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...

 

# -*- 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：