dq.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from gridworld import *\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.models import Sequential;\n",
    "from keras.layers import Dense, Activation;\n",
    "from keras import initializers;\n",
    "from keras.optimizers import Adam, SGD;\n",
    "from keras import backend as K\n",
    "import pandas as pd\n",
    "from tqdm import tqdm, trange\n",
    "%matplotlib inline\n",
    "#Deep Q-Learning\n",
    "\n",
    "#gw = gridworld((5, 5), (0, 0), (4, 0), [(o, 0) for o in range(4)], 0.9, 0, 1, 0)\n",
    "gw = gridworld((5, 5), (0, 0), (4, 0), [(o, 0) for o in range(1, 4)], 0.9, 0, 1, 0)\n",
    "# Discount Rate\n",
    "gamma = 0.95"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loss(Qtarget, Q):\n",
    "    return K.sum(K.square(Qtarget - Q))\n",
    "\n",
    "def state_tup_to_num(state_tup):\n",
    "\tstate_num = state_tup[0]*gw.M + state_tup[1]\n",
    "\treturn state_num\n",
    "\n",
    "def state_num_to_tup(state_num):\n",
    "\tstate_tup = (state_num/gw.M, state_num%gw.M)\n",
    "\treturn state_tup\n",
    "\n",
    "# Chance of random action\n",
    "e = 0.1\n",
    "learning_rate = 0.01\n",
    "# Training Episodes\n",
    "episodes = 1000\n",
    "# Max Steps per episode\n",
    "steps = 50\n",
    "trials = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/1000 [00:00<?, ?it/s]/usr/local/lib/python2.7/dist-packages/ipykernel_launcher.py:77: RuntimeWarning: divide by zero encountered in double_scalars\n",
      "100%|██████████| 1000/1000 [00:39<00:00, 25.05it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Successful episodes: 0.644\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 648x1152 with 0 Axes>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8131aff590>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f8131aba310>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Step Moving Average (100-episode window)')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Moves')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'Episode')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8131a9aed0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f8131a75550>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Reward Moving Average (100-episode window)')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Reward')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'Episode')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8131a5e150>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f8131a30750>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Loss Moving Average (100-episode window)')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Loss')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'Episode')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8131a1a350>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f81319eca50>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Random Action Parameter')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Chance Random Action')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'Episode')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x1152 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score over time: 0\n",
      "Final Q-Table Values\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.64</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E</td>\n",
       "      <td>0.61</td>\n",
       "      <td>0.73</td>\n",
       "      <td>0.74</td>\n",
       "      <td>0.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>E</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>E</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.79</td>\n",
       "      <td>0.91</td>\n",
       "      <td>0.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>G</td>\n",
       "      <td>1.05</td>\n",
       "      <td>1.02</td>\n",
       "      <td>0.96</td>\n",
       "      <td>0.89</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      0     1     2     3     4\n",
       "0  0.64  0.70  0.72  0.72  0.66\n",
       "1     E  0.61  0.73  0.74  0.75\n",
       "2     E  0.66  0.80  0.80  0.80\n",
       "3     E  0.72  0.79  0.91  0.87\n",
       "4     G  1.05  1.02  0.96  0.89"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>➡</td>\n",
       "      <td>➡</td>\n",
       "      <td>➡</td>\n",
       "      <td>⬇</td>\n",
       "      <td>⬇</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E</td>\n",
       "      <td>➡</td>\n",
       "      <td>➡</td>\n",
       "      <td>⬇</td>\n",
       "      <td>⬇</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>E</td>\n",
       "      <td>➡</td>\n",
       "      <td>➡</td>\n",
       "      <td>⬇</td>\n",
       "      <td>⬅</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>E</td>\n",
       "      <td>⬆</td>\n",
       "      <td>➡</td>\n",
       "      <td>⬇</td>\n",
       "      <td>⬅</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>G</td>\n",
       "      <td>⬅</td>\n",
       "      <td>⬅</td>\n",
       "      <td>⬅</td>\n",
       "      <td>⬅</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1  2  3  4\n",
       "0  ➡  ➡  ➡  ⬇  ⬇\n",
