DQN-Based Intelligent Scheduling in Wireless Sensor Networks Using Stable Baselines3.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Environment Setup and Library Imports\n",
        "\n",
        "This section of the code installs and imports the necessary libraries for creating and training a Deep Q-Network (DQN) in the Wireless Sensor Network (WSN) environment. The steps include:\n",
        "\n",
        "- **Installation of Stable Baselines3**: The `stable-baselines3[extra]` package is installed to provide the DQN algorithm and other tools needed for reinforcement learning.\n",
        "- **Importing Libraries**:\n",
        "  - `gymnasium`: A toolkit for developing and comparing reinforcement learning algorithms.\n",
        "  - `spaces` from `gymnasium`: For defining action and observation spaces.\n",
        "  - `numpy`: For numerical operations.\n",
        "  - `random`: For generating random numbers.\n",
        "  - `scipy.stats.poisson`: For generating sensor positions based on a Poisson distribution.\n",
        "  - `DQN` from `stable_baselines3`: For implementing the DQN algorithm.\n",
        "  - `check_env` from `stable_baselines3.common.env_checker`: For validating the custom environment.\n",
        "\n",
        "These imports set up the foundation for creating the WSN environment and training the DQN model.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dr4UZOaLPIt-"
      },
      "outputs": [],
      "source": [
        "!pip install stable-baselines3[extra]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TTYOuy7EwWD0",
        "outputId": "3482e1e6-3348-4b4c-d561-5a0d0c62e120"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torch/utils/tensorboard/__init__.py:4: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
            "  if not hasattr(tensorboard, \"__version__\") or LooseVersion(\n"
          ]
        }
      ],
      "source": [
        "import gymnasium\n",
        "from gymnasium import spaces\n",
        "import numpy as np\n",
        "import random\n",
        "from scipy.stats import poisson\n",
        "from stable_baselines3 import DQN\n",
        "from stable_baselines3.common.env_checker import check_env"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Custom Wireless Sensor Network (WSN) Environment Definition\n",
        "\n",
        "This section of the code defines a custom environment for simulating a Wireless Sensor Network (WSN) using the `gymnasium` library. The environment is designed to support the development and testing of scheduling algorithms, particularly for reinforcement learning. The key components include:\n",
        "\n",
        "- **Initialization (`__init__`)**: \n",
        "  - Sets up the environment parameters such as the number of sensors, sensor coverage, maximum steps per episode, and event generation threshold.\n",
        "  - Initializes the sensor positions and internal state variables.\n",
        "  - Defines the observation and action spaces.\n",
        "\n",
        "- **Environment Reset (`reset`)**: \n",
        "  - Resets the environment at the start of each episode, including resetting the internal state variables and the sensor selection counts.\n",
        "\n",
        "- **Step Function (`step`)**: \n",
        "  - Defines the environment's response to an agent's action, including event generation, sensor selection, reward calculation, and termination condition checks.\n",
        "\n",
        "- **Sensor Position Generation (`_generate_sensor_positions`)**: \n",
        "  - Generates random sensor positions within the environment based on a Poisson distribution.\n",
        "\n",
        "- **Render and Close Methods**: \n",
        "  - Placeholder methods for rendering and closing the environment.\n",
        "\n",
        "This custom environment serves as the foundation for testing and developing intelligent scheduling strategies in a WSN context, with a focus on reinforcement learning applications.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aRQ3jQvTwZ4M",
        "outputId": "5e4b6f3b-6aec-4dc1-bc58-57a6a8a7195a"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/ipykernel/ipkernel.py:283: DeprecationWarning: `should_run_async` will not call `transform_cell` automatically in the future. Please pass the result to `transformed_cell` argument and any exception that happen during thetransform in `preprocessing_exc_tuple` in IPython 7.17 and above.\n",
            "  and should_run_async(code)\n"
          ]
        }
      ],
      "source": [
        "class WSNEnvironment(gymnasium.Env):\n",
        "\n",
        "    metadata = {\"render_modes\": [\"console\"]}\n",
        "    def __init__(self, num_sensors=13, sensor_coverage=0.4, max_steps=1000, threshold_prob=0.3):\n",
        "        super(WSNEnvironment, self).__init__()\n",
        "\n",
        "        # Environment parameters\n",
        "        self.num_sensors = num_sensors\n",
        "        self.sensor_coverage = sensor_coverage\n",
        "        self.max_steps = max_steps\n",
