cartpole_sa.py

import gym
import random
env = gym.make('CartPole-v0')

env.monitor.start('/tmp/cartpole-experiment-3', force=True)

# its simulated annealing like

bestSteps = 0
best = [0, 0, 0, 0]
alpha = 1

for i_episode in xrange(80):
    
    test = [best[i] + (random.random() - 0.5)*alpha for i in range(4)]

    score = 0
    for ep in range(10):  # <-- key thing was to figure out that you need to do 10 tests per point
        observation = env.reset()
        for t in xrange(200): # <-- because you can't go over 200 you need to gain score hight else where
            env.render()
            if sum(observation[i]*test[i] for i in range(4)) > 0:
                action = 1
            else:
                action = 0
            observation, reward, done, info = env.step(action)
            if done:
                break

        score += t

    if bestSteps < score:
        bestSteps = score
        best = test
        alpha *= .9

    print "test:", "[%+1.2f %+1.2f %+1.2f %+1.2f]" % tuple(test), score, 
    print "best:", "[%+1.2f %+1.2f %+1.2f %+1.2f]" % tuple(best), bestSteps, alpha


print "best", best, bestSteps

env.monitor.close()

reproduced here : https://gym.openai.com/evaluations/eval_SZdv2vuThCAu6YzikbBg

i got 1 of 4 attempts to reproduce the results:
141.97 ± 5.43 https://gym.openai.com/evaluations/eval_lzji17hRtiTr9GKJKRClg
9.35 ± 0.08 https://gym.openai.com/evaluations/eval_VChwPSKS0K4SHEs0XRIeQ
9.31 ± 0.08 https://gym.openai.com/evaluations/eval_yvvMBaUESDiaCoZiA4VTqA
199.36 ± 0.34 https://gym.openai.com/evaluations/eval_MIy5Ru6lSKh9RcuNQLz1g

edit: the middle two runs had env.render() commented out