side_by_side_plot.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import matplotlib.patches as patches\n",
    "import pickle\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from IPython.display import clear_output\n",
    "from scipy.stats import norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 518.4x374.4 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(7.2, 5.2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "param = '1'\n",
    "plt.rc('font', family='serif')\n",
    "plt.rcParams['pdf.fonttype'] = 42\n",
    "plt.rc('xtick', labelsize='x-small')\n",
    "plt.rc('ytick', labelsize='x-small')\n",
    "plt.rcParams.update({'font.size': 8})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "epLen = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  2.420000            500.0   1509.0\n",
      "1609.0   291319.555556  180.944444  2.426123            500.0   1609.0\n",
      "1709.0   317235.666667  185.518519  2.426123            500.0   1709.0\n",
      "1809.0   343899.000000  190.000000  2.426123            500.0   1809.0\n",
      "1909.0   372449.000000  195.000000  2.426123            500.0   1909.0\n",
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  3.445384       243.408163   1509.0\n",
      "1609.0   291319.555556  180.944444  3.514366       247.333333   1609.0\n",
      "1709.0   317235.666667  185.518519  3.500517       249.444444   1709.0\n",
      "1809.0   343899.000000  190.000000  3.485556       252.894737   1809.0\n",
      "1909.0   372449.000000  195.000000  3.486630       259.333333   1909.0\n",
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  3.045432       226.632653   1509.0\n",
      "1609.0   291319.555556  180.944444  3.147749       228.583333   1609.0\n",
      "1709.0   317235.666667  185.518519  2.995700       228.888889   1709.0\n",
      "1809.0   343899.000000  190.000000  3.071213       230.157895   1809.0\n",
      "1909.0   372449.000000  195.000000  3.254569       233.666667   1909.0\n",
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  4.866168            500.0   1509.0\n",
      "1609.0   291319.555556  180.944444  4.879017            500.0   1609.0\n",
      "1709.0   317235.666667  185.518519  4.887259            500.0   1709.0\n",
      "1809.0   343899.000000  190.000000  4.884363            500.0   1809.0\n",
      "1909.0   372449.000000  195.000000  4.893938            500.0   1909.0\n",
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  4.904498       194.918367   1509.0\n",
      "1609.0   291319.555556  180.944444  4.917203       206.500000   1609.0\n",
      "1709.0   317235.666667  185.518519  4.916637       214.444444   1709.0\n",
      "1809.0   343899.000000  190.000000  4.913770       220.052632   1809.0\n",
      "1909.0   372449.000000  195.000000  4.915860       227.000000   1909.0\n",
      "            Unnamed: 0   iteration  epReward  Number of Balls  episode\n",
      "episode                                                               \n",
      "1509.0   264249.000000  175.000000  4.864737       177.346939   1509.0\n",
      "1609.0   291319.555556  180.944444  4.865232       187.083333   1609.0\n",
      "1709.0   317235.666667  185.518519  4.876819       197.222222   1709.0\n",
      "1809.0   343899.000000  190.000000  4.879908       205.684211   1809.0\n",
      "1909.0   372449.000000  195.000000  4.868631       207.333333   1909.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 417.6x252 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "problem = 'oil'\n",
    "typ = 'laplace'\n",
    "param = '10'\n",
    "\n",
    "\n",
    "        \n",
    "        \n",
    "        \n",
    "name_adapt = './data/'+problem+'_'+typ+'_'+param+'_adaQL.csv'\n",
    "name_adapt_model_based = './data/'+problem+'_'+typ+'_'+param+'_adaMB_Full.csv'\n",
    "# name_adapt_flag_model_based = './data/oil_'+problem+'_'+'adapt_flag_'+param+'.csv'\n",
    "name_net = './data/'+problem+'_'+typ+'_'+param+'_epsQL.csv'\n",
    "name_enet_model = './data/'+problem+'_'+typ+'_'+param+'_epsMB_One.csv'\n",
    "\n",
    "dt_adapt_model = pd.read_csv(name_adapt_model_based).groupby(['episode']).mean()\n",
    "dt_net_model = pd.read_csv(name_enet_model).groupby(['episode']).mean()\n",
    "dt_adapt = pd.read_csv(name_adapt).groupby(['episode']).mean()\n",
    "dt_net = pd.read_csv(name_net).groupby(['episode']).mean()\n",
    "\n",
    "max_val = max(dt_adapt.tail(1)['epReward'].iloc[0], dt_adapt_model.tail(1)['epReward'].iloc[0], dt_net.tail(1)['epReward'].iloc[0], dt_net_model.tail(1)['epReward'].iloc[0])\n",
