleep(0.2)
print(count)

 

效果图：