PyTorch 1.x Reinforcement Learning Cookbook

How to do it...

Let's develop the Windy Gridworld environment:

1. Import the necessary modules, NumPy, and the discrete class, from Gym:

>>> import numpy as np>>> import sys>>> from gym.envs.toy_text import discrete

2. Define four actions:

>>> UP = 0>>> RIGHT = 1>>> DOWN = 2>>> LEFT = 3

3. Let's start by defining the __init__ method in the WindyGridworldEnv class:

>>> class WindyGridworldEnv(discrete.DiscreteEnv): ...     def __init__(self): ...         self.shape = (7, 10) ...         nS = self.shape[0] * self.shape[1] ...         nA = 4 ...         # Wind locations ...         winds = np.zeros(self.shape) ...         winds[:,[3,4,5,8]] = 1 ...         winds[:,[6,7]] = 2 ...         self.goal = (3, 7) ...         # Calculate transition probabilities and rewards ...         P = {} ...         for s in range(nS): ... position = np.unravel_index(s, ...