experiment-notebooks/grid-road-env-approach-comparison-experiment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Approach Comparison Experiment\n",
    "\n",
    "In this notebook we compare our approach against the equivalent baseline MORS approach where the proposed governance variants are evaluated against 3x3 and 5x5 environment configurations for both fixed and random orientations.\n",
    "In the MORS approach implementation additional rewards are given for following the correct direction towards the goal and exchanging packages to achieve cooperation between agents.\n",
    "The _algorithm performance experiment_ explains all the existing modules present in this notebook in complete detail.\n",
    "The below _TL;DR Section Descriptions_ elaborates on the notebook implementation structure in short.\n",
    "\n",
    "### TL;DR Section Descriptions\n",
    "\n",
    "* [`Notebook Prerequisite Setup`](#notebook-prerequisite-setup): Installing `stable-baselines3` and related packages, plus defining all package imports\n",
    "* [`Gym Environment Implementation`](#gym-environment-implementation): Implementing grid road environment variants, plus setting up Gym Monitor logger\n",
    "* [`Governance Wrapper Layer Implementation`](#governance-wrapper-layer-implementation): Implementing reward shaping governance layer onto grid delivery environments\n",
    "* [`RL Algorithm Model Training`](#rl-algorithm-model-training): Training RL algorithms in simple CTCE setting for measuring approach efficacy, plus manually saving the log files locally"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id='notebook-prerequisite-setup'></a>\n",
    "#### 0. Notebook Prerequisite Setup\n",
    "\n",
    "`In this section we first install stable-baselines3 and related packages required for Gym Monitor logging. And, then we import environment implementation, RL model training, and convergence plot related matplotlib packages.`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "34ag2hg60oia",
    "outputId": "57fb7853-975b-4358-a859-4d6b9cc30ca7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
      "Collecting stable-baselines3\n",
      "  Downloading stable_baselines3-1.6.0-py3-none-any.whl (177 kB)\n",
      "\u001b[K     |████████████████████████████████| 177 kB 4.0 MB/s \n",
      "\u001b[?25hCollecting gym==0.21\n",
      "  Downloading gym-0.21.0.tar.gz (1.5 MB)\n",
      "\u001b[K     |████████████████████████████████| 1.5 MB 53.8 MB/s \n",
      "\u001b[?25hRequirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from stable-baselines3) (1.3.5)\n",
      "Requirement already satisfied: torch>=1.11 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3) (1.12.1+cu113)\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from stable-baselines3) (1.21.6)\n",
      "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from stable-baselines3) (3.2.2)\n",
      "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from stable-baselines3) (1.5.0)\n",
      "Requirement already satisfied: importlib_metadata>=4.8.1 in /usr/local/lib/python3.7/dist-packages (from gym==0.21->stable-baselines3) (4.12.0)\n",
      "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib_metadata>=4.8.1->gym==0.21->stable-baselines3) (3.8.1)\n",
      "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib_metadata>=4.8.1->gym==0.21->stable-baselines3) (4.1.1)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3) (1.4.4)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3) (3.0.9)\n",
      "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3) (0.11.0)\n",
      "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3) (2.8.2)\n",
      "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->stable-baselines3) (1.15.0)\n",
      "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->stable-baselines3) (2022.2.1)\n",
      "Building wheels for collected packages: gym\n",
      "  Building wheel for gym (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Created wheel for gym: filename=gym-0.21.0-py3-none-any.whl size=1616826 sha256=cac26a35cc701a410300a24fafbe8430180857a3a95837a1179172b381f53759\n",
      "  Stored in directory: /root/.cache/pip/wheels/76/ee/9c/36bfe3e079df99acf5ae57f4e3464ff2771b34447d6d2f2148\n",
      "Successfully built gym\n",
      "Installing collected packages: gym, stable-baselines3\n",
      "  Attempting uninstall: gym\n",
      "    Found existing installation: gym 0.25.2\n",
      "    Uninstalling gym-0.25.2:\n",
      "      Successfully uninstalled gym-0.25.2\n",
      "Successfully installed gym-0.21.0 stable-baselines3-1.6.0\n"
     ]
    }
   ],
   "source": [
    "# installing the latest stable-baselines3 library\n",
    "# !pip install stable-baselines3\n",
    "!pip install stable-baselines3\n",
    "# ignoring the restart runtime instruction and continue with the cell execution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "wJlrYdxwhBwb",
    "outputId": "6bafedea-2816-485a-ce5b-43c6603df8db"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
      "Requirement already satisfied: stable-baselines3[extra] in /usr/local/lib/python3.7/dist-packages (1.6.0)\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.21.6)\n",
      "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.5.0)\n",
      "Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.3.5)\n",
      "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (3.2.2)\n",
      "Requirement already satisfied: gym==0.21 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (0.21.0)\n",
      "Requirement already satisfied: torch>=1.11 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.12.1+cu113)\n",
      "Collecting autorom[accept-rom-license]~=0.4.2\n",
      "  Downloading AutoROM-0.4.2-py3-none-any.whl (16 kB)\n",
      "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (4.6.0.66)\n",
      "Collecting ale-py==0.7.4\n",
      "  Downloading ale_py-0.7.4-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.6 MB)\n",
      "\u001b[K     |████████████████████████████████| 1.6 MB 4.0 MB/s \n",
      "\u001b[?25hRequirement already satisfied: pillow in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (7.1.2)\n",
      "Requirement already satisfied: tensorboard>=2.2.0 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (2.8.0)\n",
      "Collecting protobuf~=3.19.0\n",
      "  Downloading protobuf-3.19.5-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB)\n",
      "\u001b[K     |████████████████████████████████| 1.1 MB 44.7 MB/s \n",
      "\u001b[?25hRequirement already satisfied: psutil in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (5.4.8)\n",
      "Requirement already satisfied: importlib-resources in /usr/local/lib/python3.7/dist-packages (from ale-py==0.7.4->stable-baselines3[extra]) (5.9.0)\n",
      "Requirement already satisfied: importlib-metadata>=4.10.0 in /usr/local/lib/python3.7/dist-packages (from ale-py==0.7.4->stable-baselines3[extra]) (4.12.0)\n",
      "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2.23.0)\n",
      "Requirement already satisfied: click in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (7.1.2)\n",
      "Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (4.64.1)\n",
      "Collecting AutoROM.accept-rom-license\n",
      "  Downloading AutoROM.accept-rom-license-0.4.2.tar.gz (9.8 kB)\n",
      "  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
      "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
      "    Preparing wheel metadata ... \u001b[?25l\u001b[?25hdone\n",
      "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.10.0->ale-py==0.7.4->stable-baselines3[extra]) (4.1.1)\n",
      "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.10.0->ale-py==0.7.4->stable-baselines3[extra]) (3.8.1)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.35.0)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.8.1)\n",
      "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.6.1)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.0.1)\n",
      "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.37.1)\n",
      "Requirement already satisfied: grpcio>=1.24.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.48.1)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (57.4.0)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.6)\n",
      "Requirement already satisfied: absl-py>=0.4 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.2.0)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (3.4.1)\n",
      "Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (1.15.0)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.2.8)\n",
      "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.2.4)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.9)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (1.3.1)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.7/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.8)\n",
      "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (1.24.3)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2.10)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2022.6.15)\n",
      "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (3.0.4)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.7/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (3.2.0)\n",
      "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (0.11.0)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (1.4.4)\n",
      "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (2.8.2)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (3.0.9)\n",
      "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->stable-baselines3[extra]) (2022.2.1)\n",
      "Building wheels for collected packages: AutoROM.accept-rom-license\n",
      "  Building wheel for AutoROM.accept-rom-license (PEP 517) ... \u001b[?25l\u001b[?25hdone\n",
      "  Created wheel for AutoROM.accept-rom-license: filename=AutoROM.accept_rom_license-0.4.2-py3-none-any.whl size=441027 sha256=f8248a2e7c323ac93aabe4f8091ee6f2839e7edb99d9308d53cff89d25e797d5\n",
      "  Stored in directory: /root/.cache/pip/wheels/87/67/2e/6147e7912fe37f5408b80d07527dab807c1d25f5c403a9538a\n",
      "Successfully built AutoROM.accept-rom-license\n",
      "Installing collected packages: protobuf, AutoROM.accept-rom-license, autorom, ale-py\n",
      "  Attempting uninstall: protobuf\n",
      "    Found existing installation: protobuf 3.17.3\n",
      "    Uninstalling protobuf-3.17.3:\n",
      "      Successfully uninstalled protobuf-3.17.3\n",
      "Successfully installed AutoROM.accept-rom-license-0.4.2 ale-py-0.7.4 autorom-0.4.2 protobuf-3.19.5\n"
     ]
    },
    {
     "data": {
      "application/vnd.colab-display-data+json": {
       "pip_warning": {
        "packages": [
         "google"
        ]
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reading package lists... Done\n",
      "Building dependency tree       \n",
      "Reading state information... Done\n",
      "cmake is already the newest version (3.10.2-1ubuntu2.18.04.2).\n",
      "The following package was automatically installed and is no longer required:\n",
      "  libnvidia-common-460\n",
      "Use 'apt autoremove' to remove it.\n",
      "Suggested packages:\n",
      "  swig-doc swig-examples swig3.0-examples swig3.0-doc\n",
      "The following NEW packages will be installed:\n",
      "  swig swig3.0\n",
      "0 upgraded, 2 newly installed, 0 to remove and 20 not upgraded.\n",
      "Need to get 1,100 kB of archives.\n",
      "After this operation, 5,822 kB of additional disk space will be used.\n",
      "Get:1 http://archive.ubuntu.com/ubuntu bionic/universe amd64 swig3.0 amd64 3.0.12-1 [1,094 kB]\n",
      "Get:2 http://archive.ubuntu.com/ubuntu bionic/universe amd64 swig amd64 3.0.12-1 [6,460 B]\n",
      "Fetched 1,100 kB in 0s (3,559 kB/s)\n",
      "Selecting previously unselected package swig3.0.\n",
      "(Reading database ... 155569 files and directories currently installed.)\n",
      "Preparing to unpack .../swig3.0_3.0.12-1_amd64.deb ...\n",
      "Unpacking swig3.0 (3.0.12-1) ...\n",
      "Selecting previously unselected package swig.\n",
      "Preparing to unpack .../swig_3.0.12-1_amd64.deb ...\n",
      "Unpacking swig (3.0.12-1) ...\n",
      "Setting up swig3.0 (3.0.12-1) ...\n",
      "Setting up swig (3.0.12-1) ...\n",
      "Processing triggers for man-db (2.8.3-2ubuntu0.1) ...\n",
      "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
      "Requirement already satisfied: stable-baselines3[extra] in /usr/local/lib/python3.7/dist-packages (1.6.0)\n",
      "Collecting box2d\n",
      "  Downloading Box2D-2.3.10-cp37-cp37m-manylinux1_x86_64.whl (1.3 MB)\n",
      "\u001b[K     |████████████████████████████████| 1.3 MB 4.1 MB/s \n",
      "\u001b[?25hCollecting box2d-kengz\n",
      "  Downloading Box2D-kengz-2.3.3.tar.gz (425 kB)\n",
      "\u001b[K     |████████████████████████████████| 425 kB 64.2 MB/s \n",
      "\u001b[?25hRequirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.5.0)\n",
      "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (3.2.2)\n",
      "Requirement already satisfied: torch>=1.11 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.12.1+cu113)\n",
      "Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.3.5)\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.21.6)\n",
      "Requirement already satisfied: gym==0.21 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (0.21.0)\n",
      "Requirement already satisfied: protobuf~=3.19.0 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (3.19.5)\n",
      "Requirement already satisfied: autorom[accept-rom-license]~=0.4.2 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (0.4.2)\n",
      "Requirement already satisfied: psutil in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (5.4.8)\n",
      "Requirement already satisfied: pillow in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (7.1.2)\n",
      "Requirement already satisfied: tensorboard>=2.2.0 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (2.8.0)\n",
      "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (4.6.0.66)\n",
      "Requirement already satisfied: ale-py==0.7.4 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (0.7.4)\n",
      "Requirement already satisfied: importlib-resources in /usr/local/lib/python3.7/dist-packages (from ale-py==0.7.4->stable-baselines3[extra]) (5.9.0)\n",
      "Requirement already satisfied: importlib-metadata>=4.10.0 in /usr/local/lib/python3.7/dist-packages (from ale-py==0.7.4->stable-baselines3[extra]) (4.12.0)\n",
      "Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (4.64.1)\n",
      "Requirement already satisfied: click in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (7.1.2)\n",
      "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2.23.0)\n",
      "Requirement already satisfied: AutoROM.accept-rom-license in /usr/local/lib/python3.7/dist-packages (from autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (0.4.2)\n",
      "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.10.0->ale-py==0.7.4->stable-baselines3[extra]) (4.1.1)\n",
      "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.10.0->ale-py==0.7.4->stable-baselines3[extra]) (3.8.1)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (3.4.1)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.35.0)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.0.1)\n",
      "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.6.1)\n",
      "Requirement already satisfied: absl-py>=0.4 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.2.0)\n",
      "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.37.1)\n",
      "Requirement already satisfied: grpcio>=1.24.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.48.1)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.6)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.8.1)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (57.4.0)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.2.8)\n",
      "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.2.4)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.9)\n",
      "Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (1.15.0)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (1.3.1)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.7/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.8)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2022.6.15)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (2.10)\n",
      "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (3.0.4)\n",
      "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->autorom[accept-rom-license]~=0.4.2->stable-baselines3[extra]) (1.24.3)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.7/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (3.2.0)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (1.4.4)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (3.0.9)\n",
      "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (0.11.0)\n",
      "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (2.8.2)\n",
      "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->stable-baselines3[extra]) (2022.2.1)\n",
      "Building wheels for collected packages: box2d-kengz\n",
