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

 

?

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

# -*- 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：