11. Building an Artificial Intelligence Algorithm

Activity 11.01: Implementing a Double Deep Q-Learning Algorithm to Solve the Cart Pole Problem

Solution

1. In the Chapter11 directory, launch a Jupyter Notebook in your Terminal (macOS or Linux) or Command Prompt window (Windows).
2. After the Jupyter Notebook is launched, create a new directory named Activity11.01. Inside the Activity11.01 directory, create a Python 3 notebook.
3. Inside the Python 3 notebook, import all necessary modules and seed the environment as shown in the following code:

   # import module
   import random
   import numpy as np
   from itertools import count
   from collections import deque
   import torch
   import torch.nn as nn
   import torch.nn.functional as F
   import torch.optim as optim
   import gym
   # make game
   env = gym.make('CartPole-v1')
   # seed the experiment
   env.seed(9)
   np.random.seed(9)
   random.seed(9)
   torch.manual_seed(9)

4. Let's define our DQN as shown in the following code:

   # define our policy...