        "        self.threshold_prob = threshold_prob\n",
        "\n",
        "        # Initialize the environment\n",
        "        self.sensor_positions, self.num_points = self._generate_sensor_positions()\n",
        "\n",
        "        # Define observation space\n",
        "        self.observation_space = spaces.Box(low=0, high=1000, shape=(self.num_points, 3), dtype=np.float32)\n",
        "\n",
        "        # Define action space\n",
        "        self.action_space = spaces.Discrete(self.num_points)  # Select sensor index\n",
        "\n",
        "        # Internal state variables\n",
        "        self.step_count = 0\n",
        "\n",
        "        # information dictionary\n",
        "        self.info = {}\n",
        "\n",
        "        # the generated event\n",
        "        self.event = None\n",
        "\n",
        "        # Number of generated events\n",
        "        self.generated_events = 0\n",
        "\n",
        "    def reset(self, seed=0):\n",
        "        # Initialize the environment at the start of each episode\n",
        "        self.step_count = 0\n",
        "        self.generated_events = 0\n",
        "        self.info = {'captured': 0, 'non-captured': 0}\n",
        "\n",
        "        for i in range(self.num_points):\n",
        "            self.info['sensor '+str(i)] = set()\n",
        "            # reset the number of selection for each sensor to 0\n",
        "            self.sensor_positions[i,-1] = 0\n",
        "        return self.sensor_positions, self.info\n",
        "\n",
        "    def step(self, action):\n",
        "        # Execute one step in the environment\n",
        "        truncated = bool(self.step_count > self.max_steps)\n",
        "        terminated = False\n",
        "        reward = 0\n",
        "\n",
        "        # Check termination condition\n",
        "        if truncated:\n",
        "            terminated = True\n",
        "            return self.sensor_positions, reward, terminated, truncated, self.info\n",
        "\n",
        "        self.event = None\n",
        "\n",
        "        # we flip a coin to get a probability and if it exceeds threshold we generate an event\n",
        "        prob_flip_coin = random.uniform(0, 1)\n",
        "\n",
        "        if prob_flip_coin >= self.threshold_prob:\n",
        "            self.event = 1\n",
        "        else:\n",
        "            self.event = 0\n",
        "\n",
        "        if self.event == 1:\n",
        "\n",
        "            self.generated_events += 1\n",
        "\n",
        "            # Generate event location\n",
        "            event_location = np.random.uniform(low=0, high=1, size=(2,))\n",
        "\n",
        "            # detect the sensors near that event\n",
        "            for i in range(self.num_points):\n",
        "                # calculate the distance between the sensor and the event\n",
        "                distance = np.linalg.norm(event_location - self.sensor_positions[i,:2])\n",
        "\n",
        "                # adding the event to the sensor events buffer in case its under the coverage\n",
        "                # the event is represented with a tuple (event, the step count in the episode)\n",
        "                if distance <= self.sensor_coverage:\n",
        "                    self.info['sensor '+str(i)].add((self.event, self.step_count))\n",
        "\n",
        "            # check the sensor buffer of the selected action\n",
        "            selected_buffer = self.info['sensor '+str(action)]\n",
        "            number_captured_events = len(selected_buffer)\n",
        "\n",
        "            #update the observation space\n",
        "            self.sensor_positions[action, 2] += 1\n",
        "\n",
        "            if  number_captured_events == 0:\n",
        "              reward = -0.4\n",
        "\n",
        "            else:\n",
        "              #calculate the reward\n",
        "              temp = 0\n",
        "              # looping through events in the selected buffer's set\n",
        "              for x in selected_buffer:\n",
        "                #age of info\n",
        "                AoI = self.step_count - x[-1]\n",
        "                if AoI == 0: temp = 1\n",
        "                else: temp += 1/AoI\n",
        "\n",
        "\n",
        "              reward = temp\n",
        "\n",
        "              # remove the selected events from the other buffers\n",
        "              for i in range(self.num_points):\n",
        "                #skip the selected sensor's buffer\n",
        "                if i != action :\n",
        "                  buffer = self.info['sensor '+str(i)]\n",
        "                  buffer.difference_update(selected_buffer)\n",
        "\n",
        "              #clear the buffer of the selected sensor\n",
        "              selected_buffer.clear()\n",
        "\n",
        "              #update the info dict\n",
