q-table.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Q-Table implementation to solve the FrozenLake RL problem\n",
    "# Adapted from https://gist.github.com/awjuliani/4d69edad4d0ed9a5884f3cdcf0ea0874"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gym\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "env = gym.make('FrozenLake-v0', is_slippery=False)\n",
    "np.set_printoptions(suppress=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.         0.77378094 0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.81450625 0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.857375   0.        ]\n",
      " [0.         0.         0.9025     0.        ]\n",
      " [0.         0.95       0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         0.         0.        ]\n",
      " [0.         0.         1.         0.        ]\n",
      " [0.         0.         0.         0.        ]]\n"
     ]
    }
   ],
   "source": [
    "# q-table\n",
    "q = np.zeros([env.observation_space.n, env.action_space.n])\n",
    "\n",
    "# discounting rate gamma\n",
    "y = 0.95\n",
    "\n",
    "# run for x amount of episodes\n",
    "for eps in range(2000):\n",
    "    # the current state (which is zero at the first timestep)\n",
    "    state = env.reset()\n",
    "\n",
    "    # run for a maximum of 100 steps\n",
    "    for _ in range(100):\n",
    "        # choose an action by selecting the action with the highest Q-value for the current state\n",
    "        # also adds noise to aid in exploration\n",
    "        noise = np.random.randn(1, env.action_space.n) * (1 / (eps + 1))\n",
    "        action = np.argmax(q[state, :] + noise)\n",
    "        \n",
    "        # get the new state and reward from the environment\n",
    "        nstate, reward, done, info = env.step(action)\n",
    "\n",
    "        # update the the q-table\n",
    "        # new q-value for action = reward + discount * maximum reward in next state\n",
    "        q[state, action] = reward + y * np.max(q[nstate, :])\n",
    "        \n",
    "        # set new state\n",
    "        state = nstate\n",
    "        \n",
    "        # stop if episode has ended\n",
    "        if done:\n",
    "            break\n",
    "            \n",
    "print(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[41mS\u001b[0mFFF\n",
      "FHFH\n",
      "FFFH\n",
      "HFFG\n",
      "\n",
      "  (Down)\n",
      "SFFF\n",
      "\u001b[41mF\u001b[0mHFH\n",
      "FFFH\n",
      "HFFG\n",
      "\n",
      "  (Down)\n",
      "SFFF\n",
      "FHFH\n",
      "\u001b[41mF\u001b[0mFFH\n",
      "HFFG\n",
      "\n",
      "  (Right)\n",
      "SFFF\n",
      "FHFH\n",
      "F\u001b[41mF\u001b[0mFH\n",
      "HFFG\n",
      "\n",
      "  (Right)\n",
      "SFFF\n",
      "FHFH\n",
      "FF\u001b[41mF\u001b[0mH\n",
      "HFFG\n",
      "\n",
      "  (Down)\n",
      "SFFF\n",
      "FHFH\n",
      "FFFH\n",
      "HF\u001b[41mF\u001b[0mG\n",
      "\n",
      "  (Right)\n",
      "SFFF\n",
      "FHFH\n",
      "FFFH\n",
      "HFF\u001b[41mG\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "# display solution\n",
    "done = False\n",
    "state = env.reset()\n",
    "env.render()\n",
    "\n",
    "while not done:\n",
    "    action = np.argmax(q[state, :])\n",
    "    state, _, done, _ = env.step(action)\n",
    "    print()\n",
    "    env.render()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}