7. Temporal Difference Learning

Activity 7.01: Using TD(0) Q-Learning to Solve FrozenLake-v0 Stochastic Transitions

1. Import the required modules:

   import numpy as np
   import matplotlib.pyplot as plt
   %matplotlib inline
   import gym

2. Instantiate the gym environment called FrozenLake-v0 using the is_slippery flag set to True in order to enable stochasticity:

   env = gym.make('FrozenLake-v0', is_slippery=True)

3. Take a look at the action and observation spaces:

   print("Action space = ", env.action_space)
   print("Observation space = ", env.observation_space)

   This will print out the following:

   Action space =  Discrete(4)
   Observation space =  Discrete(16)

4. Create two dictionaries to easily translate the actions numbers into moves:

   actionsDict = {}
   actionsDict[0] = " L "
   actionsDict[1] = " D "
   actionsDict[2] = " R "
   actionsDict[3] = " U "
   actionsDictInv = {}
   actionsDictInv["L"] = 0
   actionsDictInv["D&quot...