      "  Building wheel for box2d-kengz (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Created wheel for box2d-kengz: filename=Box2D_kengz-2.3.3-cp37-cp37m-linux_x86_64.whl size=2034345 sha256=96ff078b3f3777ddae37463e13ffb26c771b8a1129a5a1e87253623bf5addf06\n",
      "  Stored in directory: /root/.cache/pip/wheels/50/6d/6a/6ff76731fd9e8efbd1cdc6111e98b2dd0f1872184d7c28939c\n",
      "Successfully built box2d-kengz\n",
      "Installing collected packages: box2d-kengz, box2d\n",
      "Successfully installed box2d-2.3.10 box2d-kengz-2.3.3\n"
     ]
    }
   ],
   "source": [
    "# installing additional stable-baselines3 packages for plotting the agent learning outputs\n",
    "# !pip install stable-baselines3\n",
    "!pip install stable-baselines3[extra]\n",
    "!apt install swig cmake\n",
    "!pip install stable-baselines3[extra] box2d box2d-kengz\n",
    "# ignoring the restart runtime instruction and continue with the cell execution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "W5RN83LAqWyB"
   },
   "outputs": [],
   "source": [
    "# OpenAI gym related import statements\n",
    "# building a simpler environment that works with stablebaselines\n",
    "import os\n",
    "import gym\n",
    "from gym import spaces\n",
    "import numpy as np\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "0DFgWujIh6L6"
   },
   "outputs": [],
   "source": [
    "# learning curve plotting realted package imports\n",
    "from gym.envs.registration import EnvSpec\n",
    "from stable_baselines3.common.monitor import Monitor\n",
    "from stable_baselines3.common.results_plotter import load_results, ts2xy\n",
    "from stable_baselines3.common.callbacks import BaseCallback"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "NgUDplIF7SIt"
   },
   "outputs": [],
   "source": [
    "# import statements for 3D function plotting with matplotlib\n",
    "from mpl_toolkits.mplot3d import axes3d\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import cm\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id='gym-environment-implementation'></a>\n",
    "#### 1. Gym Environment Implementation\n",
    "\n",
    "This section implements different OpenAI gym's environment related classes for the two agent package delivery environment.\n",
    "The inline comments explain the implementation functionalities in detail for all the described environment and logger classes.\n",
    "\n",
    "**Note:** The description for all the classes are explained in the `Gym Environment Implementation` section of _algorithm performance experiment_ notebook, refer _grid-road-env-algorithm-performance-experiment.ipynb_ for more details."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "4bfBZtEqin78"
   },
   "outputs": [],
   "source": [
    "# utility class for saving the best model function\n",
    "class SaveOnBestTrainingRewardCallback(BaseCallback):\n",
    "    \"\"\"\n",
    "    Callback for saving a model (the check is done every ``check_freq`` steps)\n",
    "    based on the training reward (in practice, we recommend using ``EvalCallback``).\n",
    "\n",
    "    :param check_freq: (int)\n",
    "    :param log_dir: (str) Path to the folder where the model will be saved.\n",
    "      It must contains the file created by the ``Monitor`` wrapper.\n",
    "    :param verbose: (int)\n",
    "    \"\"\"\n",
    "    def __init__(self, check_freq: int, log_dir: str, verbose=1):\n",
    "        super(SaveOnBestTrainingRewardCallback, self).__init__(verbose)\n",
    "        self.check_freq = check_freq\n",
    "        self.log_dir = log_dir\n",
    "        self.save_path = os.path.join(log_dir, 'best_model')\n",
    "        self.best_mean_reward = -np.inf\n",
    "\n",
    "    def _init_callback(self) -> None:\n",
    "        # create folder if needed\n",
    "        if self.save_path is not None:\n",
    "            os.makedirs(self.save_path, exist_ok=True)\n",
    "\n",
    "    def _on_step(self) -> bool:\n",
    "        if self.n_calls % self.check_freq == 0:\n",
    "          # retrieve training reward\n",
    "          x, y = ts2xy(load_results(self.log_dir), 'timesteps')\n",
    "          if len(x) > 0:\n",
    "              # mean training reward over the last 100 episodes\n",
    "              mean_reward = np.mean(y[-100:])\n",
    "              if self.verbose > 0:\n",
    "                print(f\"Num timesteps: {self.num_timesteps}\")\n",
    "                print(f\"Best mean reward: {self.best_mean_reward:.2f} - Last mean reward per episode: {mean_reward:.2f}\")\n",
    "              # new best model, you could save the agent here\n",
    "              if mean_reward > self.best_mean_reward:\n",
    "                  self.best_mean_reward = mean_reward\n",
    "                  # example for saving best model\n",
    "                  if self.verbose > 0:\n",
    "                    print(f\"Saving new best model to {self.save_path}.zip\")\n",
    "                  self.model.save(self.save_path)\n",
    "        return True\n",
    "\n",
    "# create log dir for the agent monitoring and result plotting\n",
    "log_dir = \"/tmp/gym/\"\n",
    "os.makedirs(log_dir, exist_ok=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "2pPscl6L7Wr8"
   },
   "outputs": [],
   "source": [
    "# the 3x3 road network environment doesn't rotate in its simple\n",
    "# setting implementation\n",
    "# implementation of the Simple 3x3 road network for MultiAgent RL\n",
    "class SizeThreeSimpleGridRoadEnv(gym.Env):\n",
    "    # Defining the Driving Agent with name and gas values plus package flag.\n",
    "    class DriverAgent():\n",
    "        def __init__(self, name, gas, package, picked):\n",
    "            self.name = name\n",
    "            self.gas = gas\n",
    "            self.package = package\n",
    "            self.picked = picked # flag, package has been picked once by the agent\n",
    "\n",
    "    def __init__(self):\n",
    "        # super(SizeThreeSimpleGridRoadEnv, self).__init__()\n",
    "        # fefining different possible world configurations\n",
    "        self.world_one = np.array([[1, 0, 0],\n",
    "                      [3, 0, 2],\n",
    "                      [0, 0, 4]])\n",
    "        # self.world_two = np.rot90(self.world_one)\n",
    "        # self.world_three = np.rot90(self.world_two)\n",
    "        # self.world_four = np.rot90(self.world_three)\n",
    "        # Eeen the initial world configuration is defined to be different upon\n",
    "        # environment instantiation\n",
    "        # prob = random.uniform(0, 1)\n",
    "        # default value assignment below\n",
    "        self.world = self.world_one\n",
    "        # if prob > 0.25 and prob <= 0.25:\n",
    "        #     self.world = self.world_two\n",
    "        # elif prob > 0.5 and prob <= 0.75:\n",
    "        #     self.world = self.world_three\n",
    "        # elif prob > 0.75 and prob <= 1:\n",
    "        #     self.world = self.world_four\n",
    "        self.world_start = self.world # this 'world_start', if reset() is called, never gets used\n",
    "        # adding five actions for the environment.\n",
    "        # 0: up, 1: right, 2: down, 3: left,; 4: stay/pass chance,\n",
    "        # 5: drop -> for agent learning simplicity eliminating 4 & 5 actions\n",
    "        # when agent reaches at package location it automatically picks up the package\n",
    "        self.action_space = spaces.Discrete(4)\n",
    "        shape_0 = np.size(self.world_start, 0)\n",
    "        shape_1 = np.size(self.world_start, 1)\n",
    "        self.observation_space = spaces.Box(low=0,\n",
    "                                            high=4,\n",
    "                                            shape=(shape_0 + 1, shape_1),\n",
    "                                            dtype=np.int16)\n",
    "        self.reward_range = (-10, 10)\n",
    "        self.current_episode = 0\n",
    "        self.success_episode = []\n",
    "        # defining the driver agents in the environment\n",
    "        self.agent_one = self.DriverAgent(1,3,0,False) # 3 integer value, when carrying package\n",
    "        self.agent_two = self.DriverAgent(2,8,0,False) # 3 integer value, when carrying package\n",
    "        self.spec = EnvSpec(\"SizeThreeSimpleGridRoadEnv-v0\")\n",
    "\n",
    "    def reset(self):\n",
    "        # sequential game like formulation, each player agent moves one step at a time\n",
    "        # but resource specific delays are also added so that second agent waits for\n",
    "        # first agent to finish it complete movement in the environment\n",
    "        self.agent_one = self.DriverAgent(1,3,0,False) # instantiating agent 1 again\n",
    "        self.agent_two = self.DriverAgent(2,4,0,False) # instantiating agent 2 again\n",
    "        self.current_player = self.agent_one\n",
    "        # 'P' means the game is playable, 'W' means delivered, 'L' means no delivery\n",
    "        self.state = 'P'\n",
    "        self.current_step = 0\n",
    "        self.max_step = 32 # agent can choose not move as an alternate choice\n",
    "        # selecting a world at random to function with\n",
    "        # even the initial world configuration should be different\n",
    "        # prob = random.uniform(0, 1)\n",
    "        # if prob > 0.25 and prob <= 0.25:\n",
    "        #     self.world_start = self.world_two\n",
    "        # elif prob > 0.5 and prob <= 0.75:\n",
    "        #     self.world_start = self.world_three\n",
    "        # elif prob > 0.75 and prob <= 1:\n",
    "        #     self.world_start = self.world_four\n",
    "        # elif prob < 0.25:\n",
    "        #     self.world_start = self.world_one\n",
    "        self.world_start = self.world_one    \n",
    "        self.world = np.copy(self.world_start) # the self.world can be different from intial world\n",
    "        # no exploration_prize and bonus_reward as per the design\n",
    "        return self._next_observation()\n",
    "    \n",
    "    def _next_observation(self):\n",
    "        obs = self.world\n",
    "        data_to_add = [0] * np.size(self.world, 1)\n",
    "        data_to_add[0] = self.current_player.name # adding current player's label in the observation\n",
    "        obs = np.append(obs, [data_to_add], axis=0)\n",
    "        # observation sample provided below for reference:\n",
    "        # last row, represents 'data_to_add' vector representing the current agent\n",
    "        # array([[1, 0, 0],\n",
    "        #         [3, 0, 2],\n",
    "        #         [0, 0, 4],\n",
    "        #         [1, 0, 0]])\n",
    "        return obs\n",
    "\n",
    "    def _take_action(self, action):\n",
    "        # agent's name is matched to the array entries for index identification\n",
    "        # 'current_player.name' should be updated alongside the array values\n",
    "        current_pos = np.where(self.world == self.current_player.name)\n",
    "        # the current agent must have gas in it\n",
    "        if self.current_player.gas > 0:\n",
    "            if action == 0:\n",
    "                next_pos = (current_pos[0] - 1, current_pos[1]) # agent moving upwards\n",
    "\n",
    "                if next_pos[0] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it.\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 1:\n",
    "                next_pos = (current_pos[0], current_pos[1] + 1)\n",
    "                limit = np.size(self.world, 1)\n",
    "\n",
    "                if next_pos[1] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 2:\n",
    "                next_pos = (current_pos[0] + 1, current_pos[1])\n",
    "                limit = np.size(self.world, 0)\n",
    "\n",
    "                if next_pos[0] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            elif action == 3:\n",
    "                next_pos = (current_pos[0], current_pos[1] - 1)\n",
    "\n",
    "                if next_pos[1] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] >= 0 and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            # removing functionality for action 4 and 5 for environment simplicity.\n",
    "            # newly added logic based on new possible actions.\n",
    "            # elif action == 4 and self.current_player.name != 1: # passing is allowed for agent 2 only\n",
    "            #     pass # corresponding agent selects to not move at their chance\n",
    "            \n",
    "            # elif action == 5: # if agent is over the package, it has to pick it up, environment cases encoded above\n",
    "            #     # agent can choose to drop the package, if it is loaded with it\n",
    "            #     # after, dropping the package the agent should dissappear\n",
    "            #     if self.current_player.package == 3:\n",
    "            #         if self.world[current_pos] == 0:\n",
    "            #             self.world[current_pos] = 3\n",
    "            #             # agent dissappears from the grid after this drop\n",
    "            #         elif self.world[current_pos] == 4: # added as extra case, functionally possibly won't be triggered\n",
    "            #             self.world[current_pos] = self.current_player.name\n",
    "            #             self.state = 'W'\n",
    "            \n",
    "        else:\n",
    "            # player 1's gas is supposed to go empty first\n",
    "            # therefore, upon having empty gas tank player should be allowed to\n",
    "            # drop the package in the environment and disappear from the location\n",
    "            if self.current_player.package == 3:\n",
    "                self.world[current_pos] = self.current_player.package\n",
    "                # self.current_player.picked = False # package is dropped by the agent\n",
    "                # agent dissappears from the grid after this drop\n",
    "            else:\n",
    "                self.world[current_pos] = 0 # if gas is finished, agent should dissappear\n",
    "\n",
    "        # if gas is empty for both agents, the episode should stop at that instant\n",
    "        if self.agent_one.gas == 0 and self.agent_two.gas == 0:\n",
    "            self.state = 'L'\n",
    "\n",
    "\n",
    "    def step(self, action):\n",
    "        self._take_action(action)\n",
    "        self.current_step += 1\n",
    "        # uncomment the below statement out, while debugging\n",
    "        # print(self.world) \n",
    "\n",
    "        if self.state == \"W\":\n",
    "            reward = 2.5\n",
    "            done = True\n",
    "        elif self.state == 'L':\n",
    "            reward = 0\n",
    "            done = True\n",
    "        elif self.state == 'P':\n",
    "            reward = 0 # sparse reward encoding, only rewarded when episode ends\n",
    "            done = False\n",
    "\n",
    "        if self.current_step >= self.max_step:\n",
    "            print(f'New episode number {self.current_episode + 1}')\n",
    "            done = True\n",
    "        \n",
    "        # removing functionality for alternate play for environment simplicity\n",
    "        # agents object used to identify agent properties\n",
    "        if self.current_player.name == 1 and self.current_step > 6: # general condition > (2*gas)\n",
    "            self.current_player = self.agent_two\n",
    "        elif self.current_player.name == 2:\n",
    "            self.current_player = self.agent_one\n",
    "\n",
    "        # adding functionality of sequential agent interaction for environment simplicity\n",
    "        # if self.current_player.name == 1 and self.current_player.gas == 0:\n",
    "        #     self.current_player = self.agent_two\n",
    "\n",
    "        if done:\n",
    "            self.render_episode(self.state)\n",
    "            # self.reset()\n",
    "            self.current_episode += 1\n",
    "        \n",
    "        # if done == True and self.state == 'L':\n",
    "        #     self.reset()\n",
    "\n",
    "        obs = self._next_observation()\n",
    "\n",
    "        return obs, reward, done, {'state': self.state}, self.current_player\n",
    "\n",
    "    def render_episode(self, win_or_lose):\n",
    "        # storing the rendered episodes in a file\n",
    "        self.success_episode.append(\n",
    "            'Success' if win_or_lose == 'W' else 'Failure')\n",
    "        file = open('render.txt', 'a')\n",
    "        file.write('----------------------------\\n')\n",
    "        file.write(f'Episode number {self.current_episode}\\n')\n",
    "        file.write(\n",
    "            f'{self.success_episode[-1]} in {self.current_step} steps\\n')\n",
    "        file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "oB5PJb-2LJHS"
   },
   "outputs": [],
   "source": [
    "# the 3x3 road network environment which rotate randomly in its simple\n",
    "# setting implementation\n",
    "# implementation of the Simple 3x3 road network for MultiAgent RL\n",
    "class SizeThreeGridRoadEnv(gym.Env):\n",
    "    # defining the Driving Agent with name and gas values plus package flag\n",
    "    class DriverAgent():\n",
