Random agent (run with environment name as argument) for OpenAI gym.

rand.py

#!/usr/bin/env python
import gym
import sys

env = gym.make(sys.argv[1])
num_episodes = 200
max_timestep = 1000

env.monitor.start(sys.argv[1])
for _ in xrange(num_episodes):
    env.reset()
    for _ in xrange(max_timestep):
        a = env.action_space.sample()
        _, _, done, _ = env.step(a)
        if done:
            break
env.monitor.close()