        "              self.info['captured'] += number_captured_events\n",
        "\n",
        "        #no generated event\n",
        "        else:  reward = 0\n",
        "\n",
        "        # calculate the non-captured event in case the episode is done\n",
        "        if self.step_count == self.max_steps:\n",
        "            self.info['non-captured'] = self.generated_events - self.info['captured']\n",
        "\n",
        "        self.step_count += 1\n",
        "\n",
        "        # Return the next observation, reward, termination signal, and additional information\n",
        "        return self.sensor_positions, reward, terminated, truncated, self.info\n",
        "\n",
        "    def _generate_sensor_positions(self):\n",
        "\n",
        "        # Set a fixed random seed for reproducibility\n",
        "        np.random.seed(42)\n",
        "\n",
        "        # Define the intensity (lambda) for the Poisson process\n",
        "        intensity = self.num_sensors / (1 * 1)  # Adjust as needed based on the size of the space\n",
        "\n",
        "        # Generate the number of points based on a Poisson distribution\n",
        "        num_points = poisson(intensity).rvs()\n",
        "\n",
        "        # Generate the sensor positions uniformly at random within the space\n",
        "        sensor_positions = np.random.uniform(low=0, high=1, size=(num_points, 2))\n",
        "\n",
        "        # Create a new column of zeros with the same number of sensors to calculate the number of pulling for each one\n",
        "        zeros_column = np.zeros((num_points, 1))\n",
        "\n",
        "        # Concatenate the original array with the new column of zeros\n",
        "        sensor_positions_with_zeros = np.hstack((sensor_positions, zeros_column))\n",
        "\n",
        "\n",
        "        return sensor_positions_with_zeros.astype(np.float32), num_points\n",
        "\n",
        "    def render(self):\n",
        "        pass\n",
        "\n",
        "    def close(self):\n",
        "        pass"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xjZ01rtSweNW"
      },
      "outputs": [],
      "source": [
        "env = WSNEnvironment()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kOPr_m8SJzHp",
        "outputId": "8a62bd98-b3cb-4f62-cc5c-0535e0503706"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([[0.7319939 , 0.5986585 , 0.        ],\n",
              "       [0.15601864, 0.15599452, 0.        ],\n",
              "       [0.05808361, 0.8661761 , 0.        ],\n",
              "       [0.601115  , 0.7080726 , 0.        ],\n",
              "       [0.02058449, 0.96990985, 0.        ],\n",
              "       [0.83244264, 0.21233912, 0.        ],\n",
              "       [0.18182497, 0.1834045 , 0.        ],\n",
              "       [0.30424225, 0.52475643, 0.        ],\n",
              "       [0.43194503, 0.29122913, 0.        ],\n",
              "       [0.6118529 , 0.13949387, 0.        ],\n",
              "       [0.29214466, 0.36636186, 0.        ],\n",
              "       [0.45606998, 0.785176  , 0.        ]], dtype=float32)"
            ]
          },
          "execution_count": 5,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "env.sensor_positions"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Baseline Random Scheduler Simulation\n",
        "\n",
        "This section of the code simulates a baseline random scheduler in the custom Wireless Sensor Network (WSN) environment. The steps involved are:\n",
        "\n",
        "- **Initialization of simulation parameters**, including the number of episodes and a list to store scores.\n",
        "- **Execution of the random scheduler**:\n",
        "  - For each episode, the environment is reset, and actions are selected randomly.\n",
        "  - The agent accumulates rewards based on the actions taken and the events captured by the sensors.\n",
        "  - The episode ends when the termination condition is met (maximum steps reached).\n",
        "- **Recording and printing of scores**:\n",
        "  - After each episode, the total score is recorded and printed along with information about captured and non-captured events.\n",
        "- **Calculation and display of the average score** across all episodes, providing a baseline measure of performance.\n",
        "\n",
        "This baseline simulation is crucial for comparing the performance of more advanced scheduling algorithms, such as those utilizing reinforcement learning.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nICw2i2YzctR",
        "outputId": "d06fd7c2-e4fb-43ac-ec90-e948555e4093"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Episode:1\t Score:272.79 \t{'captured': 668, 'non-captured': 10, 'sensor 0': {(1, 986), (1, 994), (1, 978), (1, 970), (1, 999)}, 'sensor 1': {(1, 1000)}, 'sensor 2': set(), 'sensor 3': {(1, 986), (1, 970), (1, 999), (1, 994)}, 'sensor 4': set(), 'sensor 5': {(1, 994), (1, 978)}, 'sensor 6': {(1, 1000)}, 'sensor 7': {(1, 1000), (1, 999)}, 'sensor 8': {(1, 999), (1, 1000)}, 'sensor 9': set(), 'sensor 10': {(1, 999), (1, 1000)}, 'sensor 11': {(1, 986), (1, 999)}}\n",