    "        def __init__(self, name, gas, package, picked):\n",
    "            self.name = name\n",
    "            self.gas = gas\n",
    "            self.package = package\n",
    "            self.picked = picked # flag, package has been picked once by the agent\n",
    "\n",
    "    def __init__(self):\n",
    "        # super(SizeThreeSimpleGridRoadEnv, self).__init__()\n",
    "        # defining different possible world configurations\n",
    "        self.world_one = np.array([[1, 0, 0],\n",
    "                      [3, 0, 2],\n",
    "                      [0, 0, 4]])\n",
    "        self.world_two = np.rot90(self.world_one)\n",
    "        self.world_three = np.rot90(self.world_two)\n",
    "        self.world_four = np.rot90(self.world_three)\n",
    "        # even the initial world configuration is defined to be different upon\n",
    "        # environment instantiation\n",
    "        prob = random.uniform(0, 1)\n",
    "        # default value assignment below\n",
    "        self.world = self.world_one\n",
    "        if prob > 0.25 and prob <= 0.25:\n",
    "            self.world = self.world_two\n",
    "        elif prob > 0.5 and prob <= 0.75:\n",
    "            self.world = self.world_three\n",
    "        elif prob > 0.75 and prob <= 1:\n",
    "            self.world = self.world_four\n",
    "        self.world_start = self.world # this 'world_start', if reset() is called, never gets used\n",
    "        # adding five actions for the environment\n",
    "        # 0: up, 1: right, 2: down, 3: left,; 4: stay/pass chance,\n",
    "        # 5: drop -> for agent learning simplicity eliminating 4 & 5 actions\n",
    "        # when agent reaches at package location it automatically picks up the package\n",
    "        self.action_space = spaces.Discrete(4)\n",
    "        shape_0 = np.size(self.world_start, 0)\n",
    "        shape_1 = np.size(self.world_start, 1)\n",
    "        self.observation_space = spaces.Box(low=0,\n",
    "                                            high=4,\n",
    "                                            shape=(shape_0 + 1, shape_1),\n",
    "                                            dtype=np.int16)\n",
    "        self.reward_range = (-10, 10)\n",
    "        self.current_episode = 0\n",
    "        self.success_episode = []\n",
    "        # defining the driver agents in the environment\n",
    "        self.agent_one = self.DriverAgent(1,3,0,False) # 3 integer value, when carrying package\n",
    "        self.agent_two = self.DriverAgent(2,8,0,False) # 3 integer value, when carrying package\n",
    "        self.spec = EnvSpec(\"SizeThreeGridRoadEnv-v0\")\n",
    "\n",
    "    def reset(self):\n",
    "        # game like formulation, each player agent moves one step at a time\n",
    "        self.agent_one = self.DriverAgent(1,3,0,False) # instantiating agent 1 again\n",
    "        self.agent_two = self.DriverAgent(2,4,0,False) # instantiating agent 2 again\n",
    "        self.current_player = self.agent_one\n",
    "        # 'P' means the game is playable, 'W' means delivered, 'L' means no delivery\n",
    "        self.state = 'P'\n",
    "        self.current_step = 0\n",
    "        self.max_step = 32 # agent can choose not move as an alternate choice\n",
    "        # selecting a world at random to function with\n",
    "        # even the initial world configuration should be different\n",
    "        prob = random.uniform(0, 1)\n",
    "        if prob > 0.25 and prob <= 0.25:\n",
    "            self.world_start = self.world_two\n",
    "        elif prob > 0.5 and prob <= 0.75:\n",
    "            self.world_start = self.world_three\n",
    "        elif prob > 0.75 and prob <= 1:\n",
    "            self.world_start = self.world_four\n",
    "        elif prob < 0.25:\n",
    "            self.world_start = self.world_one\n",
    "        self.world_start = self.world_one\n",
    "        self.world = np.copy(self.world_start) # the self.world can be different from intial world\n",
    "        # no exploration_prize and bonus_reward as per my design\n",
    "        return self._next_observation()\n",
    "    \n",
    "    def _next_observation(self):\n",
    "        obs = self.world\n",
    "        data_to_add = [0] * np.size(self.world, 1)\n",
    "        data_to_add[0] = self.current_player.name # adding current player's label in the observation\n",
    "        obs = np.append(obs, [data_to_add], axis=0)\n",
    "        # observation Sample provided below for reference:\n",
    "        # last row, represents 'data_to_add' vector\n",
    "        # array([[1, 0, 0],\n",
    "        #         [3, 0, 2],\n",
    "        #         [0, 0, 4],\n",
    "        #         [1, 0, 0]])\n",
    "        return obs\n",
    "\n",
    "    def _take_action(self, action):\n",
    "        # agent's name is matched to the array entries for index identification\n",
    "        # 'current_player.name' should be updated alongside the array values\n",
    "        current_pos = np.where(self.world == self.current_player.name)\n",
    "        # the current agent must have gas in it\n",
    "        if self.current_player.gas > 0:\n",
    "            if action == 0:\n",
    "                next_pos = (current_pos[0] - 1, current_pos[1]) # agent moving upwards\n",
    "\n",
    "                if next_pos[0] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name # like 34 are already there, for example\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 1:\n",
    "                next_pos = (current_pos[0], current_pos[1] + 1)\n",
    "                limit = np.size(self.world, 1)\n",
    "\n",
    "                if next_pos[1] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name # like 34 are already there, for example\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 2:\n",
    "                next_pos = (current_pos[0] + 1, current_pos[1])\n",
    "                limit = np.size(self.world, 0)\n",
    "\n",
    "                if next_pos[0] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            elif action == 3:\n",
    "                next_pos = (current_pos[0], current_pos[1] - 1)\n",
    "\n",
    "                if next_pos[1] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] >= 0 and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.current_player.picked = True # package is picked once by the agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            # removing functionality for action 4 and 5 for environment simplicity\n",
    "            # newly added logic based on three new possible actions\n",
    "            # elif action == 4 and self.current_player.name != 1: # passing is allowed for agent 2 only\n",
    "            #     pass # corresponding agent selects to not move at their chance\n",
    "            \n",
    "            # elif action == 5: # if agent is over the package, it has to pick it up, environment cases encoded above\n",
    "            #     # agent can choose to drop the package, if it is loaded with it\n",
    "            #     # after, dropping the package the agent should dissappear\n",
    "            #     if self.current_player.package == 3:\n",
    "            #         if self.world[current_pos] == 0:\n",
    "            #             self.world[current_pos] = 3\n",
    "            #             # agent dissappears from the grid after this drop\n",
    "            #         elif self.world[current_pos] == 4: # added as extra case, functionally possibly won't be triggered\n",
    "            #             self.world[current_pos] = self.current_player.name\n",
    "            #             self.state = 'W'\n",
    "            \n",
    "        else:\n",
    "            # player 1's gas is supposed to go empty first\n",
    "            # therefore, upon having empty gas tank player should be allowed to\n",
    "            # drop the package in the environment and disappear from the location\n",
    "            if self.current_player.package == 3:\n",
    "                self.world[current_pos] = self.current_player.package\n",
    "                # self.current_player.picked = False # package is dropped by the agent\n",
    "                # agent dissappears from the grid after this drop\n",
    "            else:\n",
    "                self.world[current_pos] = 0 # if gas is finished, agent should dissappear\n",
    "\n",
    "        # if gas is empty for both agents, the episode should stop at that instant\n",
    "        if self.agent_one.gas == 0 and self.agent_two.gas == 0:\n",
    "            self.state = 'L'\n",
    "\n",
    "\n",
    "    def step(self, action):\n",
    "        self._take_action(action)\n",
    "        self.current_step += 1\n",
    "        # uncomment the below statement out, while debugging\n",
    "        # print(self.world) \n",
    "\n",
    "        if self.state == \"W\":\n",
    "            reward = 2.5\n",
    "            done = True\n",
    "        elif self.state == 'L':\n",
    "            reward = 0\n",
    "            done = True\n",
    "        elif self.state == 'P':\n",
    "            reward = 0 # sparse reward encoding, only rewarded when episode ends\n",
    "            done = False\n",
    "\n",
    "        if self.current_step >= self.max_step:\n",
    "            print(f'New episode number {self.current_episode + 1}')\n",
    "            done = True\n",
    "        \n",
    "        # removing functionality for alternate play for environment simplicity\n",
    "        # agents object used to identify agent properties\n",
    "        if self.current_player.name == 1 and self.current_step > 6: # general condition > (2*gas)\n",
    "            self.current_player = self.agent_two\n",
    "        elif self.current_player.name == 2:\n",
    "            self.current_player = self.agent_one\n",
    "\n",
    "        # adding functionality of sequential agent interaction for environment simplicity\n",
    "        # if self.current_player.name == 1 and self.current_player.gas == 0:\n",
    "        #     self.current_player = self.agent_two\n",
    "\n",
    "        if done:\n",
    "            self.render_episode(self.state)\n",
    "            # self.reset()\n",
    "            self.current_episode += 1\n",
    "        \n",
    "        # if done == True and self.state == 'L':\n",
    "        #     self.reset()\n",
    "\n",
    "        obs = self._next_observation()\n",
    "\n",
    "        return obs, reward, done, {'state': self.state}, self.current_player\n",
    "\n",
    "    def render_episode(self, win_or_lose):\n",
    "        # storing the rendered episodes in a file\n",
    "        self.success_episode.append(\n",
    "            'Success' if win_or_lose == 'W' else 'Failure')\n",
    "        file = open('render.txt', 'a')\n",
    "        file.write('----------------------------\\n')\n",
    "        file.write(f'Episode number {self.current_episode}\\n')\n",
    "        file.write(\n",
    "            f'{self.success_episode[-1]} in {self.current_step} steps\\n')\n",
    "        file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "dybP2J_wLJOY"
   },
   "outputs": [],
   "source": [
    "# the NxN road network environment doesn't rotate in its simple\n",
    "# setting implementation\n",
    "# implementation of the Simple NxN road network for MultiAgent RL\n",
    "class SimpleGridRoadEnv(gym.Env):\n",
    "    # defining the Driving Agent with name and gas values plus package flag\n",
    "    class DriverAgent():\n",
    "        def __init__(self, name, gas, package):\n",
    "            self.name = name\n",
    "            self.gas = gas\n",
    "            self.package = package\n",
    "\n",
    "    def __init__(self, size=5, gas=7):\n",
    "        # size: grid size, gas: values restricted between (n_min: trunc(n/2), n_max: 2*(n-1)-1)\n",
    "        # converting even size to nearest odd number\n",
    "        self.size = size\n",
    "        if size % 2 == 0:\n",
    "            self.size = self.size+1\n",
    "        # defining different possible world configurations\n",
    "        self.world_one = np.zeros((self.size,self.size), dtype=int)\n",
    "        # setting up env values for the numpy array\n",
    "        self.world_one[0][0]=1\n",
    "        self.world_one[self.size-1][self.size-1]=4\n",
    "        self.world_one[self.size//2][self.size-1]=2\n",
    "        self.world_one[self.size//2][0]=3\n",
    "        self.world = self.world_one\n",
    "        # if prob > 0.25 and prob <= 0.25:\n",
    "        #     self.world = self.world_two\n",
    "        # elif prob > 0.5 and prob <= 0.75:\n",
    "        #     self.world = self.world_three\n",
    "        # elif prob > 0.75 and prob <= 1:\n",
    "        #     self.world = self.world_four\n",
    "        self.world_start = self.world # this 'world_start', if reset() is called, never gets used\n",
    "        # Adding five actions for the environment.\n",
    "        # 0: up, 1: right, 2: down, 3: left, 4: stay/pass chance, 5: drop & dissappear\n",
    "        # when agent reaches at package location it automatically picks up the package\n",
    "        self.action_space = spaces.Discrete(4)\n",
    "        shape_0 = np.size(self.world_start, 0)\n",
    "        shape_1 = np.size(self.world_start, 1)\n",
    "        self.observation_space = spaces.Box(low=0,\n",
    "                                            high=4,\n",
    "                                            shape=(shape_0 + 1, shape_1),\n",
    "                                            dtype=np.int16)\n",
    "        self.reward_range = (-10, 10)\n",
    "        self.current_episode = 0\n",
    "        self.success_episode = []\n",
    "        # defining the driver agents in the environment\n",
    "        self.gas = gas\n",
    "        # truncating the gas to the desirable range for experimentation\n",
    "        if (self.gas < self.size//2) or (self.gas > (2*(self.size-1)-1)):\n",
    "            self.gas = random.randint(self.size//2, 2*(self.size-1))\n",
    "        self.agent_one = self.DriverAgent(1,self.gas,0) # 3 integer value, when carrying package\n",
    "        self.agent_two = self.DriverAgent(2,int(self.gas*3),0) # 3 integer value, when carrying package\n",
    "        self.spec = EnvSpec(\"SimpleGridRoadEnv-v0\")\n",
    "\n",
    "    def reset(self):\n",
    "        # game like formulation, each player agent moves one step at a time\n",
    "        self.agent_one = self.DriverAgent(1,self.gas,0) # instantiating agent 1 again\n",
    "        self.agent_two = self.DriverAgent(2,int(self.gas*3),0) # instantiating agent 2 again\n",
    "        self.current_player = self.agent_one\n",
    "        # 'P' means the game is playable, 'W' means delivered, 'L' means no delivery\n",
    "        self.state = 'P'\n",
    "        self.current_step = 0\n",
    "        self.max_step = int(self.gas*5) # agent can choose not move as an alternate choice\n",
    "        # selecting a world at random to function with\n",
    "        # even the initial world configuration should be different\n",
    "        # prob = random.uniform(0, 1)\n",
    "        # if prob > 0.25 and prob <= 0.25:\n",
    "        #     self.world_start = self.world_two\n",
    "        # elif prob > 0.5 and prob <= 0.75:\n",
    "        #     self.world_start = self.world_three\n",
    "        # elif prob > 0.75 and prob <= 1:\n",
    "        #     self.world_start = self.world_four\n",
    "        # elif prob < 0.25:\n",
    "        #     self.world_start = self.world_one\n",
    "        self.world_start = self.world_one    \n",
    "        self.world = np.copy(self.world_start) # the self.world can be different from intial world\n",
    "        # no exploration_prize and bonus_reward as per my design\n",
    "        return self._next_observation()\n",
    "    \n",
    "    def _next_observation(self):\n",
    "        obs = self.world\n",
    "        data_to_add = [0] * np.size(self.world, 1)\n",
    "        data_to_add[0] = self.current_player.name # adding current player's label in the observation\n",
    "        obs = np.append(obs, [data_to_add], axis=0)\n",
    "        # observation Sample provided below for reference:\n",
    "        # last row, represents 'data_to_add' vector\n",
    "        # a 3x3 observation grid example for reference\n",
    "        # array([[1, 0, 0],\n",
    "        #         [3, 0, 2],\n",
    "        #         [0, 0, 4],\n",
    "        #         [1, 0, 0]])\n",
    "        return obs\n",
    "\n",
    "    def _take_action(self, action):\n",
    "        # agent's name is matched to the array entries for index identification\n",
    "        # 'current_player.name' should be updated alongside the array values\n",
    "        current_pos = np.where(self.world == self.current_player.name)\n",
    "        # the current agent must have gas in it\n",
    "        if self.current_player.gas > 0:\n",
    "            if action == 0:\n",
    "                next_pos = (current_pos[0] - 1, current_pos[1]) # agent moving upwards\n",
    "\n",
    "                if next_pos[0] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 1:\n",
