Reinforcement learning cliff-v0 states

cliff-states.py

# function that takes the action and outputs state + reward
def make_step(step, episode_reward, cliff):
    new_state, reward, done, _ =  cliff.step(step)
    print(f'New state {new_state} with extra award {reward} to total {episode_reward}')
    cliff.render()
    return episode_reward + reward
  
 # walk the entire grid
env = gym.make('cliff-v0')
state = env.reset()
total_reward = 0

# 3 times up
for _ in range(3):
    total_reward = make_step(2, total_reward, env)
# all the way to the right
for _ in range(11):
    total_reward = make_step(1, total_reward, env)
# one down
total_reward = make_step(3, total_reward, env)
# to the left
for _ in range(11):
    total_reward = make_step(0, total_reward, env)
# one down
total_reward = make_step(3, total_reward, env)
# to the right
for _ in range(11):
    total_reward = make_step(1, total_reward, env)
# the last step down
total_reward = make_step(3, total_reward, env)

output.txt

New state 4 with extra award -1 to total 0
New state 8 with extra award -1 to total -1
New state 12 with extra award -1 to total -2
New state 13 with extra award -1 to total -3
New state 14 with extra award -1 to total -4
New state 15 with extra award -1 to total -5
New state 16 with extra award -1 to total -6
New state 17 with extra award -1 to total -7
New state 18 with extra award -1 to total -8
New state 19 with extra award -1 to total -9
New state 20 with extra award -1 to total -10
New state 21 with extra award -1 to total -11
New state 22 with extra award -1 to total -12
New state 23 with extra award -1 to total -13
New state 19 with extra award -1 to total -14
New state 18 with extra award -1 to total -15
New state 17 with extra award -1 to total -16
New state 16 with extra award -1 to total -17
New state 15 with extra award -1 to total -18
New state 14 with extra award -1 to total -19
New state 13 with extra award -1 to total -20
New state 12 with extra award -1 to total -21
New state 11 with extra award -1 to total -22
New state 10 with extra award -1 to total -23
New state 9 with extra award -1 to total -24
New state 8 with extra award -1 to total -25
New state 4 with extra award -1 to total -26
New state 5 with extra award -1 to total -27
New state 6 with extra award -1 to total -28
New state 7 with extra award -1 to total -29
New state 8 with extra award -1 to total -30
New state 9 with extra award -1 to total -31
New state 10 with extra award -1 to total -32
New state 11 with extra award -1 to total -33
New state 12 with extra award -1 to total -34
New state 13 with extra award -1 to total -35
New state 14 with extra award -1 to total -36
New state 15 with extra award -1 to total -37
New state 11 with extra award 0 to total -38