            "Episode:2\t Score:301.38 \t{'captured': 682, 'non-captured': 6, 'sensor 0': {(1, 999), (1, 1000)}, 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': {(1, 1000)}, 'sensor 4': set(), 'sensor 5': {(1, 999)}, 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 999)}, 'sensor 10': set(), 'sensor 11': {(1, 1000)}}\n",
            "Episode:3\t Score:304.80 \t{'captured': 704, 'non-captured': 8, 'sensor 0': {(1, 1000)}, 'sensor 1': {(1, 999)}, 'sensor 2': {(1, 1000), (1, 994)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 994)}, 'sensor 5': set(), 'sensor 6': {(1, 999)}, 'sensor 7': {(1, 999), (1, 1000)}, 'sensor 8': {(1, 999), (1, 1000)}, 'sensor 9': {(1, 999)}, 'sensor 10': {(1, 999), (1, 1000)}, 'sensor 11': {(1, 1000), (1, 994)}}\n",
            "Episode:4\t Score:289.07 \t{'captured': 710, 'non-captured': 6, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': set()}\n",
            "Episode:5\t Score:290.38 \t{'captured': 697, 'non-captured': 3, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': {(1, 999)}, 'sensor 3': {(1, 999)}, 'sensor 4': {(1, 999)}, 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': {(1, 999)}, 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 999)}}\n",
            "Episode:6\t Score:291.39 \t{'captured': 705, 'non-captured': 5, 'sensor 0': set(), 'sensor 1': {(1, 993)}, 'sensor 2': {(1, 1000)}, 'sensor 3': set(), 'sensor 4': {(1, 1000)}, 'sensor 5': set(), 'sensor 6': {(1, 993)}, 'sensor 7': set(), 'sensor 8': {(1, 993)}, 'sensor 9': set(), 'sensor 10': {(1, 993)}, 'sensor 11': {(1, 1000)}}\n",
            "Episode:7\t Score:310.45 \t{'captured': 717, 'non-captured': 7, 'sensor 0': {(1, 1000), (1, 989)}, 'sensor 1': {(1, 998)}, 'sensor 2': set(), 'sensor 3': {(1, 1000), (1, 989)}, 'sensor 4': set(), 'sensor 5': {(1, 997)}, 'sensor 6': {(1, 998)}, 'sensor 7': {(1, 998), (1, 1000)}, 'sensor 8': {(1, 997), (1, 998), (1, 1000)}, 'sensor 9': {(1, 997)}, 'sensor 10': {(1, 998), (1, 1000)}, 'sensor 11': {(1, 1000)}}\n",
            "Episode:8\t Score:272.50 \t{'captured': 680, 'non-captured': 1, 'sensor 0': set(), 'sensor 1': {(1, 1000)}, 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': {(1, 1000)}, 'sensor 7': {(1, 1000)}, 'sensor 8': {(1, 1000)}, 'sensor 9': set(), 'sensor 10': {(1, 1000)}, 'sensor 11': set()}\n",
            "Episode:9\t Score:274.23 \t{'captured': 713, 'non-captured': 9, 'sensor 0': {(1, 985), (1, 997), (1, 994)}, 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': {(1, 985), (1, 997), (1, 994)}, 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 985), (1, 997), (1, 994)}}\n",
            "Episode:10\t Score:274.84 \t{'captured': 690, 'non-captured': 9, 'sensor 0': set(), 'sensor 1': {(1, 1000)}, 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': {(1, 999)}, 'sensor 6': {(1, 1000)}, 'sensor 7': {(1, 1000)}, 'sensor 8': {(1, 1000)}, 'sensor 9': {(1, 999)}, 'sensor 10': {(1, 1000)}, 'sensor 11': set()}\n",
            "Episode:11\t Score:272.78 \t{'captured': 711, 'non-captured': 8, 'sensor 0': {(1, 993)}, 'sensor 1': set(), 'sensor 2': {(1, 996), (1, 994)}, 'sensor 3': {(1, 996), (1, 993)}, 'sensor 4': {(1, 996), (1, 994)}, 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': {(1, 996), (1, 993)}, 'sensor 8': {(1, 993)}, 'sensor 9': set(), 'sensor 10': {(1, 993)}, 'sensor 11': {(1, 996), (1, 993)}}\n",
            "Episode:12\t Score:276.68 \t{'captured': 695, 'non-captured': 7, 'sensor 0': {(1, 994), (1, 990), (1, 983)}, 'sensor 1': set(), 'sensor 2': {(1, 998)}, 'sensor 3': {(1, 998), (1, 994), (1, 990), (1, 983)}, 'sensor 4': set(), 'sensor 5': {(1, 995)}, 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': {(1, 995)}, 'sensor 9': {(1, 995)}, 'sensor 10': set(), 'sensor 11': {(1, 998)}}\n",
            "Episode:13\t Score:289.40 \t{'captured': 693, 'non-captured': 8, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': {(1, 990), (1, 1000)}, 'sensor 3': set(), 'sensor 4': {(1, 990), (1, 1000)}, 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': {(1, 1000)}, 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': {(1, 1000)}, 'sensor 11': {(1, 1000)}}\n",
            "Episode:14\t Score:288.80 \t{'captured': 700, 'non-captured': 10, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': set()}\n",
            "Episode:15\t Score:300.63 \t{'captured': 734, 'non-captured': 8, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': set()}\n",
            "Episode:16\t Score:280.93 \t{'captured': 699, 'non-captured': 6, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': set(), 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': set()}\n",
            "Episode:17\t Score:277.95 \t{'captured': 682, 'non-captured': 7, 'sensor 0': set(), 'sensor 1': {(1, 999)}, 'sensor 2': set(), 'sensor 3': {(1, 999)}, 'sensor 4': set(), 'sensor 5': {(1, 995)}, 'sensor 6': {(1, 999)}, 'sensor 7': {(1, 999)}, 'sensor 8': {(1, 999)}, 'sensor 9': {(1, 999), (1, 995)}, 'sensor 10': {(1, 999)}, 'sensor 11': set()}\n",