    "                next_pos = (current_pos[0], current_pos[1] + 1)\n",
    "                limit = np.size(self.world, 1)\n",
    "\n",
    "                if next_pos[1] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent.\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1.\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 2:\n",
    "                next_pos = (current_pos[0] + 1, current_pos[1])\n",
    "                limit = np.size(self.world, 0)\n",
    "\n",
    "                if next_pos[0] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # teducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            elif action == 3:\n",
    "                next_pos = (current_pos[0], current_pos[1] - 1)\n",
    "\n",
    "                if next_pos[1] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] >= 0 and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "        else:\n",
    "            # player 1's gas is supposed to go empty first\n",
    "            # therefore, upon having empty gas tank player should be allowed to\n",
    "            # drop the package in the environment and disappear from the location\n",
    "            if self.current_player.package == 3:\n",
    "                self.world[current_pos] = self.current_player.package\n",
    "                # agent dissappears from the grid after this drop\n",
    "            else:\n",
    "                self.world[current_pos] = 0 # if gas is finished, agent should dissappear\n",
    "        # if gas is empty for both agents, the episode should stop at that instant\n",
    "        if self.agent_one.gas == 0 and self.agent_two.gas == 0:\n",
    "            self.state = 'L'\n",
    "\n",
    "\n",
    "    def step(self, action):\n",
    "        self._take_action(action)\n",
    "        self.current_step += 1\n",
    "        # uncomment the below statement out, while debugging\n",
    "        # print(self.world) \n",
    "\n",
    "        if self.state == \"W\":\n",
    "            reward = 5.75\n",
    "            done = True\n",
    "        elif self.state == 'L':\n",
    "            reward = 0\n",
    "            done = True\n",
    "        elif self.state == 'P':\n",
    "            reward = 0 # sparse reward encoding, only rewarded when episode ends\n",
    "            done = False\n",
    "\n",
    "        if self.current_step >= self.max_step:\n",
    "            print(f'New episode number {self.current_episode + 1}')\n",
    "            done = True\n",
    "\n",
    "        # agents object used to identify agent properties\n",
    "        if self.current_player.name == 1 and self.current_step > self.gas*2:\n",
    "            self.current_player = self.agent_two\n",
    "        elif self.current_player.name == 2:\n",
    "            self.current_player = self.agent_one\n",
    "\n",
    "        if done:\n",
    "            self.render_episode(self.state)\n",
    "            self.current_episode += 1\n",
    "\n",
    "        obs = self._next_observation()\n",
    "\n",
    "        return obs, reward, done, {'state': self.state}, self.current_player\n",
    "\n",
    "    def render_episode(self, win_or_lose):\n",
    "        # storing the rendered episodes in a file\n",
    "        self.success_episode.append(\n",
    "            'Success' if win_or_lose == 'W' else 'Failure')\n",
    "        file = open('render.txt', 'a')\n",
    "        file.write('----------------------------\\n')\n",
    "        file.write(f'Episode number {self.current_episode}\\n')\n",
    "        file.write(\n",
    "            f'{self.success_episode[-1]} in {self.current_step} steps\\n')\n",
    "        file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "RBMQ5Y-yLJRM"
   },
   "outputs": [],
   "source": [
    "# the NxN road network environment that rotates in its simple\n",
    "# setting implementation.\n",
    "# implementation of the Simple NxN road network for MultiAgent RL\n",
    "class GridRoadEnv(gym.Env):\n",
    "    # defining the Driving Agent with name and gas values plus package flag\n",
    "    class DriverAgent():\n",
    "        def __init__(self, name, gas, package):\n",
    "            self.name = name\n",
    "            self.gas = gas\n",
    "            self.package = package\n",
    "\n",
    "    def __init__(self, size=5, gas=7):\n",
    "        # size: grid size, gas: values restricted between (n_min: trunc(n/2), n_max: 2*(n-1)-1)\n",
    "        # converting even size to nearest odd number\n",
    "        self.size = size\n",
    "        if size % 2 == 0:\n",
    "            self.size = self.size+1\n",
    "        # defining different possible world configurations\n",
    "        self.world_one = np.zeros((self.size,self.size), dtype=int)\n",
    "        # setting up env values for the numpy array\n",
    "        self.world_one[0][0]=1\n",
    "        self.world_one[self.size-1][self.size-1]=4\n",
    "        self.world_one[self.size//2][self.size-1]=2\n",
    "        self.world_one[self.size//2][0]=3\n",
    "        self.world = self.world_one\n",
    "\n",
    "        self.world_two = np.rot90(self.world_one)\n",
    "        self.world_three = np.rot90(self.world_two)\n",
    "        self.world_four = np.rot90(self.world_three)\n",
    "        # even the initial world configuration is defined to be different upon\n",
    "        # environment instantiation\n",
    "        prob = random.uniform(0, 1)\n",
    "        # default value assignment below\n",
    "        if prob > 0.25 and prob <= 0.25:\n",
    "            self.world = self.world_two\n",
    "        elif prob > 0.5 and prob <= 0.75:\n",
    "            self.world = self.world_three\n",
    "        elif prob > 0.75 and prob <= 1:\n",
    "            self.world = self.world_four\n",
    "        self.world_start = self.world # this 'world_start', if reset() is called, never gets used\n",
    "        # adding five actions for the environment\n",
    "        # 0: up, 1: right, 2: down, 3: left\n",
    "        # when agent reaches at package location it automatically picks up the package\n",
    "        self.action_space = spaces.Discrete(4)\n",
    "        shape_0 = np.size(self.world_start, 0)\n",
    "        shape_1 = np.size(self.world_start, 1)\n",
    "        self.observation_space = spaces.Box(low=0,\n",
    "                                            high=4,\n",
    "                                            shape=(shape_0 + 1, shape_1),\n",
    "                                            dtype=np.int16)\n",
    "        self.reward_range = (-10, 10)\n",
    "        self.current_episode = 0\n",
    "        self.success_episode = []\n",
    "        # defining the driver agents in the environment\n",
    "        self.gas = gas\n",
    "        # truncating the gas to the desirable range for experimentation\n",
    "        if (self.gas < self.size//2) or (self.gas > (2*(self.size-1)-1)):\n",
    "            self.gas = random.randint(self.size//2, 2*(self.size-1))\n",
    "        self.agent_one = self.DriverAgent(1,self.gas,0) # 3 integer value, when carrying package\n",
    "        self.agent_two = self.DriverAgent(2,int(self.gas*3),0) # 3 integer value, when carrying package\n",
    "        self.spec = EnvSpec(\"GridRoadEnv-v0\")\n",
    "\n",
    "    def reset(self):\n",
    "        # game like formulation, each player agent moves one step at a time\n",
    "        self.agent_one = self.DriverAgent(1,self.gas,0) # instantiating agent 1 again\n",
    "        self.agent_two = self.DriverAgent(2,int(self.gas*3),0) # instantiating agent 2 again\n",
    "        self.current_player = self.agent_one\n",
    "        # 'P' means the game is playable, 'W' means delivered, 'L' means no delivery\n",
    "        self.state = 'P'\n",
    "        self.current_step = 0\n",
    "        self.max_step = int(self.gas*5) # agent can choose not move as an alternate choice\n",
    "        # selecting a world at random to function with\n",
    "        # even the initial world configuration should be different\n",
    "        prob = random.uniform(0, 1)\n",
    "        if prob > 0.25 and prob <= 0.25:\n",
    "            self.world_start = self.world_two\n",
    "        elif prob > 0.5 and prob <= 0.75:\n",
    "            self.world_start = self.world_three\n",
    "        elif prob > 0.75 and prob <= 1:\n",
    "            self.world_start = self.world_four\n",
    "        elif prob < 0.25:\n",
    "            self.world_start = self.world_one\n",
    "        # self.world_start = self.world_one    \n",
    "        self.world = np.copy(self.world_start) # the self.world can be different from intial world\n",
    "        # no exploration_prize and bonus_reward as per my design\n",
    "        return self._next_observation()\n",
    "    \n",
    "    def _next_observation(self):\n",
    "        obs = self.world\n",
    "        data_to_add = [0] * np.size(self.world, 1)\n",
    "        data_to_add[0] = self.current_player.name # adding current player's label in the observation\n",
    "        obs = np.append(obs, [data_to_add], axis=0)\n",
    "        # observation Sample provided below for reference:\n",
    "        # last row, represents 'data_to_add' vector\n",
    "        # a 3x3 observation grid example for reference\n",
    "        # array([[1, 0, 0],\n",
    "        #         [3, 0, 2],\n",
    "        #         [0, 0, 4],\n",
    "        #         [1, 0, 0]])\n",
    "        return obs\n",
    "\n",
    "    def _take_action(self, action):\n",
    "        # agent's name is matched to the array entries for index identification\n",
    "        # 'current_player.name' should be updated alongside the array values\n",
    "        current_pos = np.where(self.world == self.current_player.name)\n",
    "        # the current agent must have gas in it\n",
    "        if self.current_player.gas > 0:\n",
    "            if action == 0:\n",
    "                next_pos = (current_pos[0] - 1, current_pos[1]) # agent moving upwards\n",
    "\n",
    "                if next_pos[0] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 1:\n",
    "                next_pos = (current_pos[0], current_pos[1] + 1)\n",
    "                limit = np.size(self.world, 1)\n",
    "\n",
    "                if next_pos[1] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "\n",
    "            elif action == 2:\n",
    "                next_pos = (current_pos[0] + 1, current_pos[1])\n",
    "                limit = np.size(self.world, 0)\n",
    "\n",
    "                if next_pos[0] < limit and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[0] < limit and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[0] < limit and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "            elif action == 3:\n",
    "                next_pos = (current_pos[0], current_pos[1] - 1)\n",
    "\n",
    "                if next_pos[1] >= 0 and int(self.world[next_pos]) == 0:\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos]) in (1, 2):\n",
    "                    pass # two Agents can't be at the same place\n",
    "\n",
    "                elif next_pos[1] >= 0 and (int(self.world[next_pos]) == 3):\n",
    "                    self.world[next_pos] = self.current_player.name\n",
    "                    self.current_player.package = 3 # package is also hidden now from other agent\n",
    "                    self.world[current_pos] = 0\n",
    "                    # reducing the agent's gas by 1\n",
    "                    self.current_player.gas = self.current_player.gas - 1\n",
    "\n",
    "                elif next_pos[1] >= 0 and int(self.world[next_pos] == 4):\n",
    "                    # player should only be allowed this transition to this position\n",
    "                    # when it is having the package with it\n",
    "                    if self.current_player.package == 3:\n",
    "                        self.world[next_pos] = self.current_player.name\n",
    "                        self.world[current_pos] = 0\n",
    "                        self.state = 'W' # and the episode, should end at that moment\n",
    "                        # reducing the agent's gas by 1\n",
    "                        self.current_player.gas = self.current_player.gas - 1\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "        else:\n",
    "            # player 1's gas is supposed to go empty first\n",
    "            # therefore, upon having empty gas tank player should be allowed to\n",
    "            # drop the package in the environment and disappear from the location\n",
    "            if self.current_player.package == 3:\n",
    "                self.world[current_pos] = self.current_player.package\n",
    "                # agent dissappears from the grid after this drop\n",
    "            else:\n",
    "                self.world[current_pos] = 0 # if gas is finished, agent should dissappear\n",
    "        # if gas is empty for both agents, the episode should stop at that instant\n",
    "        if self.agent_one.gas == 0 and self.agent_two.gas == 0:\n",
    "            self.state = 'L'\n",
    "\n",
    "\n",
    "    def step(self, action):\n",
    "        self._take_action(action)\n",
    "        self.current_step += 1\n",
    "        # uncomment the below statement out, while debugging\n",
    "        # print(self.world) \n",
    "\n",
    "        if self.state == \"W\":\n",
    "            reward = 5.75\n",
    "            done = True\n",
    "        elif self.state == 'L':\n",
    "            reward = 0\n",
    "            done = True\n",
    "        elif self.state == 'P':\n",
    "            reward = 0 # sparse reward encoding, only rewarded when episode ends\n",
    "            done = False\n",
    "\n",
    "        if self.current_step >= self.max_step:\n",
    "            print(f'New episode number {self.current_episode + 1}')\n",
    "            done = True\n",
    "\n",
    "        # agents object used to identify agent properties\n",
    "        if self.current_player.name == 1 and self.current_step > self.gas*2:\n",
    "            self.current_player = self.agent_two\n",
    "        elif self.current_player.name == 2:\n",
    "            self.current_player = self.agent_one\n",
    "\n",
    "        if done:\n",
    "            self.render_episode(self.state)\n",
    "            self.current_episode += 1\n",
    "\n",
    "        obs = self._next_observation()\n",
    "\n",
    "        return obs, reward, done, {'state': self.state}, self.current_player\n",
    "\n",
    "    def render_episode(self, win_or_lose):\n",
    "        # storing the rendered episodes in a file\n",
    "        self.success_episode.append(\n",
    "            'Success' if win_or_lose == 'W' else 'Failure')\n",
    "        file = open('render.txt', 'a')\n",
    "        file.write('----------------------------\\n')\n",
    "        file.write(f'Episode number {self.current_episode}\\n')\n",
    "        file.write(\n",
    "            f'{self.success_episode[-1]} in {self.current_step} steps\\n')\n",
    "        file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id='governance-wrapper-layer-implementation'></a>\n",
    "#### 2. Governance Wrapper Layer Implementation\n",
    "\n",
    "The `SimpleGovernanceKernelWrapper` class inherits a specific environment, and adds reward distribution based wrapper functionalities to give dense reward signals during agent transitions.\n",
    "Additionally, this layer also contains functionality to add in the directional governance component as well for better agent guidance and comparison with MORS baseline approach.\n",
    "Second, the `RewardShapedGridRoadEnv` class implements the MORS baseline approach with directional gradient signal and additional rewards for intermediate task completion.\n",
    "The inline comments elaborates the functionalities of the implemented governance layer in more detail for more granular understanding of different components in the governance layer.\n",
    "\n",
    "**Note:** The current design limits inheriting only a specific kind of class environment for a defined governance wrapper layer.\n",
    "Therefore, for every experiment we have to replace and update the inherited parent environment class multiple times to carry out different experiments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "id": "VSDoXxRXa197"
   },
   "outputs": [],
   "source": [
    "# basic conceptual framework of the governance wrapper:\n",
    "# * environment prior information available, like initial agent positions and corresponding gas info.,\n",
    "# * environment implementation needs change, agent info. or the whole agent instance needs to be returned as well\n",
    "# * iterative radial reward value decrease by a factor of 5 from initial movement step from 0.25 up until {gas/dim} boundary\n",
    "# * noise addition in the values, based on the grid size for size 3 noise factor zero, but with larger grid size, larger noise, for exploration promotion\n",
    "# * partial task completion reward, 0.25 reward value attained upon package pickup by the agent\n",
