chinmay8bit

def fill_in_transition(self):
    """
    Compute the transition matrix of the grid
    input: /
    output: T {np.array} -- the transition matrix of the grid
    """
    T = np.zeros((self.state_size, self.state_size, self.action_size)) # Empty matrix of dimension S*S*A

    #### 
    # Add your code here

chinmay8bit

def policy_iteration(self, threshold = 0.0001, gamma = 0.8):
    """
    Policy iteration on GridWorld
    input: 
      - threshold {float} -- threshold value used to stop the policy iteration algorithm
      - gamma {float} -- discount factor
    output:
      - policy {np.array} -- policy found using the policy iteration algorithm
      - V {np.array} -- value function corresponding to the policy 
      - epochs {int} -- number of epochs to find this policy

chinmay8bit