            "Episode:18\t Score:283.76 \t{'captured': 690, 'non-captured': 8, 'sensor 0': set(), 'sensor 1': set(), 'sensor 2': {(1, 993), (1, 1000)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 993), (1, 1000)}, 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': {(1, 993)}, 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 993), (1, 1000)}}\n",
            "Episode:19\t Score:287.99 \t{'captured': 713, 'non-captured': 6, 'sensor 0': {(1, 997), (1, 992), (1, 987)}, 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': {(1, 997), (1, 992), (1, 987)}, 'sensor 4': set(), 'sensor 5': set(), 'sensor 6': set(), 'sensor 7': {(1, 997)}, 'sensor 8': {(1, 997)}, 'sensor 9': set(), 'sensor 10': {(1, 997)}, 'sensor 11': {(1, 997), (1, 987)}}\n",
            "Episode:20\t Score:302.77 \t{'captured': 663, 'non-captured': 6, 'sensor 0': {(1, 994)}, 'sensor 1': set(), 'sensor 2': set(), 'sensor 3': {(1, 994)}, 'sensor 4': set(), 'sensor 5': {(1, 999)}, 'sensor 6': set(), 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 999)}, 'sensor 10': set(), 'sensor 11': {(1, 994)}}\n",
            "Average Score: 287.1759379951636\n"
          ]
        }
      ],
      "source": [
        "episodes = 20\n",
        "scores = []\n",
        "\n",
        "for episode in range(1, episodes+1):\n",
        "    state, info = env.reset()\n",
        "    done = False\n",
        "    score = 0\n",
        "\n",
        "    while not done:\n",
        "        action = env.action_space.sample()\n",
        "        n_state, reward, terminated, truncated, info = env.step(action)\n",
        "        score += reward\n",
        "        done = terminated or truncated\n",
        "\n",
        "    scores.append(score)\n",
        "    print('Episode:{}\\t Score:{:.2f} \\t{}'.format(episode, score, info))\n",
        "\n",
        "# Print average score\n",
        "print('Average Score:', np.mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2MSYRNl_A-Gz",
        "outputId": "b97aa58c-3735-409d-ffdc-782eab73ea29"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([[7.3199391e-01, 5.9865850e-01, 4.6000000e+01],\n",
              "       [1.5601864e-01, 1.5599452e-01, 5.6000000e+01],\n",
              "       [5.8083612e-02, 8.6617613e-01, 4.7000000e+01],\n",
              "       [6.0111499e-01, 7.0807260e-01, 5.0000000e+01],\n",
              "       [2.0584494e-02, 9.6990985e-01, 5.9000000e+01],\n",
              "       [8.3244264e-01, 2.1233912e-01, 6.0000000e+01],\n",
              "       [1.8182497e-01, 1.8340451e-01, 4.7000000e+01],\n",
              "       [3.0424225e-01, 5.2475643e-01, 7.1000000e+01],\n",
              "       [4.3194503e-01, 2.9122913e-01, 6.6000000e+01],\n",
              "       [6.1185288e-01, 1.3949387e-01, 5.7000000e+01],\n",
              "       [2.9214466e-01, 3.6636186e-01, 5.5000000e+01],\n",
              "       [4.5606998e-01, 7.8517598e-01, 5.5000000e+01]], dtype=float32)"
            ]
          },
          "execution_count": 7,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "env.sensor_positions"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Od_rc9sMA3Ef",
        "outputId": "b77c5b22-fa13-41d3-de78-370ae710ac15"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/stable_baselines3/common/env_checker.py:244: UserWarning: Your observation  has an unconventional shape (neither an image, nor a 1D vector). We recommend you to flatten the observation to have only a 1D vector or use a custom policy to properly process the data.\n",
            "  warnings.warn(\n"
          ]
        }
      ],
      "source": [
        "check_env(env, warn=True, skip_render_check=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Training the Deep Q-Network (DQN) Model\n",
        "\n",
        "This section of the code initializes and trains a Deep Q-Network (DQN) model using the custom Wireless Sensor Network (WSN) environment. The steps include:\n",
        "\n",
        "- **Model Initialization**:\n",
        "  - The DQN model is initialized with an `MlpPolicy`, which is a multilayer perceptron policy. This policy defines the neural network architecture used by the DQN to approximate the Q-values.\n",
        "  - The custom WSN environment (`env`) is passed to the model, and `verbose=1` enables detailed logging of the training process.\n",
        "\n",
        "- **Model Training**:\n",
        "  - The DQN model is trained for `100,000` timesteps. During training, the model interacts with the environment, selecting actions, receiving rewards, and updating its Q-values to learn an optimal policy.\n",
        "  - The `log_interval=4` parameter specifies that training progress should be logged every 4 timesteps, providing regular updates on the learning process.\n",
        "\n",
        "This training step is critical for enabling the DQN model to learn effective scheduling strategies in the WSN environment, optimizing sensor selection and event capture.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z-BtrdFAwlNj",