    "# * prior reward shaped grid based on basic environment information whill be matched with shared observation state for new reward calculation\n",
    "# * additionally, create a simple reward shaping environment that only assigns 0.25 reward value upon package pickup by the agent\n",
    "\n",
    "# environments to inherit from:\n",
    "# SizeThreeGridRoadEnv, SizeThreeSimpleGridRoadEnv, SimpleGridRoadEnv, GridRoadEnv\n",
    "class SimpleGovernanceKernelWrapper(GridRoadEnv): \n",
    "\n",
    "    def governance_reward(self, observation, kernel, agent_name, prev_obs):\n",
    "        agent_ind_x, agent_ind_y = np.where(observation == agent_name)\n",
    "        prev_agent_ind_x, prev_agent_ind_y = np.where(prev_obs == agent_name)\n",
    "        prev_obs = observation # saving the previous observation state\n",
    "        # if agent is not present in the observation grid, return 0 reward\n",
    "        if agent_ind_x[0] >= kernel.shape[0] or agent_ind_y[0] >= kernel.shape[1]: \n",
    "            return 0, prev_obs\n",
    "        # if agent is still at present at the previous grid position, return 0 reward\n",
    "        if agent_ind_x[0] == prev_agent_ind_x[0] and agent_ind_y[0] == prev_agent_ind_y[0]: \n",
    "            return 0, prev_obs\n",
    "        return round(kernel[agent_ind_x[0]][agent_ind_y[0]],2), prev_obs\n",
    "\n",
    "    def gradient_kernel(self, world_map):\n",
    "        grad_kernel = np.zeros((world_map.shape[0], world_map.shape[1]))\n",
    "        grad_unit = 0.025\n",
    "        for i in range(0,world_map.shape[0]):\n",
    "            for j in range(0,world_map.shape[1]):\n",
    "                if i>=j:\n",
    "                    grad_kernel[i][j] = grad_unit*(i+j)\n",
    "                else:\n",
    "                    grad_kernel[i][j] = grad_unit*(i-j)\n",
    "        return grad_kernel\n",
    "\n",
    "    def circular_kernel(self, world_map, agent_name):\n",
    "        agent_ind_x, agent_ind_y = np.where(world_map == agent_name)\n",
    "        grad_kernel = np.zeros((world_map.shape[0], world_map.shape[1]))\n",
    "        grad_unit = 0.075\n",
    "        # print(agent_ind_x, agent_ind_y)\n",
    "        for i in range(0,world_map.shape[0]):\n",
    "            for j in range(0,world_map.shape[1]):\n",
    "                if i != agent_ind_x[0] and j != agent_ind_y[0]:\n",
    "                    grad_kernel[i][j] = 1/2*(grad_unit/(i-agent_ind_x[0])**2  +  grad_unit/(j-agent_ind_y[0])**2)\n",
    "                elif i == agent_ind_x[0] and j != agent_ind_y[0]:\n",
    "                    grad_kernel[i][j] = grad_unit/(j-agent_ind_y[0])**2\n",
    "                elif j == agent_ind_y[0] and i != agent_ind_x[0]:\n",
    "                    grad_kernel[i][j] = grad_unit/(i-agent_ind_x[0])**2\n",
    "                elif i == agent_ind_x[0] and j == agent_ind_y[0]:\n",
    "                    grad_kernel[i][j] = 0\n",
    "        return grad_kernel\n",
    "    \n",
    "    # superposition of governance kernels for generating the effective kernel output for both the agents\n",
    "    # print(np.add(gradient_kernel(world_map), circular_kernel(world_map, 1)))\n",
    "    # print(np.add(gradient_kernel(world_map), circular_kernel(world_map, 2)))\n",
    "\n",
    "    def __init__(self, radial=False):\n",
    "        super().__init__()\n",
    "        self.prev_obs_agent_one = self.world_start\n",
    "        self.prev_obs_agent_two = self.world_start\n",
    "        self.radial_flag = radial\n",
    "        if self.radial_flag == False:\n",
    "            self.gov_kern_agent_one = np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 1))\n",
    "            self.gov_kern_agent_two = np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 2))\n",
    "        else:\n",
    "            self.gov_kern_agent_one = self.circular_kernel(self.world_start, 1)\n",
    "            self.gov_kern_agent_two = self.circular_kernel(self.world_start, 2)\n",
    "        self.preward_flag_agent_one = True\n",
    "        self.preward_flag_agent_two = True\n",
    "\n",
    "    def reset(self):\n",
    "        self.prev_obs_agent_one = self.world_start\n",
    "        self.prev_obs_agent_two = self.world_start\n",
    "        if self.radial_flag == False:\n",
    "            self.gov_kern_agent_one = np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 1))\n",
    "            self.gov_kern_agent_two = np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 2))\n",
    "        else:\n",
    "            self.gov_kern_agent_one = self.circular_kernel(self.world_start, 1)\n",
    "            self.gov_kern_agent_two = self.circular_kernel(self.world_start, 2)\n",
    "        self.preward_flag_agent_one = True\n",
    "        self.preward_flag_agent_two = True\n",
    "        return super().reset()\n",
    "\n",
    "    def step(self, action):\n",
    "        next_state, reward, done, info, agent = super().step(action)\n",
    "        reward_new = 0\n",
    "        if agent.name == 1:\n",
    "            # print(next_state, self.prev_obs_agent_one)\n",
    "            reward_new, self.prev_obs_agent_one = self.governance_reward(next_state, self.gov_kern_agent_one, agent.name, self.prev_obs_agent_one)\n",
    "            # print(reward_new, self.prev_obs_agent_one, agent.name)\n",
    "        elif agent.name == 2:\n",
    "            # print(next_state, self.prev_obs_agent_two)\n",
    "            reward_new, self.prev_obs_agent_two = self.governance_reward(next_state, self.gov_kern_agent_two, agent.name, self.prev_obs_agent_two)\n",
    "            # print(reward_new, self.prev_obs_agent_two, agent.name)\n",
    "        # print('\\n')\n",
    "\n",
    "        if agent.package == 3 and self.preward_flag_agent_one == True and agent.name == 1:\n",
    "            reward_new = reward_new + 0.05\n",
    "            self.preward_flag_agent_one = False\n",
    "            # print(reward_new, agent.name)\n",
    "        elif agent.package == 3 and self.preward_flag_agent_two == True and agent.name == 2:\n",
    "            reward_new = reward_new + 0.05\n",
    "            self.preward_flag_agent_two = False\n",
    "            # print(reward_new, agent.name)\n",
    "\n",
    "        reward_new  = round(reward_new, 2) + reward\n",
    "        return next_state, reward_new, done, info # , agent\n",
    "            #, agent; will give too values to unpack error for the default stablebaselines model implementation\n",
    "\n",
    "# current checks, the reward shaping with kernel is working correctly,\n",
    "# needs to add reward for successfully picking up the package as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "id": "NyC2BU_OYyTW"
   },
   "outputs": [],
   "source": [
    "# basic conceptual framework of the MORS wrapper:\n",
    "# * environment prior information available, like initial agent positions and corresponding gas info.,\n",
    "# * environment implementation needs change, agent info. or the whole agent instance needs to be returned as well\n",
    "# * iterative radial reward value decrease by a factor of 5 from initial movement step from 0.25 up until {gas/dim} boundary\n",
    "# * noise addition in the values, based on the grid size for size 3 noise factor zero, but with larger grid size, larger noise, for exploration promotion\n",
    "# * partial task completion reward, 0.25 reward value attained upon package pickup by the agent\n",
    "# * prior reward shaped grid based on basic environment information whill be matched with shared observation state for new reward calculation\n",
    "# * additionally, create a simple reward shaping environment that only assigns 0.25 reward value upon package pickup by the agent\n",
    "\n",
    "# environments to inherit from:\n",
    "# SizeThreeGridRoadEnv, SizeThreeSimpleGridRoadEnv, SimpleGridRoadEnv, GridRoadEnv\n",
    "class RewardShapedGridRoadEnv(GridRoadEnv):\n",
    "\n",
    "    def governance_reward(self, observation, kernel, agent_name, prev_obs):\n",
    "        agent_ind_x, agent_ind_y = np.where(observation == agent_name)\n",
    "        prev_agent_ind_x, prev_agent_ind_y = np.where(prev_obs == agent_name)\n",
    "        prev_obs = observation # saving the previous observation state\n",
    "        # if agent is not present in the observation grid, return 0 reward\n",
    "        if agent_ind_x[0] >= kernel.shape[0] or agent_ind_y[0] >= kernel.shape[1]: \n",
    "            return 0, prev_obs\n",
    "        # if agent is still at present at the previous grid position, return 0 reward\n",
    "        if agent_ind_x[0] == prev_agent_ind_x[0] and agent_ind_y[0] == prev_agent_ind_y[0]: \n",
    "            return 0, prev_obs\n",
    "        kernel_reward_val = kernel[agent_ind_x[0]][agent_ind_y[0]] \n",
    "        if agent_name == 1:\n",
    "            self.gov_kern_agent_one[agent_ind_x[0]][agent_ind_y[0]] = kernel[agent_ind_x[0]][agent_ind_y[0]] / 2\n",
    "        if agent_name == 2:\n",
    "            self.gov_kern_agent_two[agent_ind_x[0]][agent_ind_y[0]] = kernel[agent_ind_x[0]][agent_ind_y[0]] / 2        \n",
    "        return round(kernel_reward_val, 2), prev_obs\n",
    "\n",
    "    # working as a proxy for directionality aspects\n",
    "    # no radial governance kernel added in rewad shaping approach\n",
    "    def gradient_kernel(self, world_map):\n",
    "        grad_kernel = np.zeros((world_map.shape[0], world_map.shape[1]))\n",
    "        grad_unit = 0.01\n",
    "        for i in range(0,world_map.shape[0]):\n",
    "            for j in range(0,world_map.shape[1]):\n",
    "                if i>=j:\n",
    "                    grad_kernel[i][j] = grad_unit*(i+j)\n",
    "                else:\n",
    "                    grad_kernel[i][j] = grad_unit*(i-j)\n",
    "        return grad_kernel\n",
    "    \n",
    "    # superposition of governance kernels for generating the effective kernel output for both the agents\n",
    "    # print(np.add(gradient_kernel(world_map), circular_kernel(world_map, 1)))\n",
    "    # print(np.add(gradient_kernel(world_map), circular_kernel(world_map, 2)))\n",
    "    def __init__(self, radial=False):\n",
    "        super().__init__()\n",
    "        self.prev_obs_agent_one = self.world_start\n",
    "        self.prev_obs_agent_two = self.world_start\n",
    "        self.radial_flag = radial\n",
    "        if self.radial_flag == False:\n",
    "            self.gov_kern_agent_one = self.gradient_kernel(self.world_start) # np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 1))\n",
    "            self.gov_kern_agent_two = self.gradient_kernel(self.world_start) # np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 2))\n",
    "        else:\n",
    "            self.gov_kern_agent_one = self.gradient_kernel(self.world_start) # self.circular_kernel(self.world_start, 1)\n",
    "            self.gov_kern_agent_two = self.gradient_kernel(self.world_start) # self.circular_kernel(self.world_start, 2)\n",
    "        self.preward_flag_agent_one = True\n",
    "        self.preward_flag_agent_two = True\n",
    "\n",
    "    def reset(self):\n",
    "        self.prev_obs_agent_one = self.world_start\n",
    "        self.prev_obs_agent_two = self.world_start\n",
    "        if self.radial_flag == False:\n",
    "            self.gov_kern_agent_one = self.gradient_kernel(self.world_start) # np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 1))\n",
    "            self.gov_kern_agent_two = self.gradient_kernel(self.world_start) # np.add(self.gradient_kernel(self.world_start), self.circular_kernel(self.world_start, 2))\n",
    "        else:\n",
    "            self.gov_kern_agent_one = self.gradient_kernel(self.world_start) # self.circular_kernel(self.world_start, 1)\n",
    "            self.gov_kern_agent_two = self.gradient_kernel(self.world_start) # self.circular_kernel(self.world_start, 2)\n",
    "        self.preward_flag_agent_one = True\n",
    "        self.preward_flag_agent_two = True\n",
    "        return super().reset()\n",
    "\n",
    "    def step(self, action):\n",
    "        next_state, reward, done, info, agent = super().step(action)\n",
    "        reward_new = 0\n",
    "        if agent.name == 1:\n",
    "            # print(next_state, self.prev_obs_agent_one)\n",
    "            reward_new, self.prev_obs_agent_one = self.governance_reward(next_state, self.gov_kern_agent_one, agent.name, self.prev_obs_agent_one)\n",
    "            # print(reward_new, self.prev_obs_agent_one, agent.name)\n",
    "        elif agent.name == 2:\n",
    "            # print(next_state, self.prev_obs_agent_two)\n",
    "            reward_new, self.prev_obs_agent_two = self.governance_reward(next_state, self.gov_kern_agent_two, agent.name, self.prev_obs_agent_two)\n",
    "            # print(reward_new, self.prev_obs_agent_two, agent.name)\n",
    "        # print('\\n')\n",
    "\n",
    "        if agent.package == 3 and self.preward_flag_agent_one == True and agent.name == 1:\n",
    "            reward_new = reward_new + 0.05\n",
    "            self.preward_flag_agent_one = False\n",
    "            # print(reward_new, agent.name)\n",
    "        elif agent.package == 3 and self.preward_flag_agent_two == True and agent.name == 2:\n",
    "            reward_new = reward_new + 0.05\n",
    "            self.preward_flag_agent_two = False\n",
    "            # print(reward_new, agent.name)\n",
    "\n",
    "        reward_new  = round(reward_new, 2) + reward\n",
    "        return next_state, reward_new, done, info # , agent\n",
    "        #, agent; will give too values to unpack error for the default stablebaselines model implementation\n",
    "\n",
    "# current checks, the reward shaping with kernel is working correctly,\n",
    "# needs to add reward for successfully picking up the package as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "88yl4Hz8Fvuo"
   },
   "outputs": [],
   "source": [
    "# storing obs data into simple road env for random walk testing for the environment\n",
    "import random\n",
    "simple_obs_list = []\n",
    "simple_info_list = []\n",
    "env_simple = RewardShapedGridRoadEnv()\n",
    "# result_test = []\n",
    "\n",
    "for e_ in range(3):\n",
    "    obs = env_simple.reset()\n",
    "    for i in range(200):\n",
    "        action = random.randint(0, 4) # model_simple.predict(obs)\n",
    "        obs, reward, done, info = env_simple.step(action)\n",
    "        # print(obs, reward, done, info) # , agent.name, agent.gas, agent.package, agent.picked)\n",
    "        # print('\\n')\n",
    "        simple_obs_list.append(obs)\n",
    "        simple_info_list.append(info)\n",
    "        if done:\n",
    "            print(info['state'])\n",
    "            # obs = env_simple.reset()\n",
    "            # result_test.append(info['state'])\n",
    "            break\n",
    "# printing the output results w/ successful completions\n",
    "# result_stat = result_test.count('W') / len(result_test)\n",
    "# print(f'Success rate: {result_stat * 100} %')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id='rl-algorithm-model-training'></a>\n",
    "#### 3. RL Algorithm Model Training\n",
    "\n",
    "The `stable_baselines3` package implementation code for all the trained models as part of this experiment is presented in this section.\n",
    "The naming convention of the trained model follows the specified order `{rl-algo}_model_{grid-size}_{orientation}`, and is uniquely defined for each of the experiment model variants.\n",
    "Additionally, even though the `SimpleGovernanceKernelWrapper` and `GreedyGovernanceKernelWrapper` are the same for the models with incorporated governance layers, we manually change the inherited environment class before in the above governance layer section.\n",
    "\n",
    "This section also defines basic plotter function to quickly observe the convergence plot, and get an idea about the initial model performance.\n",
    "Also, we manually output the entire logging file generated by `Monitor` wrapper, and save it for further creating our experiment summaries in a more presentable manner."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "id": "bpbJzTgt5LcI"
   },
   "outputs": [],
   "source": [
    "# stable_baseline3 library related import statements\n",
    "# from stable_baselines3.common.vec_env import DummyVecEnv # it's usage produces 'spec' related error\n",
    "# from stable_baselines3.common.vec_env import SubprocVecEnv # it's usage produces 'spec' related error\n",
    "from stable_baselines3 import PPO\n",
    "# result plotter package function\n",
    "from stable_baselines3.common import results_plotter\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# important note: implementation limitation, currently both governance variants\n",
    "# inherit specific environments during initialization & these wrappers are hard binded to the code\n",
    "# hence, need to be updated for manual experimentation with different environments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "ZDcvSMZ570LX"
   },
   "outputs": [],
   "source": [
    "# ploting function utilities for smoothed moving average reward values\n",
    "def moving_average(values, window):\n",