    "min_val = min(dt_adapt.head(1)['epReward'].iloc[0], dt_adapt_model.head(1)['epReward'].iloc[0], dt_net.head(1)['epReward'].iloc[0], dt_net_model.head(1)['epReward'].iloc[0])\n",
    "\n",
    "dt_adapt_model['episode'] = dt_adapt_model.index.values\n",
    "dt_adapt_model = dt_adapt_model.iloc[::10, :]\n",
    "dt_net_model['episode'] = dt_net_model.index.values\n",
    "dt_net_model = dt_net_model.iloc[::10, :]\n",
    "dt_adapt['episode'] = dt_adapt.index.values\n",
    "dt_net['episode'] = dt_net.index.values\n",
    "dt_net = dt_net.iloc[::10, :]\n",
    "dt_adapt = dt_adapt.iloc[::10, :]\n",
    "\n",
    "\n",
    "fig = plt.figure(figsize=(5.8, 3.5))\n",
    "plt.subplot(2,2,1)\n",
    "\n",
    "plt.plot(dt_adapt['episode'], dt_adapt['epReward'], label='AdaQL')\n",
    "plt.plot(dt_adapt_model['episode'], dt_adapt_model['epReward'], label='AdaMB', linestyle='--')\n",
    "# plt.plot(dt_adapt_model_flag['episode'], dt_adapt_model_flag['epReward'], label='ModelAdaptiveFlag', linestyle='--')\n",
    "plt.plot(dt_net_model['episode'], dt_net_model['epReward'], label='epsilonMB', linestyle='--')\n",
    "plt.plot(dt_net['episode'], dt_net['epReward'], label = 'epsilonQL', linestyle='--')\n",
    "\n",
    "plt.ylim(min_val - .2,max_val+.2)\n",
    "plt.xlabel('Episode')\n",
    "plt.ylabel('Observed Reward')\n",
    "# plt.legend()\n",
    "plt.title('Oil Problem with Laplace Rewards')\n",
    "\n",
    "print(dt_net.tail())\n",
    "print(dt_adapt.tail())\n",
    "print(dt_adapt_model.tail())\n",
    "\n",
    "problem = 'ambulance'\n",
    "typ = 'beta'\n",
    "param = '1'\n",
    "\n",
    "\n",
    "name_adapt = './data/'+problem+'_'+typ+'_'+param+'_adaQL.csv'\n",
    "name_adapt_model_based = './data/'+problem+'_'+typ+'_'+param+'_adaMB_Full.csv'\n",
    "# name_adapt_flag_model_based = './data/oil_'+problem+'_'+'adapt_flag_'+param+'.csv'\n",
    "name_net = './data/'+problem+'_'+typ+'_'+param+'_epsQL.csv'\n",
    "name_enet_model = './data/'+problem+'_'+typ+'_'+param+'_epsMB_One.csv'\n",
    "\n",
    "dt_adapt_model = pd.read_csv(name_adapt_model_based).groupby(['episode']).mean()\n",
    "dt_net_model = pd.read_csv(name_enet_model).groupby(['episode']).mean()\n",
    "dt_adapt = pd.read_csv(name_adapt).groupby(['episode']).mean()\n",
    "dt_net = pd.read_csv(name_net).groupby(['episode']).mean()\n",
    "\n",
    "max_val = max(dt_adapt.tail(1)['epReward'].iloc[0], dt_adapt_model.tail(1)['epReward'].iloc[0], dt_net.tail(1)['epReward'].iloc[0], dt_net_model.tail(1)['epReward'].iloc[0])\n",
    "min_val = min(dt_adapt.head(1)['epReward'].iloc[0], dt_adapt_model.head(1)['epReward'].iloc[0], dt_net.head(1)['epReward'].iloc[0], dt_net_model.head(1)['epReward'].iloc[0])\n",
    "\n",
    "dt_adapt_model['episode'] = dt_adapt_model.index.values\n",
    "dt_adapt_model = dt_adapt_model.iloc[::10, :]\n",
    "dt_net_model['episode'] = dt_net_model.index.values\n",
    "dt_net_model = dt_net_model.iloc[::10, :]\n",
    "dt_adapt['episode'] = dt_adapt.index.values\n",
    "dt_net['episode'] = dt_net.index.values\n",
    "dt_net = dt_net.iloc[::10, :]\n",
    "dt_adapt = dt_adapt.iloc[::10, :]\n",
    "\n",
    "plt.subplot(2,2,2)\n",
    "\n",
    "plt.plot(dt_adapt['episode'], dt_adapt['epReward'], label='AdaQL')\n",
    "plt.plot(dt_adapt_model['episode'], dt_adapt_model['epReward'], label='AdaMB', linestyle='--')\n",
    "# plt.plot(dt_adapt_model_flag['episode'], dt_adapt_model_flag['epReward'], label='ModelAdaptiveFlag', linestyle='--')\n",
    "plt.plot(dt_net_model['episode'], dt_net_model['epReward'], label='epsilonMB', linestyle='--')\n",
    "plt.plot(dt_net['episode'], dt_net['epReward'], label = 'epsilonQL', linestyle='--')\n",
    "max_val = max(dt_adapt.tail(1)['epReward'].iloc[0], dt_adapt_model.tail(1)['epReward'].iloc[0], dt_net.tail(1)['epReward'].iloc[0], dt_net_model.tail(1)['epReward'].iloc[0])\n",
    "\n",
    "\n",
    "plt.ylim(min_val-.2,max_val+.2)\n",
    "plt.xlabel('Episode')\n",
    "plt.ylabel('Observed Reward')\n",
    "plt.legend()\n",
    "plt.title('Ambulance Problem with Beta Arrivals')\n",
    "\n",
    "print(dt_net.tail())\n",
    "print(dt_adapt.tail())\n",
    "print(dt_adapt_model.tail())\n",
    "\n",
    "plt.tight_layout()\n",
    "fig.savefig('combine_plot_paper.pdf', bbox_inches = 'tight',\n",
    "     pad_inches = 0.01, dpi=900)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}