        "outputId": "a8c22b67-f948-426d-9b37-33fa2a15b01e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Using cuda device\n",
            "Wrapping the env with a `Monitor` wrapper\n",
            "Wrapping the env in a DummyVecEnv.\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 294      |\n",
            "|    exploration_rate | 0.619    |\n",
            "| time/               |          |\n",
            "|    episodes         | 4        |\n",
            "|    fps              | 4604     |\n",
            "|    time_elapsed     | 0        |\n",
            "|    total_timesteps  | 4008     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.238    |\n",
            "| time/               |          |\n",
            "|    episodes         | 8        |\n",
            "|    fps              | 4656     |\n",
            "|    time_elapsed     | 1        |\n",
            "|    total_timesteps  | 8016     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 12       |\n",
            "|    fps              | 4522     |\n",
            "|    time_elapsed     | 2        |\n",
            "|    total_timesteps  | 12024    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 16       |\n",
            "|    fps              | 4559     |\n",
            "|    time_elapsed     | 3        |\n",
            "|    total_timesteps  | 16032    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 289      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 20       |\n",
            "|    fps              | 4576     |\n",
            "|    time_elapsed     | 4        |\n",
            "|    total_timesteps  | 20040    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 24       |\n",
            "|    fps              | 4592     |\n",
            "|    time_elapsed     | 5        |\n",
            "|    total_timesteps  | 24048    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 288      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 28       |\n",
            "|    fps              | 4610     |\n",
            "|    time_elapsed     | 6        |\n",
            "|    total_timesteps  | 28056    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 289      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 32       |\n",
            "|    fps              | 4608     |\n",
            "|    time_elapsed     | 6        |\n",
            "|    total_timesteps  | 32064    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 36       |\n",
            "|    fps              | 4342     |\n",
            "|    time_elapsed     | 8        |\n",
            "|    total_timesteps  | 36072    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 289      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 40       |\n",
            "|    fps              | 4090     |\n",
            "|    time_elapsed     | 9        |\n",
            "|    total_timesteps  | 40080    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 290      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 44       |\n",
            "|    fps              | 4135     |\n",
            "|    time_elapsed     | 10       |\n",
            "|    total_timesteps  | 44088    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 289      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 48       |\n",
            "|    fps              | 4168     |\n",
            "|    time_elapsed     | 11       |\n",
            "|    total_timesteps  | 48096    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 284      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 52       |\n",
            "|    fps              | 2928     |\n",
            "|    time_elapsed     | 17       |\n",
            "|    total_timesteps  | 52104    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.311    |\n",
            "|    n_updates        | 525      |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 275      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 56       |\n",
            "|    fps              | 2302     |\n",
            "|    time_elapsed     | 24       |\n",
            "|    total_timesteps  | 56112    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.169    |\n",
            "|    n_updates        | 1527     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 272      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 60       |\n",
            "|    fps              | 1988     |\n",
            "|    time_elapsed     | 30       |\n",
            "|    total_timesteps  | 60120    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.192    |\n",
            "|    n_updates        | 2529     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 268      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 64       |\n",
            "|    fps              | 1750     |\n",
            "|    time_elapsed     | 36       |\n",
            "|    total_timesteps  | 64128    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.181    |\n",