    "    \"\"\"\n",
    "    Smooth values by doing a moving average\n",
    "    :param values: (numpy array)\n",
    "    :param window: (int)\n",
    "    :return: (numpy array)\n",
    "    \"\"\"\n",
    "    weights = np.repeat(1.0, window) / window\n",
    "    return np.convolve(values, weights, 'valid')\n",
    "\n",
    "def plot_results(log_folder, title='Learning Curve'):\n",
    "    \"\"\"\n",
    "    plot the results\n",
    "\n",
    "    :param log_folder: (str) the save location of the results to plot\n",
    "    :param title: (str) the title of the task to plot\n",
    "    \"\"\"\n",
    "    x, y = ts2xy(load_results(log_folder), 'timesteps')\n",
    "    y = moving_average(y, window=100)\n",
    "    # Truncate x\n",
    "    x = x[len(x) - len(y):]\n",
    "\n",
    "    fig = plt.figure(title)\n",
    "    plt.plot(x, y)\n",
    "    plt.xlabel('Number of Timesteps')\n",
    "    plt.ylabel('Rewards')\n",
    "    plt.title(title + \" \")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "LJjDiyrTkN8O"
   },
   "outputs": [],
   "source": [
    "# creating the environment for code compilation testing only\n",
    "# env_simple = RewardShapedGridRoadEnv()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "# env_simple = Monitor(env_simple, log_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "dx5z0giWkOC8"
   },
   "outputs": [],
   "source": [
    "env_simple = SimpleGovernanceKernelWrapper()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_simple = Monitor(env_simple, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_3x3_simple = PPO(\"MlpPolicy\", env_simple, verbose=0)\n",
    "ppo_model_3x3_simple.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "G_kGxZyl8Z2_",
    "outputId": "0d257618-4b88-4205-b304-9d1481d9ebc7"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 3x3 Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "AaSBY3oS8hVa",
    "outputId": "7b45c53d-d756-4b17-92cc-44ccbdd0ad55"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 3x3 Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ZOAF0NEKK2G8"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "qjutkXgF8Wp8"
   },
   "outputs": [],
   "source": [
    "env_simple = RewardShapedGridRoadEnv()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_simple = Monitor(env_simple, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_3x3_reward = PPO(\"MlpPolicy\", env_simple, verbose=0)\n",
    "ppo_model_3x3_reward.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "drZd00wm9yLI",
    "outputId": "aab0379f-1302-4ca6-f5d2-e695da56830a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 3x3 Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "eGQAVbSS8Wth",
    "outputId": "67373f41-3c91-4293-cb6f-d4ca5f564163"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 3x3 Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "G5wbIlsTL88p"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "a2tMCdn98Wvj"
   },
   "outputs": [],
   "source": [
    "env_random = SimpleGovernanceKernelWrapper()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_random = Monitor(env_random, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_3x3_random = PPO(\"MlpPolicy\", env_random, verbose=0)\n",
    "ppo_model_3x3_random.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "u5N_rhihBvnx",
    "outputId": "b32d7607-0374-4b0a-c407-4979da03fbfd"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 3x3 Random Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "R6IORUpBBv3v",
    "outputId": "2ac53ca6-df33-4f91-ba15-60e0f99b49aa"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 3x3 Random Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "F2TQ4XnO8WxG"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "C8chkfvb8Wzy"
   },
   "outputs": [],
   "source": [
    "env_random = RewardShapedGridRoadEnv()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_random = Monitor(env_random, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_3x3_reward_random = PPO(\"MlpPolicy\", env_random, verbose=0)\n",
    "ppo_model_3x3_reward_random.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "WJtUoz0_5A-N",
    "outputId": "86b11a98-80b2-454e-fd3e-4e2e4c612d53"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 3x3 Random Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "7plf_2xjhSnm",
    "outputId": "e4c9638e-7e82-4d64-df65-4ee13b43e4bb"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 3x3 Random Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "zSNMlPjLhSvv"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "UCz8DeTjR4Fg"
   },
   "outputs": [],
   "source": [
    "env_simple = SimpleGovernanceKernelWrapper()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_simple = Monitor(env_simple, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_5x5_simple = PPO(\"MlpPolicy\", env_simple, verbose=0)\n",
    "ppo_model_5x5_simple.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "A5pTgiXqR3-Q",
    "outputId": "5bbbe130-e290-4d72-b592-5f4ae2410bb0"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 5x5 Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "k1EsmzEJR3zi",
    "outputId": "d2e23a2d-6f7c-4f37-b253-f2386edd78fb"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 5x5 Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "JOO-8C0RR3ru"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "aCecnF-PNtTP"
   },
   "outputs": [],
   "source": [
    "env_simple = RewardShapedGridRoadEnv()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_simple = Monitor(env_simple, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_5x5_reward = PPO(\"MlpPolicy\", env_simple, verbose=0)\n",
    "ppo_model_5x5_reward.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "Vut_onVrdRmy",
    "outputId": "ce12620d-1833-4d3c-eb39-44c8a4b6486e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkIAAACICAYAAAALS4T7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2deXhdVbXAf+vezEOTlHQeKRTaUkpLC2UGZaoVtQoIMqOIqAiI4gMFB0BBeCr1ATIJgoJMIjKPWgoIpS0UOkFbOrcpbdpMTdKM6/2x9705ubm5uWmTJmnW7/vud8/Z01l7n2Gvs/Y6e4uqYhiGYRiG0RsJdbUAhmEYhmEYXYUpQoZhGIZh9FpMETIMwzAMo9diipBhGIZhGL0WU4QMwzAMw+i1mCJkGIZhGEavxRQhw+jGiMhfROTGrpajJyAis0Tkop3MqyKybyfIdLSIfNLR5XYVInKciKwP7C8WkeO6UKQOQ0R+KSJ/62o5jN2PKUJGpyIiq0WkWkS2i8hnvmPP8XGzRGSHjysWkadEZFAg7xEi8m8RqRCRMhF5VkTGJTjWcSLS6MuL/M5PUs6/iEhtTN5wK2kvEJEGn6ZcRD4UkVPa2za7Sowcwd/gDj5OsF0rROQTEbmwI4/R2YhIvojcLyKbfB2WicjVnX1cVX1TVffvjLJj7p8yEZktIgd2xrFaQ1UPUNVZHV1uzP24TUReFZExHX2cBMcf6ZXjlN11TKPrMEXI2B18SVVzgIOBKcC1gbhLfdx+QD7wBwARORx4BfgXMBjYG/gQeFtERiU41kZVzQn8HmyHnLfE5G1IkPYdL3c+cCfwqIjkt+NYHcU7MTLnqOrGTjjORl/fPsAPgXtFpFM6+LYQR3ufXX8AcoCxQB7wZWBFR8vWBUTun77ALOCvXStOh3KLr9sQYAPw5y6Wx9hDMUXI2G2o6gbgRWB8nLhtwD8CcbcAD6nqTFWtUNVtqnot8C7wy/YeW0T28W+WB/v9wSKyZVfN+qraiOt8soHRvuw8EXnIl79GRK6NdNxejn+LyFZvBXs4qECJyCQRed9bLR4DMnZWNm+N+7GIfOQtBo+JSIaPWxq0YolIipf34Dbqq6r6ArANmODzhkTkahH51NfrcRHp6+MeFJEf+e0h/i37+4G22ObzF4jIc16GEr89NCDfLBH5tYi8DVQBo0TkRBH52NftdkASiH4I8Iiqlqhqo6p+rKpPxqQ5QUSWi0ipiNwhIhKQM9E5Wy0i14jIEi/7A4F2jh1KavWc+PifiEiRiGwUkYskySE7r7Q/CkQtpiJyqIi84+tTJCK3i0iajxMR+YOIbBZn1VwoIuN9XLqI/K+IrBVnxb1LRDLjHdfX5wS//Ut/7h/y1+9iEZkSSDtYRP7hz/EqEbmsrXr5ulUDjwMTkynL13uer9dnIvJ7H97sXMTKH8Ns/18qzip1uIjsKyJv+PNW7O9PYw/AFCFjtyEiw4DpwAdx4gqBU4EPRCQLOAJ4Ik4xjwMnJjhMf//wW+Uf9NkAqvop8D/A33z5DwAPxpj1v+c75vkicmqSdQoDFwJ1wBof/H84q8Mo4FjgPJ8GXGd9E87KNRYYhlfsfCf1NE6x6uvrn5QcCfg6MA1nUZsAXODD/w58I5DuZKBYVd9PVJhXWr4MFNJkUfkBMANX18FACXCHj3sDOM5vHwusBI4J7L/plckQ7pyMAIYD1cDtMYc/F7gYyAXKgKdw1sVC4FPgyASivwv8WkQuFJHRraQ5BacwTcC128mRatPKOQtwtk+/D866eS2tE/eciMg04ErgBGBfmtqtTfy1czaunhEacNa7QuBw4Hjgez7uJNx52A93rX4d2OrjbvbhE70cQ4CfJynKl3EKWT7wDP4cinsReBZn1R3iZblCRE5upZxg3bJx1+qKJMuaCcxU1T648/F4krIHiVyj+d7K+g5wA85KXQAMxd3nxp6AqtrPfp32A1YD24FSnKJwJ5Dp42bh3u5Lcabvh4F+uIeMAmPilDcNqGvlWANxb8QhXCczG7g7Js0zwELgIyA9EH4wsBeQglPWKoAjWznOBUC9l7sO12l/3ceFgVpgXCD9d4BZrZQ1A/jAbx8DbAQkEP9f4MYk5Ij8Po1p+3MC+7cAd/ntfX0ds/z+w8DPWznOcUCjL78G18FeEYhfChwf2B/k2yUF1xGV+HNyl2+L9T7dg8CVrRxzIlAS2J8FXB/YPw94N7AvwHrgolbKywR+Csz3sq0AvhCIV+CowP7jwNVtnbNAO18S2J8eOQ++7dYneU7uB24KxO3r5dq3FTlm0XT/1OCUw+PjpfXprwD+6bc/DywDDgNCMe1YCewTCDscWJWgPif47V8CrwXixgHVfnsqsDZGnmuAB1qR9S/ADl+3RmAVMCGZsnD3/a+AwjjX8fqYsFj5/+a3R/q2TwmkfQi4BxjaWhvbr2f+zCJk7A5mqGq+qo5Q1e+pM3VHuMzHDVHVs1V1C67jbMR1qLEMAorjHURVN6nqEnVDH6uAn9DSonIvbvjt/1S1JpD3fVXdqqr16oZ+Hga+lqBO76pqPu7t8BngaB9eCKTSZB3Cbw8BEJEBIvKoiGwQkXLgbz4POIvDBlXVmLyJeNe3X+S3T0z8psB2Fc5PBlVdgVNgvuQtZF8GHklwnI2+vn2AP+I60ggjgH/6IZhSX24DMECdJa4Sp9gcDTwHbBTnX3QszmKEiGSJyN3ihhLLcZ1ZvjR3WF8X2B4c3PdtFoxvhqpWq+pvVHUyTuF9HHhC/BBeorZq45zFk22Nl6814h4ntk6J6hPgMn9eMnEWrSdFJDJkuZ+4IcZNXu7fRORW1X/jrDV3AJtF5B4R6YN7EckC5gfO50s+PBli65YhzuF4BDA4UqYv96fAgARl/a+v20jcy0bEJ62tsr6Fs2h9LCJzpeM+ZPgJTlF8zw/7fbODyjW6GFOEjG6HqlYC7wCnx4n+OvB6skURuMbFfa12G87p8pcxnWC8vIl8TiKybge+C5wrIpNwSlod7mEdYTjO4gWuM1LgQHWm+3MCxykChoiIxOTtLCLDY18BlnjlKCFeefwf4EARmeGD1+GsK0GFLEOdTxg4Zec0IM2HvQGcj1MiF/g0P8J1dFN9u0SGJoJtEVQQi3BDVC6Ra7NhJIGqRpSCbJzlsC0SnbMIwWMPx1n22ksRzhoar8yEeOX/TZyl6yQf/CfgY2C0l/unBORW1T96xXAcTnG4Cnf9VgMHBM5lnjqn5V1hHc6qFLxGclV1ehJ1WwtcDsz0vkoJy1LV5ar6DaA/8FuccpiNU8izIuV6Jbs1BU9bBLgXrW+r6mCcZfNO6YQpF4zdjylCRnflauB8EblMRHLFOdPeiDPT/ypeBhH5nIiMEMcwnK/DvwJJZgLzVPUi4HncUE0k72kikuN9YE7CdXbPJCOoOkfv+3BDSw04a8OvvdwjcH4fkflJcnFDhWUiMgTX+UR4BzfUdZmIpIrI14BDk5FhJ3kU12l+l8TWoGaoai3wO5r8Ru7C1XcEgIj0E5GvBLK8AVxKkwPqLL//ljZ9mZeL64BLvYL6izbEeB44QES+5i0Ol+GGRuMiIteJyCEikibOOfly3LBLMnP8JDpnEb4vIkO97D8DdsaR9nHgQhEZ661017Uns7gvLccBiwNylwPbxX16/t1A2kNEZKqIpOIUhB1Aozp/rXuBP4hIf592SDK+PG3wHlAhIv8jIpkiEhaR8SJySDKZVfVVnHJ5cVtlicg5ItLP16XUF9GIGwrMEJEv+npfC6S3csgtPk/0C1UROV2aHPhLcMpSYzvawOimmCJkdEtU9S2c8+nXcG/Ka4BJOD+O5a1km4Tzqan0/wtxHSS+Y55GU2dwJXCwiJzt9y/HWW1KgVuBb2v75ke5DZjuhyV+4GVYCbyFUzLu9+l+hfNHKsN15k8F6lzr63sB7qusM4LxrXC4tJxHKNnOpQinfB1B+zvu+4HhIvIlnIL5DPCKiFTgHHanBtK+geuUI4rQW7g389mBNLfhhneKff6X2pC9GGcxvBnn5DsaeDtRFpwzdjGuQz0R+KK36LVFq+cswCM4R9qVOMftdk+Cqaov4oYd/4Oz7EQcn2tazQS3R847zsn+Wl8OwI+Bs3C+YPfS/Bz38WEluHtrK+66B2fxWwG864fUXqNpWGqn8ArvKbgh0lW483AfzlE7WW7FDU+ltFHWNGCxb5OZwJl+aLQM5yx+H+5er8T5lcWTtwr4NW66jlIROQznSD/Hl/sMcLmqrmyH/EY3RZq7IxiGYRjtQURW45y0X+vgcscCi3BO/fUdWbZhGE2YRcgwDKObICJfFTePTwHOv+VZU4IMo3NplyLk/Sf6dJYwhmEYvZzvAJtxw2sNBPx6DMPoHNocGhORR4BLcDflXNzY8kxVvTVhRsMwDMMwjG5OMhahcf5z0xm45RH2xs3wahiGYRiG0aNJZmXdVP+p4QzgdlWtE5Eu87AuLCzUkSNHdtXhDcMwDMPoYcyfP79YVePOG5WMInQ3bhryD4HZfq6Q8o4Tr32MHDmSefPmddXhDcMwDMPoYYhIq7P0t6kIqeofcXNbRFgjIp/rCMEMwzAMwzC6klZ9hETkytZ++EnqDGNPYP6aEs778xzmrylpV5pk8rXnWMG4nSm7rePOuP0tZtzxdoeV2RnH2NV6d3V+wzB6HoksQrn+f3/cjJqR5Qa+hJvi3Ohk5q8pYeZry7j8hP2YPKKgq8Vpk/bIG0wLcMOzi0GE604Z12I/UlYkz7Txg3hpUVH0OMHwx+eupbiylm3baxlSkMnNp05oIUtsOeU76lmwrpTy6joAKmsbyE5PaSbLss+2U1XXwMot26msbeCMKcN4bN46SqrqKK+uo09magu5Hpmzlltf/phj9+vHG8u2cNXJY3hpURGzlxfz9opiRvXLaSbfDc8uZsH6MopKqymurKWkyslz+Qn7MfO1ZYwb1IfH5q3jqpPHcNbU4S3qHZH761OG8dKiIvpmp/HshxvJSgtTUdMQrf+5973LkAK35FKknhF5b3phCXlZaRTmpEfr39o5nb+mhBueXUxlrSt7Q0k1VXVu+4L755CXlUZZVR3XTB/L/gNzueHZxRRX1lJWVcfZU4ezpKi8RZ0ibVBeXcfTlx4VbcOrTh7D/gNzW8gSkSFyrcx8bRmzlxczb/U2QKLXwCebKqLlnDU1/vJt89eUcMH9c6ioaaCotJpB+ZnR6zPYzohE2zi27WPbLFH7tXYvxKYPtkFrsseW0968bZUZr56t3VexbRZpp+C90Z46dBTtLbsz0reWprPqvSsydUTZO8PubiNI7vP52bip6Cv8fi7wvKoekzBjJzFlyhTdU3yE2jqx5/15DrOXF3PM6EIe+tbUOCXsfPm7eoPEi2uPvMG0ALOXuwXlC7JSGbFXNgvWuSWCgmVF8hRkpVJSVReNO/F3s1i+pZKs1BBVdc2X/slND3PN9HHRh/Gd/1nO+tIdgFt9UoGh+RkU5qTz6ZZKKmqa5q6bODSPytoGlm9OvApDSKBRm8rLSg2TnRFmS0Vti7RDCzJZX1Id3RcgJSSERECgpr65/JGyWxzT/3fFQkchJyoNNim9YRg9hKIHr6CmaHnchbSTcZYeAASf6LU+zEhAMopG5O0ViKs4RN6sIv/tJVH5bR17Z/K3R97YtOXVdXy6pZKSqjpG9NWoElJUtoMZd7zNdaeMi6YNvl0CbCqPLMUkTByaR3FlLRtLq2lUqKhp4KYXllBR08DCDWVRKws0LS9dVl3PqH45VNTUk5UaYkd9I40Kn26ppCGeFhJDJEkkZVVdQ9QyEktQCYrkqWtU4ix23azsFuFtStV5JNEkhmEYPYZkFKGHgPdE5J9+fwbwl06TaA8hGUWjLcVh8oiCnbIEJVN+MkpLe/O3Jm9rQ1rBtE9felQL5TFiAQLXng99a2o0T9DMf830sS3M/8FhjoF5mUzKy4hahDaW7qAwN420lHCzoRsgOkyWEhIqauoZ3S+bDaXVNDRCZnqIiup6vnTQYBZvKGPtNqfUZKaHKK+qpzA3jbzMNABKq2tbWIRSw8LIvbI5YHAfnl6wMRoesSQB9MtJo7Kmnpr6xl22uAiQl5VCaVV9dB8gPyuF7TsavAJmGIbRu0k4NCYiAgwF+gFH++DZqvrBbpAtLj1laKyn+fd0Jq0NabVFrP9He9txZ85Ba0rbzjB/TQln3P0O9V7hCNY71vclorRNHJZPn4yUqAIY4TdfPZDH567low1lNCqEBW6YcWDUbwOa/H2C/jARn6SCrFTuO/+QqIIekSVS36Cvzv4Dc7nowbmUVNWREpJm8r+1ophGdcNjT1xyRDRdcIgzLG7YLCzw+CVHRP1+Jg7N4+lLj2JHXQPn3z+HOauaHJLTwiHGDMrlo/VlTBqWxw0zDmRkYTYrt2ynvlH56ztrWFpUTl1DIyERtlXWsrXSKZr7D8jlk88qomVFhhMPHVnA45cc0ey8RhTdeO3xf2cdzOriSn78xIfNhkMFOOmAAfTLTUcVVnxWQXpaCrOXbUn6Whicl8EJ4wbw6Nx11NY3kp4S4r7zp3Dv7JVxfYMmXf8KJVV1FGSl8sHPT4p7bUXSjyrM5sTfv0FxZS0FWalMHJbP+2tLKauua5EvwiEjC5i7urlDeGFOGsXbaxmS714aRhZmc+5hI1otY2eG7nfluRjpq8778xzeXLGVY0YXRv3nYv0Fd/a+VVU+/7tZrCquYkh+BicdMJBgFxmRIRKUnZ7CFSeMJj0lnFT57W2z2PTxXAp21nWiPcedcftbze7hziLYV9x3/iEd2neKyHxVnRI3LgkfoYWqemCHSbOL9BRFyGiiI5WLnkbEAXlgXnzH7QjxnMc3lFWzbXst3z56FFdPHxstL9b6lejhGtsxJNtRxDrJRpTRVxdv4r63VnHRUXtz9fSxcR1kW3PqvvyE/dhSUcMlf5vf7Fh9MlIo39H2uqJpKSFOHDuARlUaGpVXlnzGkfvuRU56Ci8v/gyAjNQQOwJ+Ysft34/hfbNoVEUQJgzN49kPNzZrj+ufXczAvIxoGQAD+2SwtbKGrLQU7jl3MlNH7dVCnn8v/Yw/zfqUK07cjyH5mSiuo1y8sZwH/7uKcw8fSf/cDP40a0ULB/qgkhrvvLXXyTneeX1tyWc8+M5qhuRn8unm7cxdU8Lo/jn075NOYyOEQvDlgwYzeUQBWWkpFJXt6FRn4o6is4/bnR2223Ks7yw5Z9zxNgvWlTJxWD5Pf//IDj1WouN2JLuqCD2Im1F6bodKtZOYImQYzeku1sfSqlreXbmNkEA4JKSGQ/zt3TWs3lrJSeMG8ti8dWypqImmz8tM5bUrj6UwJ433Vm1zflUKjao8vWAj768pYVjfTL5z7D6ERJg8ooCc9Pij+ZsrdpCeEiYvM5V126p459Ot/O7VT9hR1+icu70VKUJExrrA+GN2WpiLjh7FUaMLmTKiAGcQ7zx253nrLteI0TPZE66fXVWEPgb2BdYAlXiXBlWd0NGCJoMpQobRfVBVNpXv4L1V27jy8Q/bdC4/a+pwLv3cvgzOz9xNEjbx/toSZi/bQqNCY6Py1opiFqwrZXjfLG47cyIHD++ZD3jDMNomkSKUjLP0yR0sj2EYPYwtFTXkZabSqMrdb6zkjEOGkZ+Vyhdmvsmq4spouguOGMlpk4cy8/XlAHz/c/uSl5lKWXUd/XLTGdIFClCEg4cXNFN2fnTSfmytrGWv7LROt/4YhtF9SWaJjTUAItIfyOh0iQzD6Fas3LKd43//BimBoaRH3luDIGwq38Fe2WlcM30s4wb1YdzgPgDce17cF69uhYhQmJPe1WIYhtHFtKkIiciXgd8Bg4HNwAhgKXBA54pmGEZXs72mnrPvm4MqzfxpPvNzN40qzOb1Hx1rFhXDMHosra41FuAG4DBgmaruDRwPvNupUhmG0eXU1Dcw6fpXKCpzM3HP/dkJ3DhjPB/9sulz7hevONqUIMMwejTJ+AjVqepWEQmJSEhV/yMityVTuIisBiqABqA+1lHJz1M0E5gOVAEXqOr77aqBYRidwiNz1katQI9/53D65aZzjp9X5qNfnkRJZW3S86cYhmF0V5JRhEpFJAeYDTwsIptxX48ly+dUtbiVuC8Ao/1vKvAn/28YRhdRWVPPvW+u5LbXnMPzJzdOa6Hw9MlIpU9GaleIZxiG0aEkowh9BagGfgicDeQB13fQ8b8CPKTuG/53RSRfRAapalEHlW8YRjt4d+VWzrynaeT7wiNHmtXHMIw9mmQUoTNxy2osBx5sZ/kKvCIiCtytqvfExA8B1gX21/uwZoqQiFwMXAwwfHjbs6wahtE+auob+Os7a7jx+aUA5Gel8s0j9+aSY/fpYskMwzA6l2QUoeHA3SKyNzAPN0T2pqouSCLvUaq6wX96/6qIfKyqs9srpFeg7gE3oWJ78xuGkZiv3vFflhSVA3DUvoXce94UMtPMEmQYxp5PMvMI/QJARDKBbwNXAbcBbT4lVXWD/9/sV68/FKdIRdgADAvsD/VhhmF0EJU19Sz7rIJGhbLqWj4/ZkA07oO1JXz1zv8CMLp/Dv931iTGDOzTVaIahmHsdpKZR+ha4EggB/gA+DHwZhL5soGQqlb47ZNo6Vv0DHCpiDyKc5IuM/8gw+g4Hpu7lp//azE19Y0t4gqyUimpalqh/K/fmsrAPJsz1TCM3kUyQ2NfA+qB54E3gHdUtSZxFgAGAP/0c4ykAI+o6ksicgmAqt4FvID7dH4F7vP5C9tdA8MwAHh58SYenrOW/rnpPDl/PTfOGM+1Ty+Kxo8ZmMvHmyqi+xElaPyQPjz+ncPJSkvmcWAYhrFn0eaiqwAi0gdnFToKOB3YrKpHdbJscbFFVw3D0dConHPfHEqr6zhlwiBuffmTuOkevfgwDhu1FwCriyupb2zk/bWl5GWmcvyY/qSEk5lX1TAMo+eyS4uuish44GjgWGAK7iuvNofGDMPoOFZs3s5try1jwtA8tlbWcvW0Mfxj/nreWbkVgKXe0TnCgUPyGDsol+F9s6JKEMDIwmwA9u2fu/uENwzD6MYkYwu/Gaf4/BGYq6p1baQ3DKMD2VZZywm/fwOA5z5yLnRvfLKFjzdVkJ4S4rUrj2X11kr6ZKRy0LD8rhTVMAyjx5HMV2On+C/GhpsSZBi7nwfeXgXA2EF9OPXgIdz4/NKor885h41gWN8shvXN6koRDcMweixtOgeIyJeABcBLfn+iiDzT2YIZhuH8gJ5fWMTEYfm8ePnRXHT0KFbdND0a/9PpY7tQOsMwjJ5PMkNjv8TN/zMLQFUX+MkVDcPoRLbX1HPcrbMo3l7D5cePjoaLCB/+4iTCISEcspXfDcMwdoVkV58v85/BR7DZnQ2jkxn/i5cBmDA0j28d3fzdIy/TFjw1DMPoCJJRhBaLyFlAWERGA5cB/+1csQxjz+XR99Zy9VMLyc1I4e5zJ3PEPoU0NCpvrShGVZkysi+PzW1agu9f3z+SmBcRwzAMo4NIRhH6AfAzoAb4O85X6IbOFMow9mSufmohABU76jnr3jkJ0/77R8eaEmQYhtGJJPPVWBVOEfoZgIjsD9yOW3fMMIw2WLetijtnrWB432wG57slLIbkZ7K5Ygd1Dc1HmQ8Y3IfFG8vJSgvz4uVHM2Kv7K4Q2TAMo9fQqiIkIhOA/wUGA08Dd+AUoKnA73aLdIaxB3Dfmyv5+3vrmoU9+M1D2bd/Du98upWlReUMzs/kwKF5DMhN541lW/j8mP5mCTIMw9gNJLII3Qv8CXgH+ALuE/oHgbNVdcdukM0wejx1DY088t5aAE4YO4DXln7G+CF92Ld/DgCH77MXh++zV7M8x48d0KIcwzAMo3NIpAilq+pf/PYnInKZqv5kN8hkGHsEqsrvXllGXYPy21MP5IxDhlNVW09KyNb2MgzD6C4kUoQyRGQSELHP1wT3VfX9zhbOMHoy98xeyV1vfArA6ZOHAdgK74ZhGN2MRE/lIuD3gf1NgX0FPt9ZQhlGT2L+mm28u3Ibk4bnU1PXyOfG9EdVeWL+egB+eMJ+hGziQ8MwjG5Jq4qQqn5udwpiGD2VHz72IWu3VUX3v3jgIEYPyGHF5u385qsHctbU4V0onWEYhpEIs9Mbxk5SUlnLj59orgQBPL+wCNxUQXzxwEFdIJlhGIaRLJ2mCInIMOAhYABuKO0eVZ0Zk+Y44F/AKh/0lKpe31kyGUZH8rU//ZdVxZWMGZjLg988lPpGJTUk/OP9Dfz2pY+54oTR5GXZUhiGYRjdmc60CNUDP1LV90UkF5gvIq+q6pKYdG+q6imdKIdhdAhLi8opqarlj68v5ysTh7CquJL9B+Ty0hXHNEv33eP24bvH7dNFUhqGYRjtoU1FSNysbmcDo1T1ehEZDgxU1fcS5VPVIpzDNapaISJLgSFArCJkGN2eO2et4JaXPonuf7C2lNSw8OR3D+9CqQzDMIxdJZkJTe4EDge+4fcrcLNMJ42IjAQmAfEWVjpcRD4UkRdF5IBW8l8sIvNEZN6WLVvac2jD2GVUlUf9zNBD8jMBGDOoD/97+kHkZtjQl2EYRk8mmaGxqap6sIh8AKCqJSKSluwBRCQH+AdwhaqWx0S/D4xQ1e0iMh23lMfo2DJU9R7gHoApU6ZobLxhdBaNjcpJt81m7bYqbj1tAqdPGdbVIhmGYRgdSDKKUJ2IhHEOz4hIP6AxmcJFJBWnBD2sqk/FxgcVI1V9QUTuFJFCVS1OSnrD6ARUlQsemMv6kio2lu6guq6BKSMK+NJBg7taNMMwDKODSUYR+iPwT6C/iPwaOA24tq1M3rfoz8BSVf19K2kGAp+pqorIobihuq3JCm8YHY2qcunfP+CNZc2HYB+48BAyUsNdJJVhGIbRWbSpCKnqwyIyHzget7zGDFVdmkTZRwLnAgtFZIEP+ykw3Jd7F06p+q6I1APVwJmqakNfxm7nreXFLNpYxh9fX05VbQM56SnM/dkJXPboB4wqzDZfIMMwjD0UaU3vEJG+iTKq6rZOkagNpkyZovPmzeuKQxt7GKrKis3beeS9tTzw9upo+OfH9Of2sybZumCGYRh7CCIyX1WnxItL9KSfj0A7bZ4AABcUSURBVPMLEpwVp8Rv5wNrgb07WE7D2C2oKg/PWcu1Ty9qFn7Q0DxuO3MSexdmd5FkhmEYxu4m0VpjewOIyL3AP1X1Bb//BWDG7hHPMDqWlVu2c/OLH/PKks+iYXedM5nJIwrol5vehZIZhmEYXUEytv/DVPXbkR1VfVFEbulEmQxjl1FV7n97NTe9sJT6RmXcoD5cccJoLv7r/GiaaQcM5OTxA5g2fmAXSmoYhmF0JckoQhtF5Frgb37/bGBj54lkGLtGSWUtj7y3lltfbpoJeklReVQJCoeEmWdO5JQJ9jm8YRhGbycZRegbwC9wn9ADzKZplmnD6Hac/8B7fLS+DIDXrjyWlJBQVl3Hz55eSL+cdO6/4BDc7A6GYRhGbyeZz+e3AZf7hVNVVbd3vliGkTzlO+qYs3Ib5dV1vLhoU1QJ+s6xo9i3f0403XM/OLqrRDQMwzC