       "1  E  ➡  ➡  ⬇  ⬇\n",
       "2  E  ➡  ➡  ⬇  ⬅\n",
       "3  E  ⬆  ➡  ⬇  ⬅\n",
       "4  G  ⬅  ⬅  ⬅  ⬅"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Qt_list = []\n",
    "rt_list = []\n",
    "\n",
    "for t in range(trials):\n",
    "\n",
    "    # Initialize history memory\n",
    "    step_list = []\n",
    "    reward_list = []\n",
    "    loss_list = []\n",
    "    e_list = []\n",
    "\n",
    "    #Model\n",
    "    initializer = initializers.random_uniform(0, 0.1, seed=1)\n",
    "    model = Sequential([Dense(4, input_dim=gw.M*gw.N, \n",
    "                        kernel_initializer=initializer,\n",
    "                        use_bias=False\n",
    "                                 )])\n",
    "    model.compile(loss=loss, optimizer=Adam(lr=learning_rate))\n",
    "\n",
    "\n",
    "    #Transform into one-hot vector\n",
    "    def OH(l, x):\n",
    "        return np.identity(l)[x].reshape(1,l)\n",
    "\n",
    "\n",
    "    for i in trange(episodes):\n",
    "        # Initialize\n",
    "        state = gw.reset()\n",
    "        reward_all = 0 #Reward counter\n",
    "        done = False\n",
    "        s = 0 #Step counter\n",
    "        l = 0 #Loss\n",
    "\n",
    "        for s in range(steps):\n",
    "            # Choose action randomly or through agent model\n",
    "            Q = model.predict(OH(gw.M*gw.N, state_tup_to_num(state)), batch_size=1)\n",
    "            if np.random.rand(1) < e:\n",
    "                action = np.random.choice([0,1,2,3])\n",
    "            else:\n",
    "                action = np.argmax(Q)\n",
    "\n",
    "            # Take action and obtain new state and reward    \n",
    "            new_state, reward, done = gw.step(action)\n",
    "\n",
    "            # Adjust reward if done without reaching end\n",
    "            #if done and reward == 0.0: reward = -1\n",
    "\n",
    "            # Find max-Q for future state\n",
    "            Q1 = model.predict(OH(gw.M*gw.N, state_tup_to_num(new_state)), batch_size=1)\n",
    "            maxQ1 = np.max(Q1)\n",
    "\n",
    "            # Bellman Equation\n",
    "            # Update target for training by adding reward for action and discounted max next state Q-value\n",
    "            targetQ = Q\n",
    "            targetQ[0, action] = reward + (gamma * maxQ1)\n",
    "    #         if new_state in gw.error_states:\n",
    "    #             targetQ[0, action] -= 1\n",
    "\n",
    "\n",
    "            # Train on target Q value\n",
    "            history = model.fit(OH(gw.M*gw.N, state_tup_to_num(state)), targetQ, verbose=False, batch_size=1)\n",
    "\n",
    "            # Update history and set current state\n",
    "            l += history.history['loss'][0]\n",
    "            reward_all += reward\n",
    "            state = new_state\n",
    "\n",
    "            if done == True:\n",
    "                # Reduce e if current episode is successful\n",
    "                if reward > 0: \n",
    "                    e = 1./((i/50) + 10)\n",
    "                break\n",
    "\n",
    "        # Update history\n",
    "        step_list.append(s)\n",
    "        reward_list.append(reward_all)\n",
    "        loss_list.append(l/s)\n",
    "        e_list.append(e)\n",
    "    print('\\nSuccessful episodes: {}'.format(np.sum(np.array(reward_list)>0.0)/float(episodes)))\n",
    "\n",
    "    window = int(episodes/10)\n",
    "\n",
    "    plt.figure(figsize=[9,16])\n",
    "    plt.subplot(411)\n",
    "    plt.plot(pd.Series(step_list).rolling(window).mean())\n",
    "    plt.title('Step Moving Average ({}-episode window)'.format(window))\n",
    "    plt.ylabel('Moves')\n",
    "    plt.xlabel('Episode')\n",
    "\n",
    "    plt.subplot(412)\n",
    "    plt.plot(pd.Series(reward_list).rolling(window).mean())\n",
    "    plt.title('Reward Moving Average ({}-episode window)'.format(window))\n",
    "    plt.ylabel('Reward')\n",
    "    plt.xlabel('Episode')\n",
    "\n",
    "    plt.subplot(413)\n",
    "    plt.plot(pd.Series(loss_list).rolling(window).mean())\n",
    "    plt.title('Loss Moving Average ({}-episode window)'.format(window))\n",
    "    plt.ylabel('Loss')\n",
    "    plt.xlabel('Episode')\n",
    "\n",
    "    plt.subplot(414)\n",
    "    plt.plot(e_list)\n",
    "    plt.title('Random Action Parameter')\n",
    "    plt.ylabel('Chance Random Action')\n",
    "    plt.xlabel('Episode')\n",
    "\n",
    "    plt.tight_layout(pad=2)\n",
    "    plt.show()\n",
    "    \n",
    "    r_av = sum(reward_list)/episodes\n",
    "    print \"Score over time: \" +  str(r_av)\n",
    "    rt_list.append(r_av)\n",
    "    print \"Final Q-Table Values\"\n",
    "    Q = np.zeros((gw.M,gw.N,4))\n",
    "\n",
    "    for x in range(gw.M):\n",
    "        for y in range(gw.N):\n",
    "            Q[x,y,:] = model.predict(OH(gw.M*gw.N, state_tup_to_num((x, y))), batch_size=1)\n",
    "\n",
    "    v = np.max(Q, axis = 2)\n",
    "    gw.print_v(v)\n",
    "\n",
    "    pi = np.argmax(Q, axis = 2)\n",
    "    gw.print_policy(pi)\n",
    "    \n",
    "    Qt_list.append(Q)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}