            "|    n_updates        | 3531     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 266      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 68       |\n",
            "|    fps              | 1616     |\n",
            "|    time_elapsed     | 42       |\n",
            "|    total_timesteps  | 68136    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.204    |\n",
            "|    n_updates        | 4533     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 264      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 72       |\n",
            "|    fps              | 1489     |\n",
            "|    time_elapsed     | 48       |\n",
            "|    total_timesteps  | 72144    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.185    |\n",
            "|    n_updates        | 5535     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 263      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 76       |\n",
            "|    fps              | 1413     |\n",
            "|    time_elapsed     | 53       |\n",
            "|    total_timesteps  | 76152    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.253    |\n",
            "|    n_updates        | 6537     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 261      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 80       |\n",
            "|    fps              | 1325     |\n",
            "|    time_elapsed     | 60       |\n",
            "|    total_timesteps  | 80160    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.154    |\n",
            "|    n_updates        | 7539     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 259      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 84       |\n",
            "|    fps              | 1277     |\n",
            "|    time_elapsed     | 65       |\n",
            "|    total_timesteps  | 84168    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.196    |\n",
            "|    n_updates        | 8541     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 256      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 88       |\n",
            "|    fps              | 1201     |\n",
            "|    time_elapsed     | 73       |\n",
            "|    total_timesteps  | 88176    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.457    |\n",
            "|    n_updates        | 9543     |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 253      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 92       |\n",
            "|    fps              | 1169     |\n",
            "|    time_elapsed     | 78       |\n",
            "|    total_timesteps  | 92184    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.163    |\n",
            "|    n_updates        | 10545    |\n",
            "----------------------------------\n",
            "----------------------------------\n",
            "| rollout/            |          |\n",
            "|    ep_len_mean      | 1e+03    |\n",
            "|    ep_rew_mean      | 253      |\n",
            "|    exploration_rate | 0.05     |\n",
            "| time/               |          |\n",
            "|    episodes         | 96       |\n",
            "|    fps              | 1132     |\n",
            "|    time_elapsed     | 84       |\n",
            "|    total_timesteps  | 96192    |\n",
            "| train/              |          |\n",
            "|    learning_rate    | 0.0001   |\n",
            "|    loss             | 0.165    |\n",
            "|    n_updates        | 11547    |\n",
            "----------------------------------\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<stable_baselines3.dqn.dqn.DQN at 0x7afcb6cf4e80>"
            ]
          },
          "execution_count": 9,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model = DQN(\"MlpPolicy\", env, verbose=1)\n",
        "model.learn(total_timesteps=100_000, log_interval=4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Evaluation of the Trained DQN Model\n",
        "\n",
        "This section of the code evaluates the performance of the trained Deep Q-Network (DQN) model in the Wireless Sensor Network (WSN) environment. The evaluation process includes:\n",
        "\n",
        "- **Initialization of evaluation parameters**, including the number of episodes and a list to store scores.\n",
        "- **Execution of the DQN model**:\n",
        "  - For each episode, the environment is reset, and the DQN model predicts actions based on the current state.\n",
        "  - The agent accumulates rewards by selecting actions, and the episode continues until the termination condition is met (maximum steps or other criteria).\n",
        "  - The total score for each episode is recorded.\n",
        "- **Recording and printing of scores**:\n",
        "  - After each episode, the score is printed alongside the information about captured and non-captured events.\n",
        "- **Calculation and display of the average score** across all episodes, providing a measure of the DQN model's effectiveness in optimizing sensor scheduling.\n",