6Kcksunog8BDQ1+8XA+er6qKEGQ2jk1BVyqrraFRoaFTOv/89lhQ1rd7SLzed2Vd9jsw0mwDRMAzDSEwyQ2N3A1eq6n8AROQ43LpfR3SiXIYBwJPz1/PhulJyMlIY0TeL0uo6nlmwsZniA3Do3n255dQJpISFwpx0mwXaMAzDSIpkFKHsiBIEoKqzRMQmWjE6jVcWb2LBulJWb63khYWb4qbZKzuNH3x+X8LhEOkpIU6ZMMgmQDQMwzDaTTI9x0oRuQ74q98/B1jZeSIZvZGy6jrufuNTXlnyGSs2Oze0kEBqWPj3j44jJSzc+Z9PGTMol1MPHkp6Ssgcng3DMIxdJhlF6JvAr4DI6vGzfZhh7DSriit5cVERpx08lBcXbeKWlz6msrYBgLOmDudHJ+5H3+w0gKjCc8OM8V0mr2EYhrFnksxXYyXAZQAiEsYNlZUnzmV0R+avKWHma8u4/IT9mDyioM1008YP4qVFRXHT76hrYHtNPauKK/n1c0uiSswZhwxj9IBcBBicn8m+/XOorm3gg7UlPL1gAx+uK6O+sZFPt1QCcMtLTXP9nD55KFedvD/rSqr54WMLWpUz2Xq0l44otzNk66z6GoZhGMl9NfYIcAnQAMwF+ojITFW9tbOF6w0k6uSS6QCDaQBueHYxlbUNZKenMHVkAQ/PWcOg/CxuPnUCM19bxuzlxQD86ZzJPDFvPc9/tJGrp4+N5kWEypp6lm/eztxV26iub2TOqm1MGz+QgX0yAFhXUs0LC4sAt/hccNneG59f2ky+1JBQ36jN0vTLSWO//jms2lrJQUPzmTgsn59MG0NaSoj5a0q46MG5lFTVsXBDGfedf0iLukfqUV5dR5/M1Gbtk6zS8Mictdz0whLystIozEnnulPGccNzS1iwrpTyHfVcd8q4uOXElj9/TUm03SJ5Im380Lemtnr82HMYKePrU4bx+Ny10fImjyiIlhmvPeLJ012VxERlArtdgeyNCmZvrLNhtEWrq89HE4gsUNWJInI2cDBwNTBfVSfsDgFjmTJlit79j1c79QHXWmdXXFlLWVUd10wfy1lTh8c9TmthwY7ugbdWsqF0B31z0ti2vZaqugZy01P4yzcP5ZNNFdz0whIG5mVS39DIqq1VpIaE9NQwO+oa2H9gLukpIVZuqWRAnwzWbaukqq6RrNQw2RlhtlTUxq1TWlhICYeoqm0gLSzUNjSd97AAIjQ0tn4thARSQiHqGhpJdMXEU3yChAUaYiL75aZRWllHXSvHP2Z0IfNWbwOE8w4fwetLP2NTeQ0D8zJYvnk7IZ+uMSbfJceMAuC+t1ZxxD57sXBDGWdMGcaSonLmrNpGTX1TjtH9c9hQUkVVXSMCpKWEqKlvjJ4XcB11UdkOlm/eTljg20eP4rF56yipqgMgPSyEQyGyM8LsqG2MXifQdA1ElNTrThkHOOXz0y2VVNTUA5CVGqKqzsk1cWgeT196VDPl8JjRhTz0ranR6yxWnofeWU1VXSPp4RBjB/dpdpygohUrR6xCEmsNPO/Pc5i9vJhjRhdy+Qn7NVP+dvYeDJYJRLeD9duVDjtYfrwyY+PbS1crFW0d/5E5a7n15Y+56uQx0etwV+vcUXR128WjO8pkdByJVp9PRhFaDEwEHgFuV9U3RORDVT2o40VtmylTpui4796R1ANu4tC8qMUAaPXNGYj7Vh/JX76jngXrSqMy5KaH2adfTrMO7JJjRnHiAQOjHVZBVirH7tePZxZsbNZBh2jZYUfjBFRJqGh0FWGvbTS0JnwXkB4OUbMbBEoN4eYs0uaKXKw1LJaUkHD9V8bz0qKiqMISITc9TEMjVNU1NMuTlRqOhuWmp3DN9LG8tKiIaeMH8fjctVFlPDMtzOaKGtK9wtYaE4fmAbDATzKZmx6moqahWfyabVVRJQucQlKQldpC8YrcI0D0fmitQ41VnNqyrgHtVlJiXzAi7RRR4Noqc2ctRpHwyHMh+Jzp6A40kYxttdGk61+JPos++PlJbZa3O+mOClnsc7+r22hPpauuwUSKULLzCK0GPgRmi8gIoEt9hCIPuWnjB3Hen+dEH0iRoZJp4wcBUL6jntnLi/lgbQkgVNTUM2/1NsKhEPlZqawrqaZ8Rz19MlKiHcUF989hYF4mE4flU1lTz4L1ZQjNO73tNQ3R9BHufXNl9G0coKSqjqcXtFySLVG3ncAg0+XsrL4hwFcmDubFhUXUxJqC2iCe9SjIripBYYFQSKhrQ666wGGCSYO50r2VLRhW36jc+NziqMUuSFAZiZCV6rTNoQWZlFXVUVFTz8//tYh6f2H0yUyNXncRBTwsMLpfNiuLK2nQJkU7cr1GLD/gFKu8rFQqaqqjFiMg2llG7iugmUIReXAhwoJ1pYzul+1llei9FkukrGAHA7RQQFpTSIL5W2Pma8ui7bFma2V0ODXyf9XJY1qVCWDyiIKEnXBrw5zBDvOY0YXR50xsuo4g0VBr7HMwtmO56uQxUYtQhLbqvLuYNn4QCzeUtXr97C6C7Rtpz848n0bia7qraNMiFDeTSIqq1ieRbhowEwgD96nqzTHx6bhZqycDW4EzVHV1ojLHTZikUy6/i3GD+nDvmytpUDesgWq0MzhmdCHTxg/ipheWUNug0TfmYMca6TCyUkOcd/hIHnh7VbOOOtlOcmdJDXdO2VmpIbIzUqJDZLFWi3hWjHgKR1vWDmiyyKSGmisLQYbmZ1BZ28AZU4bx+tLP2FC6o4UVJEJuepiBeZkAbCrbEe3s2yIkcOOMA9l/YG4zC8O5970bVUxjjwMSHVoLMrpfNhceNYqbXljCjvrG6DkKC3zpoMG8sWwLx+7Xj+c+KqK+UclKDREOhWho1Lj1ilh4huZnsK2yNipPWGBUvxw2lFRHh0bBKTgFWancd/4hzayL951/CABXP/lh9DqPWJyCw7QRBSZigZo4NI/rvnRAtF1ueHYxC9aXNRt2S3YIK56FNJk3+raGpHbWOtCaRejWlz+OtlvQqtVe2rIIdbZfViIZgnQX60p76C4yJ+ve0FWy7Il0R4tQq4qQiJyjqn8TkSvjxavq79s4aBhYBpwIrMc5Wn9DVZcE0nwPmKCql4jImcBXVfWMROX2Gba/9j27+aGzUt3wQLBzCL5JuzRhhhRksnzzdiKzz7RXDUlLCVGbYBgiHm1ZNYAW8qSGhfysVLZW1Ca0IMVjaEEmb/3P51t0jCuLK1lfUh1VXtLDzpcn0mb1jdpsuGZ0v2xnSRDhsL37ctdsN3VUeliiCmNuephJwwuaDUOMG9SHh+esiTohLy0qp6a+kZA4i9cxowsZN6gP9721ir7ZaWyuqCErNcx+A3Ob+ZtEnJkjiuzo/jkUb6+hpKqOiUPzKN5ew/rSHUCTL02QoE9OhIhPTPCN3p2Apo60Nd+uYL7gcE9EKRjdL5viytpmSlJBVipXnTyGlxYVtUgX8duYcftbUcXk64cMj/qH3XzqhGibttURxyNW9l15yLflIN7eh1lnKxLJfPW4J9ETO9CeKHNn012Uwz2VnVWEvqOqd4vIL+LFq+qv2jjo4cAvVfVkv3+Nz3dTIM3LPs07IpICbAL6aQIzVebg/XTAeX8AnJKREm7ykQgBqd6ZJXbYpF9OGuXVdS2GL9pDSEAQGpK0okU6/8ibdNCKFSHi4zG6XzbL/SflAAVZqRTmpLN88/aor0jEyRWaLCbBdOCUk4W/mtbiQTPjjrddR9w/h0F5GVHlJZ7VIbaTi9yg4KxvFx65dwsnzNYYd91LVNU1kB4OMXVU36TfvNrr59Ue/4mdfQi3xzk+2bfMjnbgTbYNDMPoXphy2LnskrP0Lhz0NGCaql7k988FpqrqpYE0i3ya9X7/U5+muLVyR42doBmn/Za6BiUkQr020tDorCgNjdrhfjYCZAYcWKFpWC09LIwdnBf93Byahjwi+xEFI9jJRYZmKmrqm3X0kS/GIlaZRE57sY6mVz/5IZvKa6JfKiXrGBovPF7enbUAxPtyJRk6+su/nkJHyt1T28AwDKOj2dWvxkbh/HwOw43evAP8UFUTLrPRkYqQiFwMXAyQWTBg8sjv/6VVP5MIqWEhNRRqNV3scFRKSBiUl8G6kupm/iBXnTyG/QfmNvv0OXauF2jyRxndL5ubTzuozWEJaH3elK6e2M86UMMwDGNPYlcVoXeBO4C/+6AzgR+oakJbe2cNjaUPGq2Dzr8toczgrDYZ3pKTniLU1GszP6HIV2HLN28nKzXMXy9y1UlWAWjvp7iGYRiGYXQNu6oIfRQ7eWIy8wh5xWYZcDywAecsfZaqLg6k+T5wYMBZ+muq+vVE5eYPH6OHXHF31DLz0qKiZg66AEWl1VE/nIjDauRLkkQ+J+3BFB/DMAzD6BnsqiL0W6AEeBQ3knQGUADcCqCq2xLknQ7chvt8/n5V/bWIXA/MU9VnRCQDt6r9JGAbcGZbQ25TpkzRefPmJZQ5nk+LKS6GYRiG0TvZVUVoVYJoVdVRuyJce0lGETIMwzAMw4iwSzNLq+reHS+SYRiGYRhG15NoHqGfqOotfvt0VX0iEPcbVf3pbpIxVq4K4JOuOHY3pxBoddqBXoq1SXysXVpibRIfa5f4WLu0pLu3yQhV7RcvIpEi9L6qHhy7HW9/dyIi81ozb/VmrF1aYm0SH2uXllibxMfaJT7WLi3pyW0SShAnrWzH2zcMwzAMw+hxJFKEtJXtePuGYRiGYRg9jkTO0geJSDl+lQm/jd/P6HTJWueeLjx2d8bapSXWJvGxdmmJtUl8rF3iY+3Skh7bJp221phhGIZhGEZ3J9HQmGEYhmEYxh6NKUKGYRiGYfRaepQiJCLTROQTEVkhIld3tTwdjYjcLyKbRWRRIKyviLwqIsv9f4EPFxH5o2+Lj0QkOL3B+T79chE5PxA+WUQW+jx/FJFu//WfiAwTkf+IyBIRWSwil/vw3t4uGSLynoh86NvlVz58bxGZ4+vymIik+fB0v7/Cx48MlHWND/9ERE4OhPfI+01EwiLygYg85/etTURW+2t8gYjM82G9+h4CEJF8EXlSRD4WkaUicnhvbhcR2d9fI5FfuYhcsce3iar2iB9uvbJPgVFAGvAhMK6r5ergOh4DHAwsCoTdAlztt68Gfuu3pwMv4pzXDwPm+PC+wEr/X+C3C3zcez6t+Lxf6Oo6J9Emg4CD/XYubiHfcdYuCJDjt1OBOb4Oj+PW7AO4C/iu3/4ecJffPhN4zG+P8/dSOrC3v8fCPfl+A64EHgGe8/vWJrAaKIwJ69X3kJf7QeAiv50G5Fu7RNsmDGwCRuzpbdLljd2Ok3I48HJg/xrgmq6WqxPqOZLmitAnwCC/PQj4xG/fDXwjNh3wDeDuQPjdPmwQ8HEgvFm6nvID/gWcaO3SrE2ygPeBqbiZXVN8ePSeAV4GDvfbKT6dxN5HkXQ99X4DhgKvA58HnvN17NVt4mVdTUtFqFffQ0AesAr/0ZC1S4v2OQl4uze0SU8aGhsCrAvsr/dhezoDVLXIb28CBvjt1tojUfj6OOE9Bj90MQln/ej17eKHgBYAm4FXcdaKUlWt90mCdYnW38eXAXvR/vbq7twG/ARo9Pt7YW0Cbu63V0Rkvohc7MN6+z20N7AFeMAPpd4nItlYu0Q4E/i7396j26QnKUK9HnUqdK+c70BEcoB/AFeoankwrre2i6o2qOpEnBXkUGBMF4vUpYjIKcBmVZ3f1bJ0Q45StyzSF4Dvi8gxwcheeg+l4FwR/qSqk4BK3LBPlF7aLng/ui8DT8TG7Ylt0pMUoQ3AsMD+UB+2p/OZiAwC8P+bfXhr7ZEofGic8G6PiKTilKCHVfUpH9zr2yWCqpYC/8EN3eSLSGSi1GBdovX38XnAVtrfXt2ZI4Evi8hq4FHc8NhMenebAKCqG/z/ZuCfOMW5t99D64H1qjrH7z+JU4x6e7uAU5jfV9XP/P4e3SY9SRGaC4wW9wVIGs5s90wXy7Q7eAaIeNyfj/ORiYSf5732DwPKvOnyZeAkESnwnv0n4fwaioByETnMe+mfFyir2+Jl/TOwVFV/H4jq7e3ST0Ty/XYmzm9qKU4hOs0ni22XSHudBvzbv9k9A5wp7guqvYHROGfGHne/qeo1qjpUVUfi5P23qp5NL24TABHJFpHcyDbu2l9EL7+HVHUTsE5E9vdBxwNL6OXt4vkGTcNisKe3SVc7KbXnh/NQX4bzhfhZV8vTCfX7O1AE1OHeVr6F81l4HVgOvAb09WkFuMO3xUJgSqCcbwIr/O/CQPgU3APwU+B2YpwEu+MPOApnhv0IWOB/061dmAB84NtlEfBzHz4K12mvwJm10314ht9f4eNHBcr6ma/7JwS+4OjJ9xtwHE1fjfXqNvH1/9D/Fkfk7u33kJd7IjDP30dP475w6tXtAmTjLKN5gbA9uk1siQ3DMAzDMHotPWlozDAMwzAMo0MxRcgwDMMwjF6LKUKGYRiGYfRaTBEyDMMwDKPXYoqQYRiGYRi9FlOEDMPYLYhb6ft7fnuwiDzZiceaKCLTO6t8wzD2HEwRMgxjd5GPW/EdVd2oqqe1kX5XmIib88cwDCMhpggZhrG7uBnYR0QWiMgTIrIIQEQuEJGnReRVEVktIpeKyJV+Icx3RaSvT7ePiLzkFw59U0TG+PDTRWSRiHwoIrP9rM/XA2f4Y53hZ1e+X0Te8+V+JXDsf4nILBFZLiK/8OHZIvK8L3ORiJzRJS1mGEank9J2EsMwjA7hamC8qk4UkZHAc4G48cAk3GzPK4D/UdVJIvIH3DT8twH3AJeo6nIRmQrciVtP7OfAyaq6QUTyVbVWRH6Om+X2UgAR+Q1uCY1v+qVJ3hOR1/yxD/XHrwLmisjzwAhgo6p+0efP66xGMQyjazFFyDCM7sB/VLUCqBCRMuBZH74QmCAiOcARwBNuiSIA0v3/28BfRORx4CnicxJuQdYf+/0MYLjfflVVtwKIyFO4ZV1eAH4nIr/FLdXxZkdU0jCM7ocpQoZhdAdqAtuNgf1G3HMqBJSq6sTYjKp6ibcQfRGYLyKT45QvwKmq+kmzQJcvdp0hVdVlInIwzs/oRhF5XVWv35mKGYbRvTEfIcMwdhcVQO7OZFTVcmCViJwO4Fe7Pshv76Oqc1T158AWYFicY70M/MCveI2ITArEnSgifUUkE5gBvC0ig4EqVf0bcCtw8M7IbRhG98cUIcMwdgt++Olt7yR9604UcTbwLRGJrKL+FR9+q4gs9OX+F7fK+n+AcRFnaeAGIBX4SEQW+/0I7wH/wK1A/g9VnQcciPMjWgD8ArhxJ+Q1DKMHYKvPG4bRaxGRCwg4VRuG0fswi5BhGIZhGL0WswgZhmEYhtFrMYuQYRiGYRi9FlOEDMMwDMPotZgiZBiGYRhGr8UUIcMwDMMwei2mCBmGYRiG0Wv5fwnqhyocJSBWAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 5x5 Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "VZSaGoeVORy4",
    "outputId": "177a2f56-b80a-4b07-e9f7-3f82be56cb76"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 5x5 Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "3Az2WnNvOR1w"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ofbHPLu2NtVb"
   },
   "outputs": [],
   "source": [
    "env_random = SimpleGovernanceKernelWrapper()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_random = Monitor(env_random, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_5x5_random = PPO(\"MlpPolicy\", env_random, verbose=0)\n",
    "ppo_model_5x5_random.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "uBJSEYRtSUcT",
    "outputId": "b8225abb-32c8-4e35-ac34-9da93b365440"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 5x5 Random Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "jlIjD1v9NtYR",
    "outputId": "46a13d6b-8875-4315-c85a-981aabe27d69"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 5x5 Random Road Env Governance Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "JbCFt4cohS1A"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "wkMIZobSSWBv"
   },
   "outputs": [],
   "source": [
    "env_random = RewardShapedGridRoadEnv()\n",
    "# logs will be saved in log_dir/monitor.csv\n",
    "env_random = Monitor(env_random, log_dir)\n",
    "# training sample environment policy for checking the env functionality\n",
    "callback = SaveOnBestTrainingRewardCallback(check_freq=1000, log_dir=log_dir)\n",
    "\n",
    "ppo_model_5x5_reward_random = PPO(\"MlpPolicy\", env_random, verbose=0)\n",
    "ppo_model_5x5_reward_random.learn(total_timesteps=75000, callback=callback)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "id": "5479k3ype49X",
    "outputId": "bbd1707d-9e11-49be-b8e3-60e0ea319511"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper function from the ploting library\n",
    "results_plotter.plot_results([log_dir], 1e5, results_plotter.X_TIMESTEPS, \"PPO 5x5 Random Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "foR8syoSSXel",
    "outputId": "488cd799-e547-4de3-cf4e-c4ea152c1a59"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZQAAAEWCAYAAABBvWFzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2dd3gc1fWw37PqtmTLtty7cccYY4zBVFNMaKEFAoSQkBAIJAQIaZBAQkh+CSEN8kHoCRB6EkLvxWCajW3cC7h3W7bcZNXdvd8f985qdrVareRdaSWd93n22Zm5d2bO1DP3nHvPEWMMiqIoirK/BFpbAEVRFKV9oApFURRFSQmqUBRFUZSUoApFURRFSQmqUBRFUZSUoApFURRFSQmqUNoYImJEZHhry5FK2uMxpQMRGeLOVXYz1r1URD5Ik1z3isjN6dh2ayAiD4vIb930MSKyvLVlShUiskZETkrX9ltUobiDqRSRchHZ6i5coSubLiJVrmy7iDwrIn196x4pIu+IyF4R2S0iL4rI2AT7mioiYbc97/fNJOV8WERqYtbNaqDupSIScnX2iMh8ETmjqecm02jserSSHN7vxTTsx3/Ny0TkTREZner9pBMROVpEPnLPR5mIfCgih6V7v8aYK40xv0n1dn0K1LvuW0Xk7yKSk+p9NYQxZoYxZlQ6tu2ObZ87to0i8peG3jNp2v8tIvJYKrfZGi2ULxtjCoGJwCTgJl/Z1a5sJFAM/BVARKYAbwDPA/2AocB84EMRGZZgX5uMMYW+3yNNkPP2mHVDCep+7OQuBv4OPCUixU3YV6biXY/hQCHwp9aUw/f7cpr2c7s73v7ARuChNO2nUZraChGRLsBLwP8DumOP4ddAdeqla3GK3XU5CJgCfL+V5UklB7tjOw64APh2K8uzX7SaycsYsxF4FRgXp6wM+K+v7HbgUWPMncaYvcaYMmPMTcAnwC1N3beIHOC+4Ca6+X4iUioiU5t3NBG5w8C/gM7ACN++3hGRHe5L/3G/snGtth+LyAL3Zfm0iOT7yn8iIptFZJOIRN1sItJVRB51sq8VkZtEJODKLnVfqH8VkV0issq18i4VkfUisi3ZFpsxZhfwHDDBt+8jReRTJ/OnInKkr+xbIrLUtSZXich3Y+Ru8JiagmuFbhCRH7nj2Swi33Jlh4vIFv8Xn4icIyILkjjeSuCZmOPtJyL/ded6tYhc45bni211l7j5X4hI0L3gEZHfiMgdbvp0EfnMtWTXi8gtvu17X+OXicg64B0RyRKRP7n7ZhVwegKxRzrZnzTGhIwxlcaYN4wxUcfrtrfTHcOpvuUNXjPfef65k2WNiFzsK/ebiBq8Jq68h1jrwh533/xWkjTFGWO2AW8CEcuEiNwgIiud3EtE5Bxf2XARec/do9tF5Glf2WixrdAyEVkuIl+Nt0/veHzzjT2vZ4jIPPfMfSQi45M8thXAh0Tfcw1uS0R+JrZVs9fJf6JbHrkW8eT3LT8F+DlwgdgW0ny3/FJ3/fe6e+Ti2HUbO5AW+wFrgJPc9EBgMfAbNz8d+I6bLgHewb6cOwEh4Pg42/sWsLmBfU0FaoCtwGpsa6ezr/xyYInb/uvAn3xlDwNl7jcH+EqCY7oU+MBNZ2G/nmqAXm7ZcGAakAf0BN4H7og5J7OwLa/uwFLgSld2ipN/HFZJPQEYYLgrfxTbaisChgCfA5f55Aq6c5QF/BZYB9ztZDkZ2AsUNnBc/uvRA3gLeN7Ndwd2ApcA2cBFbr6HKz8dOAAQ7JdXBTAxmWNKJEcD1zgI3ArkAKe5fXVz5SuBab76/wZuaGBbDwO/ddOdsffefDcfcPfBL4FcYBiwCviSK3/fu0ewLemVwKm+snN88h7ktjfenYezXdkQdx4edfsvAK4ElmGfle7Au65Odhz5uwA7gEeAU71zEHOf1mLv+yzgKmATIElcM+88/wV77xwH7ANGxTl3jV2Tp9yvE1YxrMc9P3GOyTsn2W6+H9Yy8W1fnfPd8gD2C38f0NeVPQn8wpXlA0f7ru967LORDRwCbAfGNnA8G5J8Xg8BtgGHu3P8TVc/r4Hj8z/Lo4HNwA8b2xYwysnfz3eeDoiVPYH83jv4FuAxX1lnYI/vuvYFDmzSO74plff35w6mHNgFrMWahwp8L44KV7YReBz7Ah7gTvzoONs7BahtYF993A0bwJrI3gfui6nzArAQWOC/6FhzXA93s52GffEe1cB+LsU+QLuwD2wl8NUE5+Bs4LOYc/J13/ztwL1u+h/Abb6ykd5N6G6yGtxD4Mq/C0z3yfWFr+wgt25v37IdwIQG5PSux2633jxgkCu7BJgVU/9j4NIGtvUccG1jx9SIHLt8P+8jZKo739m++tuAI9z0b4F/uOki7MtmcAP7eRioctsPYz9Cxruyw4F1MfVvBP7ppn8D/M3dL1uAa4HbsC+xSpyijbPPO4C/uukh7jwM85W/g3tZufmTaUChuPIx7jg2YO/JF7zr7e6HFb66ndy2+iRxzaa67fk/yJ4BbvadO/8LOO41wd6ztbgXlu8aNaZQvOtugI+ALgmer3nAWW76UeB+YEBMnQuAGTHL7gN+1cDxxL6QG3pe78Hdm77y5cBxDchqsC/wfW76Sdx7KNG2sM//NuAkICfOfbw/CmUX8BXce7mpv9YweZ1tjCk2xgw2xnzPWPOCxzWurL8x5mJjTCn2yzeM1Zax9MV+WdTDGLPFGLPEGBM2xqwGfoo9UX4ewH4p/z9jTLVv3bnGmB3GmKAx5hWscjs3wTF9YowpBrphH+JjvAIR6S0iT7nm6R7gMWwLzM8W33QF1l8B9itova9srW+6BPsFuDamvL9vfqtvutIdW+yyQhrmGmNMV+zXdDescvfkWhtTN7JvETlVRD5x5oRdWKVc4lu3oWNKJEex7+fvUbTDGBP0zfvP3xPAuSKSh71+c40xifb3J3cdh2DPjeeMHQz0c6aHXe6Yfg70duXvYR/cidgPlDexD/4R2Jf4DoiY4d51ZrPd2BZI7L3gPzdNOlfGmKXGmEuNMQOw93U/rNLy2OKrW+EmvU4xia4ZwE5jzL4YWfo1IEpD16QnVun6j8k/3RAl7rp0wpqFXvcKROQbPrPQLuxxe3L/FNvimiUii6XOvDoYODzmel6M/QhNhoae18HAj2K2O5CGzxPYe6YQq+QOx77UE27LWPPYdViFsM29XxLtIync9b0Ae19uFpGXpYkdUzK+27A7yI+xTdtYvgq8neym8B2v2N5ld2Adr7eISPdG1pUkZC3HmhIuEZFD3OLfufUPMsZ0Ab6ezLYcm7E3kccg3/R27Nfe4JjyjUluO2mMMQuxX5J3i4hgTSWDY6oNAja6l/d/sQ783u5F8Ap1x5zomFIt9xLsi+9U4GtYBZPMeuuwrYw7RaQA+9JbHaPUiowxp7lVPsIqn3OA99x+B2Ffyu/5Nv0E9oNjoFPU91L/XjC+6WafK2PMMuzXaj0fZSxJXDOAbiLS2Tc/CHsfNIVSbEtngG/ZwAbq1sN9fD4MHCEiJSIyGPtReDW2FVgMLPLkdh+Vlxtj+mFb738X2z19PfY6+a9noTHmqiYeTyzrgf+L2W4nY8yTjRyXMcY8g33P/TKZbRljnjDGHI19Dg3wB7fePqzi9UikJE29Bca8boyZhv1YX4Y9v0mT8QrFcQPwTRG5RkSKRKSbczxNwfZkqYeIHC8ig8UyEGuCeN5X5U5gtjHmO8DL2IfbW/c8ESkUkYCInIxVAi8kI6ixHQoepO7GKMKa+XaLSH/gJ0047meAS0VkrIh0An7l20/Ilf+fOyeDgeuxLaB08Aj2i/xM7MtmpIh8TUSyReQCrHnxJayPIQ/38hDr+D05mWNKE09glcOxWB9KUhhj3sS+MK/A2sz3OkdogVhn+ThxXXLd1/4crP/MUyAfYb/0/AqlCCgzxlSJyGSskkvEM8A1IjJARLphn4O4OCfzj0RkgJsfiPVtfZLE4TZ2zTx+LSK5InIMcAZNOJ8QuWefxX7AdXJfv99Idn2n+C7BthB2YL/mjZMbsc7/cb7653vnA2vpMFhrx0vY+/cSEclxv8NEZExTjicODwBXupaoiEhnsR0xipJc/zbgchHpk2hbIjJKRE5w56MK25oOu23MA04Tke5uO9cl2N9WYIjUdeTpLSJnuQ+Haux7K5xg/Xq0CYVijPkA+BLWbLEZ+9V5CNbJ9kUDqx2Cfaj3uf+FgNcz5yys/8X7IrkemCh1PRquxX7p7wL+CFxujJneBJHvwF7U8ViFNxHri3gZ+0AlhTHmVbetd4AV7t/PD9zxrQI+wL48/9EEOZPGGFODVcI3OxPOGcCPsA/2T4EzjDHbjTF7sef5GexD/DV8yjiJY4rHXRI9DmVOE0R/Emt+escYE9c8moA/Yo8tG3u8E7C+le3Yj4auvrrvYU2Qs3zzRVjfncf3gFtFZC/2g+OZRvb/ANa8Mx+YS+J7Zy/WZDJTRPZhFcki7DVKSGPXzLHFlW3CmoCvdK2gpnI19rxtwXZ8eJLGuzbvEpFy7AtwCnCm+6pfAvwZ+2W/Fesn/NC33mHY81HujudaY8wqd7wnAxe649mC/cLPa8bxRDDGzMZ2ergLe65WYH1Xya6/EHu//KSRbeVhlc92J3svrE8PXGcSrK/kDSDSsy0O3gfBDhGZi9UH12PPSRn2uWlSq83r4aEoihIXsd3pH3O+mVRv+w/YjgHfTPW2lZanTbRQFEVpHzjT3HhnxpkMXAb8r7XlUlJDk2MCKYqi7AdFWDNXP6yZ6s9E+zaVNoyavBRFUZSUoCYvRVEUJSWkzeQlNr7N+9geCdnAf4wxv4qpk4cdzXootrfQBcaYNYm2W1JSYoYMGZIOkRVFUdotc+bM2W6M6ZnOfaTTh1INnGCMKRcbbvoDEXnVGOPvF38ZdgTucBG5ENt174JEGx0yZAizZ89On9SKoijtEBFJJirFfpE2k5frJ17uZnPcL9ZhcxZ2wBzAf4AT3UhsRVEUpY2RVh+KG1E8DxvI7E1jzMyYKv1xsXxc7J/d2KCMsdu5QkRmi8js0tLSdIqsKIqiNJO0KhRj8zJMwMbumSwijcYVamA79xtjJhljJvXsmVYToKIoitJMWqSXl7EJmt7FhjvxsxEXHE5shrquWOe8oiiK0sZIm0IRkZ7iMhO6iK3TsNEr/byATRwDcB423pIOjFEURWmDpLOXV1/gEbEpWAPAM8aYl0TkVmyU3xewoeP/JSIrsMHILkyjPIqiKEoaSZtCMTaX9SFxlv/SN11F/DwniqIoShtDR8oriqLsB+Gw4eOVO/jdK0sJhsJUB0NU1oRaW6xWQYNDKoqiNJMX5m/imic/i8z36JzL4zPXsa6sgjW3nR5ZXlUbIj8nqzVEbFG0haIoitJMZnwePS5uy54q1pVVRC1bsGEXo29+jSE3vMzuitqWFK/FUYWiKIrSTDrlRrc6akP1M+Yu37I3Mj1jRfsemK0KRVEUpZnUhqNHOQRDdfPeCAi/qSsUbt+jItSHoiiK0kyemLkuaj4nq+4bvaImxPqdFcxZuzOyTBWKoiiKEpe87AD9uxVw6ZFD+OXzi+lRmBsp+2zdLr7+UHT4Qn8Lpj2iJi9FUZRmIgLTxvTmG1OGkJsdiOouvGbHvnr1c7LbdzB1VSiKoijNpDZkImau/OwAa3fU9fDaWxWsV7+ypr7Tvj2hCkVRFKUZhMKGULhOoeypCvLa4i2R8j+8Fhu6ECpr2/eAR1UoiqIozcDrItwUM1ZlTZCaYPttpahCURRFaQaeQsnNSvwavfjwQXx10gCyAsK9761i5E2vUrq3uiVEbHG0l5eiKEozqHU9tnIaUShXnzCcvl0LeHXRlohfZcPOCnoW5aVdxpZGWyiKoijUDURMluc+2whAdpY1efXtmh+3nqdw/KPqw+007ZMqFEVROjy7K2oZddNr3PZqfUd6Q9z60hIANu2qBOCRb0/mwH5d6NMlWrHkBOxrtiBqxPz+SpyZqEJRFKXDs21vFTWhMPe+t7LJ6wbd6PeRvYt4+Zpj+OTnJ0aVe077jhCCRRWKoigdnpr9aDKEG1EOnskr6KsXDLfPJooqFEVROjzJduX1/Cz+MPSN6aLsgG2h+P0mVbWqUBRFUdodK0vLOefvHyWs8+maMobc8DJn3/0hAFc+NidSdsyIkoTriliF4vfDV7XTAY6qUBRF6dCc+Of3Gq3z0YodAMzfsBuAj1fZ+eNG9uT40b3q1X/r+mO5YNJAsgJ1gx79TnlVKIqiKO2Yk8f2BiAYx4YV8vk8wmGDpycaGoMyvFcRfzhvPCt/d1pk2e/OPYjTD+oLQHU7HS2vCkVRFAUoys8BoCrOy77GF3a+sjZEtusKHGqCc33CwGJ+d+5Bdh/aQlEURWm/FHeyCuWJmWvrlflbLftqgpFeYUNLCpu0j/wc+8rVFoqiKEo75vxJAwBYvqW8Xpk/V/y+6rrWxQ9OGN6kfeRGIhPXNlKzbaIKRVGUDs3Qks4AjO7ThZLCPHKz678W/Sav3ZVWGfz45JF065xbr24ivB5fbyze2lxxMxpVKIqidGg652VxouuplZ8ToDqOf8Nv8tpVUQM0HhSyIQZ0K2D19vrZHNsDqlAURenQ1AbrkmRt3VPFs59t5MEZq6KUyPqddZkYn5q1HiBuSyYZjhnRk5LCprVs2gqqUBRF6ZBUB0M89MFqqoKhSMTggd07AfDbl5fyzOwNkbq7K4OUFNpw815WxkmDuzdrvwU5WWwvr2HFNuurWVVazksLNjX7ODKJtOVDEZGBwKNAb8AA9xtj7oypMxV4HljtFj1rjLk1XTIpiqJ4PDhjNX98fTkAhw7qBsAlRwzm1y/aKMJ+x3lNMMThQ7vz1tKtVAfDDO9VyEEDujZrv4V5doDjSX+xAypLCnPZXl7DEcN6RJRWWyWdCbaCwI+MMXNFpAiYIyJvGmOWxNSbYYw5I41yKIqi1KO8OhiZ9loo3lgUiI69VRsy5GQJ+TlZVAfDdC2oq9dULpg8iF2VtTz6se2evL3c+mQqqkPQtF7IGUfaTF7GmM3GmLluei+wFOifrv0piqI0hSypC4vi+VAK8+q+sb0owhU1QdaVVZCbHYjUyw4kn0c+lv7FBdzy5QPrLa8Ktv3Bji3iQxGRIcAhwMw4xVNEZL6IvCoi9c+yXf8KEZktIrNLS0vTKKmiKB2FQKC+QvG3PEJhq0zG/vJ1AP732caIM31dWQX7QyCOQmoPo+fTrlBEpBD4L3CdMWZPTPFcYLAx5mDg/wHPxduGMeZ+Y8wkY8yknj17pldgRVE6BP/6eE1kemzfLgD08aXxDRlD6d7qyHxtyPCzU0cDyYe7bwqfuICTbZm0KhQRycEqk8eNMc/Glhtj9hhjyt30K0COiCSOBa0oipICdrqcJn+94ODIKHkvNArYOF1CdEsiP9s61IMpyLjoDaj0mLW6bL+32dqkTaGIHRL6ELDUGPOXBur0cfUQkclOnravphVFaTMcM6JnZAS7pzDAjoivjQn+6CmceBGJm8oFhw2Mmm8PSbfS2cvrKOASYKGIzHPLfg4MAjDG3AucB1wlIkGgErjQGNM+ky0ripKR5ATqvqsL8+teiY99so48n4IB6OJ8LIcP67Hf+82PGRjZHnwoaVMoxpgPgIRdIYwxdwF3pUsGRVGUxsjJjnbOr/rdaby7fBuXPTKbJ2aui5R9dvM0ijvl8PC3DmN4r/3v35ufE62stJeXoihtCjUAWEI+H0in3Ojv6kBAOHFMb44eXkKlazU8cfnhdOuci4gwdVQvBnTrtN8y+FtDw3p2bhcmL1UoitJBeHvpVobe+Aqfb93b5HV/98pSvv/EXOav35UGyVoeLxz9T08Z1WAdv4M+1vSVCkb2LopMTxhQ3C5MXqpQFKWD4IVMn7N2Z5PXvf/9Vby8YDNn3f0hO8qrG18hw/ESZOUmiBic5zNJ5TUzEGQi/NvMy8nSFoqiKG0HbzBdqBldXof5urhu3l2VMplaih3l1dz1zhdc8tBMVpaWE3T5TRKFoPePpPe3VlJFtA/FsL28OqFJctpf3uO+91amXI5UogpFUToI3rvTH8PKT6IsgjWhcCRwYXUanccfrdjO7orUZjOcvaaMQ3/7Fn9643NmfLGdE//8XsTklUihnHlwv8h0blbqTV5eyuHvH38AZftsPK9NCZT1qu372FsV/9plCqpQFKWD4L08b3t1Wb2yGV+UMv6WNxhyw8v1vpKrakNs2FnJXqdw0mWa2ba3iq89OJML7v84pdu96IFP6i3zRrrnZDXcEbUg12fySkMLJS87i0W//hLXTxvF6eOt8qpoQNmHwoZQ2DQ7qVdLkdnSKYqSMga7XB9F+fVHC3zqG6VdGeMcXrLZRkzywpKkq4Xi5WpftqXpnQYSEc+K5LVQEiXJyk+zDwVsMMqsgFDg9tWQsvbkzU6gADMBVSiK0kHwwoXEjn8AwOcviH2p1bqv+W9OGRK3PFXUpmD0eTziBWKsTcKH0tOZ+DrlZkVFIU4Hno8mVpkbY1i9fV/EJJaoE0EmkN6zpChKxuC9RL2w7Nv2VHHevR9z+bHD8L9zY7uveut5LZt0dW9tqkLZVx3k0zVlHDuiZ1yl4RHP0Z2MD2VQj05M//FU+ncrIDvNL/KeRVZ53ffeSiYPtZkgX1m4me89PheAA3raThFZ+xE2vyVQhaIo7Yi9VbU8O3cjwbDhsqOHRpU9P28jADv21bCrooYVpeWsK6vg5ucWRdWLVRjb9lpHsadQqtMQaRfqFFey3P3uCv4+fSX/vWoKhyZIx+tXiJ5TuybUuA8FYEhMAMd0MbpPFyYP7c7SzXUB2R/9eE1kemXpPgBG9Skik1GFoijtiPPu+ZjlbuDixYcPipi3yvbVRPkmlm/ZG+k6G0uswvByn/crLgAyp4Xi+XZ2VybuFTa4RyeG9OjMOYf057qnbVhBz4yXSSakEb0KI+caIDtQXzZ1yiuK0mIs942Cr/b5Oipq7Jf5BZNshNuqYLjBlkaswvCUkhduPVN8KF7WxIYUY2S7wTC9ivKizEVeqyXdpqymkJ+TRbXv3MczbzXWomptMudsKoqSUvzBBmP9IGu27+POtz+Pv16MwqgJhgkIdM5Nnw/ljcVbuOudFZH5N5dsjSqfvaYs4vuJpbGBmjWhMDnZAQZ1r4u/VVZhndyZ9ILOzwlQFQyzeNNurn3qM977vH52Wm2hKIrSKvzsvwt4ffEWaoJh7pluX9ZF+XYw3S0vLmbRxtgEqpaZq3dw6p0z+L+Xl/DMp+tZtmUvudkBAgEhNyuQFh/KFf+aw0cr61IhXf7o7EjCqfc/L+W8ez/mHx+ujlrnraXbgMaj9NYEw+RmBTh4YDHfPXYYAKtKrWkpk17Q+dlZhMKGKx+bw/PzNkWVeY2VRN2cMwH1oShKO2Ddjgp/z18Api8vZfryUqaN7R354vdaKImCDt/x1hcAUQ5ib72cLOHJWeu45sTh9aL07i8H9utCTTDMF86PULbPxgzbvLsSoMGglo2Z4GpDJvIiPnZkT+57fxU7vW64GfSC9kyLtcH6F8drhGVneC+vzDmbiqI0ixlflHLsH9/lR8/Mj1vuNx/FG9SYDN7Avn01IXZX1vLSgs3N2k48vG69J47uxdmH9I8s91wqAacpG0q725gJrjYUjpi2vOMvdQEuM6mF4o3GT+RLyiR545HZ0imK0ihLNtmWxPwNjYeWb65C8XpD3XHBBCC1eVU8RZGTFYgadLnT+Tm80eH+bsV+f0qiFko4bAj6QpZ4AxRfWbjF7TNzvvi99MM7XOspHpnUooqHmrwUpY3jvVob+oL34/lQ4tGjc26DLzPvRXbMiBIgtT29IpF/swNRUX1fXrCZLgU5rNthTV7z1teF3ff7cWJHl/vxcsJ7CqVv1wJG9i7k863WrJZJ3Yb31TQe+FFbKIqipBWvsZBMWHp/wMNY/vCV8Q2WedkF8yIxp1LX06vGN2o935fI6uNVO7jmyc9Ytd2+/Ef36RIp8++/OoEsNTHjTQpys3jjh8dFyjPpBR3v+l00eVDUvMbyUhQlrYRjzE8f3nACz33/KP7y1YM5d2L/qLLcrAA/PGlk3O2cOKYXB/brEr9sdG8A8l1LJZU9vSKBGrMkbpyx0r2ev6PuZerffyLlVhezK/6LOCeDTEjxxp3kxsidSS2qeGS2dIqiRKgJhuN+xXb2tTqyAkL/4gImDCzm3IkDKIoJapiTFaAgN/5jLyIceUCPyPw5h/RncA87dsOLNJydFSA7ICltodRF0g3ETWS1r8buq6YBJbJxV1VUWbxt5zaQwjeTfCjxFEpWIBAV6TiTWlTxyGzpFEWJMPKmV7nqsTn1lvu/1mNfkLFf/DlZkjjtre/FGzYm0sPK/yLLT1G62nDYcPtry5jy+3ci+4jXQtnuWij+45zny23/1tKtXPf0Z3H30VDek6umHkBxpxxy4oQ3aS36dMmvt6xrQQ7//NZhAPz1goM1OKSiKKnjDV8X4NpQmJWl5VFO6dgAi7F5PHKyAhw0oDjKMX3rWQfStcA66/0thImDurFo424guneRHdG9/y2UVdvL+fv0upS2OVkSt4WycZd1yvtbJd7AxIDYMRper61YGsp78rNTRvOTk0cljFLc0kwb2ztq/pnvTmFM3yKK8nNYc9vprSRV08gc9awoSpO47dVlnHLHjMhARKjv2O0Ux+R16OBuUY7pb0wZwlkTrK/FG6z4tcMH8Y0pgyMtFL8tPy87KyUmr1g/TG5WIGGqXX+ryGBTuAzrWZhwH4nynmSSMgFrcvQzeWj3hL3yMhFtoShKGyDeuI+563bGqRnNJUcMZkSvQi57ZDbQuM/gwskD6VGYy5cO7IOINGDyCkQFnmwusa2pnKwAiSxQfiVWEwqTkxUgKzY8QAx1Ji/9dm4J9CwrShvAP8bEM+MEGnmZAnTOy+bEMXWmlMZ6NXXKzeasCf0jvgzvKz7a5NX0Fko4bHhzydZopRDTQsnOkoTHVFkb4sZnF3D23R+yoayS3Aac+H5qQnZ/meR8T8S5h/RvvFIGowpFUdoA/nAcXs6Mxr7O/XhV/Q75E0f34syD+yVcz7MKxYTYN0YAACAASURBVDrlK2tD1IbCDUYAjuWjlTu4/NHZ3PbqMuas3cnbS7fy1fs+BuyASrBZC3sU2um+XfO55IjBUYpge3k1T85az7z1u3h54Ways4TzDh0AwKje8RNPrSurAKBbp9yk5Gxtbj5jbGuLsF+oyUtR2gD+gIGe76EJ+oQpw3rw0codUcEFH7r0sEbX83oV+V/sO8qrWbOjghG/eBWA/151JIcO7pZwO+XVdhT4ytJyvnLPR1Flf71gAgO7d4rkW5nx0+Pp2zWf7KwARw3vwadrdrK9vLpeBN6crACXTBnCf+dubDCkjJeh0UsOlul069w2FF9DaAtFUdoANaH6YzDGD+gaWeYNYIxN++tx/zcm8eq1xzQ5oZTnKPY7jLeXR4dn8XqCJcJrYcUOwgTo3jk3okwABnbvFJHzlHF9ufmMsZFxMH5ynLJL5NOJjJTPoAGM7Zm0tVBEZCDwKNAb2ynjfmPMnTF1BLgTOA2oAC41xsxNl0yK0hZ5bdFmbn9teWTeUygBEfKyA3x600kU5mbzl69OaHAbhXnZjOkbfxR8IryGid+0Fds5KrZrcjw8hRIvu2JzHeaePyg/J4vpy0s56+4PeeryI6LCy3ituWRkVPafdJq8gsCPjDFzRaQImCMibxpjlvjqnAqMcL/DgXvcv6Iojisfi/7GuvSfn3LhYQMpyM0iNytAlzR2LfWc5H5XSezgumQc9As27I7anp+COIMZY/HGycSTzYv/NX/9Lh6YsYprThwRqTPjC5v1MNNDlviZ8dPj26wCTJtCMcZsBja76b0ishToD/gVylnAo8b2ifxERIpFpK9bV1E6PA2FiX/q0/Utsv/jRvZk9tqd9O6SF1mWFdO39+7pK7n0qPimNg/Pd+OFpPfIyRIGdm/cv/Hto4Zy8tg+DO9VyKKNu3lm9noO6m9Nfv6eXq8v3hKlUDxnfKaNOUnEQF+q4rZGizjlRWQIcAgwM6aoP+B/Mja4ZVEKRUSuAK4AGDQoOvqmorRnGgtJHxurK9V8//jhfOXQAVFO7diPfS94YyI8k9eyLdFZF39x2ph6A/rikZ+TxfBedhDjuP5dGde/a1SZR6xJrTZkGNsMU5/SPNLerhKRQuC/wHXGmPhJrBvBGHO/MWaSMWZSz549UyugomQwibL3FeRk1YsmnGoCAanXQ6op3ZU9auL4TiA10X79PpiK2uicIrWhcEZFFG7vpPVMi0gOVpk8box5Nk6VjcBA3/wAt0xRFOLnF/eorA01uddWKshqxiDBhhRjKoIznjmhbixNOGY3taFwvRDwSvpI293oenA9BCw1xvylgWovAN8QyxHAbvWfKEodXnfhG08dzcyfn1ivvDVelfFaKMnkdY/H8N6JY3Elw2FDurP4118CYMLA4qiyYMho2JUWJJ1n+ijgEuAEEZnnfqeJyJUicqWr8wqwClgBPAB8L43yKEqbI+g+ubsW5NA7TnjzUX3ijxBPJ14oF2+UOsCctYnjitWGwpQU1g3au/yYoTz8rcOYOCjxgMhk6ZyXzeg+RZHz5VETCrdKK66jks5eXh/QyAeU6931/XTJoChtHc/k1dBX9hnjE4dOSQc3njqay44eSr/iAr45ZQhfvusDKmsSt1BqgmGK8nMigyJ/cXrqQ4zkZgfYETPoUk1eLYuqbkVJIx+u2M6Ym1/jhfmbGq8ch0i+9QYcy60R9DA7KxBx1Oe5LruJ8qOEw4a3lm5j9fZ9aZWrKD+b2Wt38vjMtQBUB0Ms3rSH7AxKotXe0TOtKGnkhXmbqKwNscCXYdBj574aKmqCcdaqw/M9eGFGYoMgtnYGP29Q4Q3/XchF93/CrooaNu+u5JYXFhN0sr+4wCrTUw7sA8TPTJgKbjp9LAO6FfDEzHUAbNxpE3NlawulxdDgkIqSRhJ1+z3kN29ywuhe/CNBkMZgTIKo568+ilWl+9i8u5I3l2xNagxHOil0QRnLq4N8vGoHn63fxSMfrWH68lJOGtObo0eUsGa7jfh719cO4fXFW+lZlJdok81mTN8uHDygmGVb7OgEL9/KqeP6pmV/Sn1UoShKGql1AxMbMgm9s2xbwvVjTV75OVmM7deFsf26ROU5aS26x0THra4NR7JGeg7y2lCYgFhT2enj0/tyz8sJROJ3RVp32kJpMdTkpShpIhgK86LznTz2ybqEdf8zZwPTl9dXLqV7q4A6k1emUx0MRWJseVFjakLhFov2W1kTYsPOSuat38WP/z0fSM3gSSU59EwrSprwkjvFIzYx1Y//PZ9L//lpvXpeDpBeXdJjJko1+6pDkWjE731eijGGRz9eE5UPPp0McnGwzr77w0iYl7YUGLKto2daUdKEPw7XAT07R5V9sGJ7UtvYVVELwPBeLT/eJFn8TvY1O/bRo9Aqv4c/WsOaHRVU1YaTCgCZCn508ii6dYqOTKwDG1uOpM60iFwrIl3ciPaHRGSuiJycbuEUpS3j2fCL8rIjdn2PRz5ak9Q2CnKzIlF1M5X/97VD+NP5B1NSmMfeqiD9fbG/yl3GxJvTMO4kHrnZAQZ0i47Wqz6UliNZ1f1tF9jxZKAbdgT8bWmTSlHaOOGw4T9zNgA2wOKGnZVRoej3ue7CsV/TsdQEwxn/QjxsSHfOO3SAy5wYisouubfatrBaspXgD2ff0vvu6CR7pr07+jTgX8aYxbROGCFFaRPc+tIS/vnhGqAup4k/da7XpXVPVZDz7/2o3voeNcGWc2jvL/k5WVQFQ9T6WmNeTveWVSjRCbvayvlrDyR7pueIyBtYhfK6y8DYMl42RWmDeKPCX7/uWG46w5p7/AEUvUF/obDh0zU2DlZJYX3Hu+0h1XhGw0xg865KXlm4hW2+/Ch1CqXlvj9jsx1qC6XlSPZMXwbcABxmjKkAcoFvpU0qRWnjZAWEcf27MKpPUeSLedveqohSic0PMm1sb+KNUawJtp1YVKe4AYRrdtSFWCmvsiavlgzQmBfTQsl0k2F7IuFVFpGJIjIRmOAWDXPzg9FBkYoSl7J9NbyzbFskhpT3xfyVez7mh0/PA+paKB5FedkRZVO6t5rdlfZF3JJjOPaXrxxqk3350/x6LZSW7LqbH9Oi027DLUdjSuHP7j8fOBRYgPWdjAdmA1PSJ5qitE0e+mAVAIs27gZs6HmPVxdtAeqHZCnKz6bajdU4++4P2birkreuP5YV28oZ169tpLAtyrPHuXNfbWTZnkgLpQVNXuqUbzUSKhRjzPEAIvIscKgxZqGbHwfcknbpFKUN4jnc/++ccQBMHtKd3KxAVO+n2hiTV+e8bGpCNmzJxl02qOHLC6zyOWls64dYSQZ/XC+PB2asBlrW7HTBpIF8sXUv5x86kD1VtXSLCQ+jpI9kzVajPGUCYIxZJCJj0iSTorRpakNhivKzueCwQYDtNnzimF6R1sntry1j465KhvcqZMW2cn5wwnA659lHsdoX88szHbWV4IZ9uuQzsHsB68sq65UN6t45zhrp4eCBxfz7yiNbbH9KHcm2BReKyIMiMtX9HsCavxRFiaE6GCYvxo7vN3H9ffpKAH76pVGs/v1p/OjkUeQ7P4k/RMnOihpEWj9EfbIU5Gbx9vVT45a1FT+Qsn8k20K5FLgKuNbNvw/ckw6BFKWts21PNZ1yoxVKqa8rrce4/l0j4ee9nmDeyHKoi+PVltAeVR2bRj8bRCQLeNUY81djzDnu91djTFULyKcobY6te6oYWhJt4rn3kkMBGOZb7ndUewqltLy+4mlLtHZ+FqV1abSFYowJiUhYRLoaY3a3hFCK0papDobqtVD6di3gKxMH8MmqHZFl/u6sXrgQr1eUhz8uVlth2tjejO5TRDBsuGf6Sq47aURri6S0EMmavMqxfpQ3gcioJWPMNWmRSlHaMNaHUr/xn58TiPTgAiKOeKgbjLenMlqhDOjW9hTKA9+YFJn+2SmjW1ESpaVJVqE8636KojRCVW2onlMeiFr2/k+OjxofUeAUyszVZdHbCmqEI6XtkJRCMcY8km5BFKU9sG1vFdv2VlNSVH/sg2fWmjy0O4N6RIdY93wuy11SqNPH9+XlBZvjtnQUJVNJSqGIyAjg98BY7Kh5AIwxw9Ikl6K0SbbtqcYYGD+guF6Z10KJ1wu4S74dZb7X+VB++qVRnHtIf4b3KkyfsIqSYpI1ef0T+BXwV+B4bGBI/XRSlBi80fDxxl14LZSY7L9AXbyvPZXByPonjmkbI+QVxSNZpVBgjHkbEGPMWmPMLcDp6RNLUdomXi6QeAEJvZ5f8coCrtmyZY/tja/xp5S2SLItlGoRCQBfiMjVwEZA2+KKEoMXoyteC+WUcX3ZXl7DSUm0PIoLEmdyVJRMJFmFci3QCbgG+A3W7PXNdAmlKG0VL8RKvBZGz6I8fjhtZKPb6F9c0KL5QxQlVSSrUMqMMeXY8ShJJdYSkX8AZwDbjDHj4pRPBZ4HVrtFzxpjbk1SHkXJSKqDnkJp/ohxjXultFWSVSj/EJEBwKfADOB9f/ThBngYuAt4NEGdGcaYM5KUQVEyHq+Fsj9Jnfzh3xWlLZHsOJTjRCQXOAyYCrwsIoXGmO4J1nlfRIakQkhFaSvUJujllSw9NH+H0kZJdhzK0cAx7lcMvIRtqewvU0RkPrAJ+LExZnEKtqkorUYiH0pjTP/xVBZs3M0pB/ZJtViK0iIka/KaDszBDm58xRhTk7h6UswFBhtjykXkNOA5IG4UORG5ArgCYNCgQSnYtaKkh5pg8xXKkJLODClpuURUipJqkr3rS4BbsTnkXxORt0TkN/uzY2PMHufoxxjzCpAjIiUN1L3fGDPJGDOpZ8+e+7NbRdlvjDFU1MT3c9Qk6DasKO2dpO56Y8wuYBW2R9Zm4ADg2P3ZsYj0EZc8QUQmO1l2JF5LUVqfv729grG/fJ3dFXWRgY0xTF++jb+9/QWwf055RWmrJOtDWQUsAz7AZmr8VmNmLxF5EuvALxGRDdjQLTkAxph7gfOAq0QkCFQCFxpj4gSlUJTM4vl5GwGbDKtrJzsA8a2l27j80dmROpq5UOmIJOtDGW6MaVIcbWPMRY2U34XtVqwobQovx3vIF5RrZ0VN3DqK0pFItl0+XETeFpFFACIyXkRuSqNcipKxxFMowVB041pT4SodkWQVygPAjUAtgDFmAXBhuoRSlEwm4JTFgg27IstWlpa3ljiKkjEkq1A6GWNmxSzT4bxKh+TQwd0AmL68FIDV2/fx0AerE62iKB2CZH0o20XkAMAAiMh52N5eitLhyHYO96pgCIBdzn8yeWh3vnP0UHoU5rWabIrSmiSrUL4P3A+MFpGN2O7DF6dNKkXJUBZt3M0/P1wDQHWt7afiDWa89sQRHDU87lAqRekQJBvLaxVwkoh0xprJKrA+lLVplE1RMo7bXl0WmfZaKIlyoChKRyLhEyAiXUTkRhG5S0SmYRXJN4EVwFdbQkBFySTG9e8amS7bZ01dNSGrWDTLotLRaewJ+BcwClgIXA68C5wPnGOMOSvNsilKxlETDFOYl82Vxx3Apl2VkWWgo+MVpTGT1zBjzEEAIvIg1hE/yBhTlXbJFCUD+WTVDsqrgxTkZFEbMoTCJpJUS01eSkensScgEqzIGBMCNqgyUToqK7aVs2TzHgDycuyjUx0MRVooeapQlA5OY0/AwSKyx/32AuO9aRHZ0xICKkprcfNzi/jqfR+z0Zm2dldan8lNp48h3ymPv775OTUpSKqlKO2BhE+AMSbLGNPF/YqMMdm+6S4tJaSitAb/+mQts1aX8dL8TUBdb64D+3WlT9d8AB6YsZp1ZRWA+lAURZ8ARWmEylqve7DXEhFOGdc3Ur6nstYt18dJ6dgkO7BRUToMuytrefSjNZF5z+m+fMteALID0YpjT6WNQqQKReno6BOgKDG8vXQrf37z88j8PdNX8tm6nRGTV2ya3pcX2ihE2RqyXungqEJRlBi8FkmJLybX9x6fGzF5FebFb9hryHqlo6MKRVGAYKguf5ynOF677hjm3HQSAGP6dqE2FCYgdflQfnP2uMg6f794YgtKqyiZifpQlA7PEzPX8fP/LQRs66O8us4n0iU/h/EDuhI2ht2VtVHhVc4/dAA3P7cIgCE9OtffsKJ0MFShKB2eRZt2R6Y9ZQJ1AxVXle5jwQZbp39xQb1y0BzyigKqUBQFYwy9ivJ44eqjmb58G+XVQXp3yScvOwuIVjL9ivMj036fSVF+TssJrCgZiioUpcOzpypIbnaAPl3zuXDyoHrlF00exJOz1gFElIzHd48dRlZAIgMdFaUjowpF6dCUVwd5eUHi5KP+7sCx8bpuPG1MWuRSlLaI9vJSOjTb91YDcJAvz0ks/uEl+blZDdZTlI6OKhSlQ+N1Eb7i2GEN1jlsaPfI9HeOHpp2mRSlraImL6VD441+T9RL64zx/Zg0uLv6SRSlEbSFonRo3l66FagfnysWVSaK0jiqUJQOzaw1ZQCM6F3YypIoSttHFYrSocnJCjCufxcG60h3RdlvVKEoHZrqYKje2BJFUZpH2hSKiPxDRLaJyKIGykVE/iYiK0RkgYhodD2lxamsCWkueEVJEel8kh4GTklQfiowwv2uAO5JoyyKEpd1ZRV0ytXOjoqSCtKmUIwx7wNlCaqcBTxqLJ8AxSLSN0F9RUk5e6uCDOxe0HhFRVEapTXb+v2B9b75DW5ZPUTkChGZLSKzS0tLW0Q4pWNQGwo3mDBLUZSm0SaMx8aY+40xk4wxk3r27Nna4ijthGAoTNhAblabeAwUJeNpzSdpIzDQNz/ALVOUFuHDlTsAyFGnvKKkhNZ8kl4AvuF6ex0B7DbGJA77qigpZOYqq1BOGtOrlSVRlPZB2ozHIvIkMBUoEZENwK+AHABjzL3AK8BpwAqgAvhWumRRFI9Q2ERywhusuWt4r6LWFUpR2glpUyjGmIsaKTfA99O1f0WJ5Z8frubXLy7himOH8fPTxlAbDJOtqXsVJWWo8VjpMLw4fxMA97+/ippgmDeXbqWiJtTKUilK+0EVitJhCPhywFcFQ+yprKUoX7sMK0qq0KdJ6TAEfKkXq2pDZAUCnDKudytKpCjtC22hKB2GWavrAjdU14apDYXJVR+KoqQMbaEoHYJQ2ETNn/63GeypCraSNIrSPtEWitIh8HLHXzX1APp1zY8ok7wcDV2vKKlCWyhKh8BTKD065/LOj6eyrqyCeet3MW2M+lAUJVWoQlHaPZU1Iab8/h0AtpfXkJ+TxcjeRYzsrQMaFSWVqMlLafds3VNFebU1cZ0wWsOsKEq6UIWitHuCYWvu+ttFhzB5aPdWlkZR2i+qUJR2T03Q9vDSLsKKkl7Uh6K0S7bsruLWlxYzolcRuS48fXZAv58UJZ2oQlHaHc98up7fvrzEdQ3eAkBRfjZDSjq3rmCK0s5RhaK0G8r21XD3uyt46IPV5GYHuOtrh7C7spaDBxQzsnddS0VRlPSgCkVpN9z83CJeXmhztL167TEc0LOwlSVSlI6FKhSl3eAlzvrs5ml065zbytIoSsdDbQBKu6FzXha9ivJUmShKK6EKRWkXLNq4mydnrY+EWFEUpeVRk5fSpnnv81J+/cJiVm3fB0BRfk4rS6QoHRdVKEqLs21PFX+fvpLdlbV8tHI7edlZXH7sMD5bu5MfThvJwO6dktrGQx+u5r73VgFw9oR+FHfK5cwJ/dItvqIoDaAKRWlRgqEwVz0+lzlrd9KnSz7GwLqyCm5+bhEAz362kbzsAKGw4XtTD+DSo4ayq6KGPl3z6ZRbd7t+7/G5zF67E4CXfnA04/p3bZXjURSlDlUoSovy+dZy5qzdyU++NIrvHz8cgKWb9/DkrHUcOrgb1z09j+qgi731zgr+9s4KAC6aPJDfnzs+sp2yfTWUFOby/NVH07+4oOUPRFGUeqhCUVqUytoQAAf26xJZNqZvF249axwAp4zrQ0CEtTsqeOiD1cxfv4slm/ewvqwyajtVtSGmjuqlykRRMghVKEqLUh20CiUvO36mRG/58F6F/P7cgwC46P5PqA6GMMYQChuyswJUB8Pk6ch3RckoVKEoLcK6HRXc8fbnDHIO97yc5JVBIACfrtrJ0BtfYUC3Am46fSw79tlEWYqiZA6qUJQW4T9z1vPs3I2R+aa0Lmav2RmZ3rCzko9WbgeseUxRlMxBFYrSInTtZEevHzq4G+MHdGV4r+TjbN3z9Yms3LaPsooaHpyxitqQoaQwj8OGaLIsRckkVKEoaaM6GCIgQk5WIOI7efw7hzfZVHXC6N6cMBrufOsLakOGqtqQ+k8UJQNRhaLsN5U1IR6YsYqi/GxOH9+Xsn01XPqPT9mypwqA40b2ZNMu20trfxSB53d5eeFmBmjvLkXJONKqUETkFOBOIAt40BhzW0z5pcAfAc+4fpcx5sF0yqSkjqraEC8v2MyP/j0/suzXLy6JTPfpks+WPVW893kpAN065SDS/DS8x4/qxVOz1rFmR0WTTGaKorQMaVMoIpIF3A1MAzYAn4rIC8aYJTFVnzbGXJ0uOZTU8fHKHTz96TpysgL8e86GyHIRO66kf3EB05eX0rtLPvddcihj+tqxJi/O38S89bs48+D9C4syqk8Rb11/HKu272Nwj8bDsyiK0rKks4UyGVhhjFkFICJPAWcBsQpFyVDWbN/Hpt2VZIlww7MLWe0CMBZ3qgvA+M0pg7nxtDEJ/SJfPrgfX95PZeKRnRVgZO+ilGxLUZTUkk6F0h9Y75vfABwep95XRORY4HPgh8aY9bEVROQK4AqAQYMGpUHUjkFNMEx2QNiws5I3lmxhwsBixvTtwubdlUxfXsqxI3vy2CdreWfZNjbsrKy3/sjehTx5+RF0d/lGNu2u0pHqiqJEaG2n/IvAk8aYahH5LvAIcEJsJWPM/cD9AJMmTTItK2Lbpbw6yFOz1jGydxEfrtjOfe+vSrzCy0ujZicMLGZYz84c0LOQo4aXMKp3EQW5dS0RVSaKovhJp0LZCAz0zQ+gzvkOgDFmh2/2QeD2NMrTbgmHDYa6FLgAO/fVcMhv3oxbv1NuFmdN6M/A7gXs3FfDsi176VmUR3FBLlXBEGeM78uRB5S0kPSKorQX0qlQPgVGiMhQrCK5EPiav4KI9DXGbHazZwLRn8hKQkJhw0sLNvH7V5axs6KGU8f1wQDPz9sUVe/cif351pFDGduvS5TSURRFSSVpUyjGmKCIXA28ju02/A9jzGIRuRWYbYx5AbhGRM4EgkAZcGm65GmLGGMIhg05WXb8xaKNu/nVC4tZsmkPI3oXsmDD7qj6z/kUycDuBVw/bSTnHDKgRWVWFKXjIsa0LZfEpEmTzOzZs1tbjLSyo7yaP73xOU/OWgdAfk6A3KwAe6qC9epeMGkg/3fOOJZt2csfXlvG148YzKGDu9Gjc+5+jflQFKV9ISJzjDGT0rmP1nbKd3g++GI789bvpFNuNseMKOG9z0v5rXOOD+reiXVlFVTVhqmqtUmnHv/O4YzoVcif3ljOuP5dueSIwYgI4/p35V+XxetEpyiK0jKoQmklaoJhfvj0PF5euDlu+Z0XTuCsCf0Ba+o6oGchWQEh14Uuuf28g1tMVkVRlGRQhdKCzFm7k1tfXMyWPVVs3VMdWX7/JYfy1tKt7KkMMrJPEVcdd0BU91zNl64oSltAFUqK2bCzgrvfXcmxI0o4ZVwfRARjDBc/OJOPVu6IqjttbG/u/tpEcrMDnHyg5vZQFKVt0yEVitcRwUsn21xqQ2EWbdzN4B6d6dYphysfm8Pri7cC8OSsdUwa3I1pY3tz17sr2Osc6n88bzxfPrifZhtUFKXd0eEUyu9eWcq7y7ZRFQzRv7iAp66Y0ug6Nz67gGfnbuRHJ4/kd68sA+oi6cbj6OElfLZuJ7PX2h/AWRP6cdHkQRwxrEfqDkZRFCWD6FAK5d1l27jfF35kfVn9eFV+KmqC/O+zjTw5y4YX85QJEFeZXHfSCK4+fnik1TN33U6mL9vGkcNLVJEoitLu6TAK5fXFW/juv+Ywuk8R35gyhJ//byGdneO7qjbEr55fzBfb9lJZG6Z/cT5b91SzcGPdwME/njeev775OddNG8npB/Vl654qSory6JKf09AumTioGxMHdUv7sSmKomQCHUah9C8u4Ipjh/H1wwczqEcnVpaW89SsdXzwxXa+/tDMqLpLN++Jmn/h6qMYP6CY8yfVhSYb1lMTPCmKovjpMAplXP+uUd1v83MC7KsJRZTJ0cNL+NK4Plw8eRBPz15PTlaA0w/qG9V9V1EURWmYDqNQYjlqeAl3v7sSgN+ePY6vHzE4UnbRZM25oiiK0lQ6rEKZMqwHPztlNNkBiVImiqIoSvPosApFRLhq6gGtLYaiKEq7ofmj+hRFURTFhyoURVEUJSWoQlEURVFSgioURVEUJSWoQlEURVFSgioURVEUJSWoQlEURVFSgioURVEUJSWIl2yqrSAipcA+YHtry5IEJaicqUTlTC0qZ2rJdDkHG2N6pnMHbU6hAIjIbGPMpNaWozFUztSicqYWlTO1tBU504mavBRFUZSUoApFURRFSQltVaHc39oCJInKmVpUztSicqaWtiJn2miTPhRFURQl82irLRRFURQlw1CFoiiKoqSENqVQROQUEVkuIitE5IYW2uc/RGSbiCzyLesuIm+KyBfuv5tbLiLyNyffAhGZ6Fvnm67+FyLyTd/yQ0VkoVvnbyIizZRzoIi8KyJLRGSxiFybibKKSL6IzBKR+U7OX7vlQ0Vkptv20yKS65bnufkVrnyIb1s3uuXLReRLvuUpu09EJEtEPhORlzJVThFZ467LPBGZ7ZZl1HV32ykWkf+IyDIRWSoiUzJUzlHuXHq/PSJyXSbKmnEYY9rED8gCVgLDgFxgPjC2BfZ7LDARWORbdjtwg5u+AfiDmz4NeBUQ4AhgplveHVjlMSnVbwAAB9dJREFU/ru56W6ubJarK27dU5spZ19gopsuAj4HxmaarG7dQjedA8x023wGuNAtvxe4yk1/D7jXTV8IPO2mx7p7IA8Y6u6NrFTfJ8D1wBPAS24+4+QE1gAlMcsy6rq77TwCfMdN5wLFmShnjMxZwBZgcKbLmgm/VhegCRd2CvC6b/5G4MYW2vcQohXKcqCvm+4LLHfT9wEXxdYDLgLu8y2/zy3rCyzzLY+qt58yPw9My2RZgU7AXOBw7Ajj7NhrDbwOTHHT2a6exF5/r14q7xNgAPA2cALwkttvJsq5hvoKJaOuO9AVWI3rCJSpcsaR+2Tgw7Ygayb82pLJqz+w3je/wS1rDXobYza76S1AbzfdkIyJlm+Is3y/cOaWQ7Bf/xknqzMjzQO2AW9iv9R3GWOCcbYdkceV7wZ6NEP+5nAH8FMg7OZ7ZKicBnhDROaIyBVuWaZd96FAKfBPZ0J8UEQ6Z6CcsVwIPOmmM13WVqctKZSMxNhPjIzpey0ihcB/geuMMXv8ZZkiqzEmZIyZgG0BTAZGt7JI9RCRM4Btxpg5rS1LEhxtjJkInAp8X0SO9RdmyHXPxpqO7zHGHIKNxxflN8oQOSM4/9iZwL9jyzJN1kyhLSmUjcBA3/wAt6w12CoifQHc/za3vCEZEy0fEGd5sxCRHKwyedwY82wmywpgjNkFvIs1/xSLSHacbUfkceVdgR3NkL+pHAWcKSJrgKewZq87M1BOjDEb3f824H9YJZ1p130DsMEYM9PN/werYDJNTj+nAnONMVvdfCbLmhm0ts0t2R/2C2cVtunsOTEPbKF9DyHah/JHop1zt7vp04l2zs1yy7tj7cfd3G810N2VxTrnTmumjAI8CtwRszyjZAV6AsVuugCYAZyB/Qr0O7u/56a/T7Sz+xk3fSDRzu5VWAdqyu8TYCp1TvmMkhPoDBT5pj8CTsm06+62MwMY5aZvcTJmnJw+eZ8CvpWpz1Im/lpdgCZe4NOwvZdWAr9ooX0+CWwGarFfWZdhbeNvA18Ab/luEgHudvItBCb5tvNtYIX7+W/SScAit85dxDgtmyDn0dgm+AJgnvudlmmyAuOBz5yci4BfuuXD3EO2AvvSznPL8938Clc+zLetXzhZluPrJZPq+4RohZJRcjp55rvfYm87mXbd3XYmALPdtX8O+5LNODndtjpjW5hdfcsyUtZM+mnoFUVRFCUltCUfiqIoipLBqEJRFEVRUoIqFEVRFCUlqEJRFEVRUoIqFEVRFCUlqEJRWg0RMSLyZ9/8j0XklhRt+2EROS8V22pkP+e7yLnv+pYd5ItUWyYiq930WyJypqQxUraInC0iY9O1fUVJRHbjVRQlbVQD54rI740x21tbGA8RyTZ18boa4zLgcmPMB94CY8xC7JgLRORh7BiW//jWeSFVssbhbGwgyyVp3IeixEVbKEprEsTm4f5hbEFsC0NEyt3/VBF5T0SeF5FVInKbiFwsNsfKQhE5wLeZk0Rktoh87mJzeYEp/ygin7rcFd/1bXeGiLxAnJexiFzktr9IRP7glv0SO6D0IRH5YzIHLCKXishdvmO8R0Q+cccyVWz+naVOEXnrnCwiH4vIXBH5t4vXhjv2Je44/iQiR2JjT/3RtYgOcL/XXODIGSIy2rfve+OcnwPduZzntjsimeNSFNAWitL63A0sEJHbm7DOwcAYoAwbvuRBY8xksUnFfgBc5+oNwca1OgB4V0SGA98AdhtjDhORPOBDEXnD1Z8IjDPGrPbvTET6AX8ADgV2YiP7nm2MuVVETgB+bIyZ3eQjt3TDxjI7E9tyOQr4DvCpiEzARme4CTjJGLNPRH4GXC8idwPnAKONMUZEio0xu5xCjLSIRORt4EpjzBcicjjwd2xcsobOz5XAncaYx11wxKxmHpfSAVGForQqxpg9IvIocA1QmeRqnxoXRlxEVgKeQlgIHO+r94wxJgx8ISKrsFGNTwbG+1o/XYERQA02BlOUMnEcBkw3xpS6fT6OTbz2XJLyJuJFpxAWAluduQwRWYx94Q/AJun6UGxSv1zgY2x4/Cps6+glrJkrCteSORL4t9QlBMzzVYl3fj4GfiEiA4BnjTFfpOAYlQ6CKhQlE7gDm2jrn75lQZxJVkQC2BepR7VvOuybDxN9T8fGFTLYuEs/MMa87i8QkanYkOotjV/22OPKBkLAm8aYi2JXFJHJwInAecDV1LU8PALY/C0TGth3vfNjjHlCRGZiAx6+IiLfNca805QDUjou6kNRWh1jTBk2te5lvsVrsCYmsOagnGZs+nwRCTi/yjBscMbXgavEhvpHREaKTfSUiFnAcSJSIiJZ2Ax77zVDnubwCXCUM0chIp2dzIXYwIWvYH1QB7v6e7EpoDE2H85qETnfrSsicrBv2/XOj4gMA1YZY/6Gzfo5vgWOUWknqEJRMoU/AyW++QewL/H5WB9Dc1oP67DK4FWsH6EKeBDrdJ8rIouwaVkTttSdee0GbO6W+cAcY8zzzZCnyTgz26XAkyKyAGuSGo1VGi+5ZR8A17tVngJ+IjYr4gHAxcBl7jwuBs7ybT7e+fkqsEhsRs1x2JQIipIUGm1YUTogEr87s6LsF9pCURRFUVKCtlAURVGUlKAtFEVRFCUlqEJRFEVRUoIqFEVRFCUlqEJRFEVRUoIqFEVRFCUl/H/4IB1ExLTJDwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# reward function value curve\n",
    "plot_results(log_dir, \"PPO 5x5 Random Road Env Reward Shaping Baseline Results\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "wUQDE24De9Un"
   },
   "outputs": [],
   "source": [
    "# outputting the logged reward values\n",
    "!cat /tmp/gym/monitor.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}