        "\n",
        "This evaluation step is essential for assessing how well the DQN model has learned and its ability to outperform baseline scheduling strategies.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M1O52Cu8xJua",
        "outputId": "24f0185f-31d2-4ef9-be24-97016f2cac00"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Episode:1\t Score:317.66 \t{'captured': 651, 'non-captured': 43, 'sensor 0': {(1, 626), (1, 490), (1, 560), (1, 842), (1, 781), (1, 218), (1, 671), (1, 622), (1, 994), (1, 843), (1, 997), (1, 771)}, 'sensor 1': {(1, 577), (1, 249)}, 'sensor 2': {(1, 632), (1, 644), (1, 854), (1, 561), (1, 439), (1, 451), (1, 42), (1, 890), (1, 835), (1, 292), (1, 844), (1, 795), (1, 432), (1, 612), (1, 267), (1, 807), (1, 999), (1, 657), (1, 837), (1, 309), (1, 617), (1, 385)}, 'sensor 3': set(), 'sensor 4': {(1, 632), (1, 644), (1, 854), (1, 561), (1, 439), (1, 451), (1, 42), (1, 890), (1, 835), (1, 292), (1, 844), (1, 795), (1, 432), (1, 612), (1, 267), (1, 807), (1, 999), (1, 657), (1, 837), (1, 309), (1, 617), (1, 385)}, 'sensor 5': {(1, 671), (1, 560), (1, 994), (1, 843), (1, 997), (1, 842), (1, 781), (1, 218)}, 'sensor 6': {(1, 577), (1, 249)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 997)}, 'sensor 10': set(), 'sensor 11': {(1, 632), (1, 844), (1, 795), (1, 612), (1, 267), (1, 999), (1, 561), (1, 439), (1, 451), (1, 837), (1, 309), (1, 617), (1, 890)}}\n",
            "Episode:2\t Score:299.74 \t{'captured': 643, 'non-captured': 53, 'sensor 0': {(1, 525), (1, 885), (1, 833), (1, 946), (1, 711), (1, 241), (1, 247), (1, 201), (1, 805), (1, 701), (1, 350), (1, 713), (1, 563), (1, 331), (1, 767), (1, 242), (1, 910), (1, 501)}, 'sensor 1': {(1, 276), (1, 567), (1, 123)}, 'sensor 2': {(1, 928), (1, 986), (1, 870), (1, 275), (1, 159), (1, 98), (1, 235), (1, 781), (1, 860), (1, 445), (1, 503), (1, 335), (1, 518), (1, 402), (1, 646), (1, 588), (1, 234), (1, 182), (1, 310), (1, 255), (1, 313), (1, 618), (1, 624), (1, 865), (1, 465), (1, 349), (1, 858)}, 'sensor 3': set(), 'sensor 4': {(1, 928), (1, 986), (1, 870), (1, 275), (1, 159), (1, 98), (1, 235), (1, 781), (1, 860), (1, 445), (1, 503), (1, 335), (1, 518), (1, 402), (1, 646), (1, 588), (1, 234), (1, 182), (1, 310), (1, 255), (1, 313), (1, 618), (1, 624), (1, 865), (1, 465), (1, 349), (1, 858)}, 'sensor 5': {(1, 525), (1, 885), (1, 563), (1, 833), (1, 711), (1, 946), (1, 331), (1, 247), (1, 767), (1, 805), (1, 910), (1, 501), (1, 713)}, 'sensor 6': {(1, 276), (1, 567), (1, 123)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 986), (1, 870), (1, 275), (1, 235), (1, 781), (1, 860), (1, 503), (1, 335), (1, 518), (1, 646), (1, 234), (1, 182), (1, 310), (1, 618), (1, 624), (1, 865), (1, 349)}}\n",
            "Episode:3\t Score:292.47 \t{'captured': 647, 'non-captured': 43, 'sensor 0': {(1, 333), (1, 351), (1, 575), (1, 935), (1, 462), (1, 288), (1, 526), (1, 500), (1, 593), (1, 829)}, 'sensor 1': {(1, 897), (1, 595)}, 'sensor 2': {(1, 992), (1, 841), (1, 960), (1, 876), (1, 725), (1, 995), (1, 969), (1, 255), (1, 650), (1, 740), (1, 17), (1, 946), (1, 325), (1, 337), (1, 976), (1, 834), (1, 428), (1, 936), (1, 951), (1, 980), (1, 588), (1, 54), (1, 893)}, 'sensor 3': {(1, 995)}, 'sensor 4': {(1, 992), (1, 841), (1, 960), (1, 876), (1, 725), (1, 995), (1, 969), (1, 255), (1, 650), (1, 740), (1, 17), (1, 946), (1, 325), (1, 337), (1, 976), (1, 834), (1, 428), (1, 936), (1, 951), (1, 980), (1, 588), (1, 54), (1, 893)}, 'sensor 5': {(1, 935), (1, 575), (1, 125), (1, 462), (1, 288), (1, 526), (1, 500), (1, 593)}, 'sensor 6': {(1, 897), (1, 595)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 876), (1, 995), (1, 969), (1, 255), (1, 650), (1, 740), (1, 946), (1, 325), (1, 428), (1, 980), (1, 588)}}\n",
            "Episode:4\t Score:293.50 \t{'captured': 630, 'non-captured': 49, 'sensor 0': {(1, 992), (1, 751), (1, 998), (1, 618), (1, 287), (1, 996), (1, 369), (1, 999), (1, 732), (1, 465), (1, 320), (1, 326), (1, 509), (1, 358), (1, 898), (1, 315)}, 'sensor 1': {(1, 423), (1, 626), (1, 773), (1, 817), (1, 87), (1, 899)}, 'sensor 2': {(1, 333), (1, 513), (1, 693), (1, 815), (1, 760), (1, 531), (1, 180), (1, 967), (1, 552), (1, 634), (1, 939), (1, 771), (1, 179), (1, 563), (1, 520), (1, 453), (1, 761), (1, 410), (1, 709), (1, 733), (1, 852), (1, 196), (1, 205)}, 'sensor 3': {(1, 996), (1, 998), (1, 999)}, 'sensor 4': {(1, 333), (1, 513), (1, 693), (1, 815), (1, 760), (1, 531), (1, 180), (1, 967), (1, 552), (1, 634), (1, 939), (1, 771), (1, 179), (1, 563), (1, 520), (1, 453), (1, 761), (1, 410), (1, 709), (1, 733), (1, 852), (1, 196), (1, 205)}, 'sensor 5': {(1, 992), (1, 751), (1, 618), (1, 369), (1, 465), (1, 320), (1, 509), (1, 358), (1, 898), (1, 315), (1, 112)}, 'sensor 6': {(1, 423), (1, 626), (1, 773), (1, 817), (1, 87), (1, 899)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 333), (1, 693), (1, 760), (1, 998), (1, 563), (1, 180), (1, 453), (1, 520), (1, 967), (1, 552), (1, 761), (1, 709), (1, 634), (1, 733), (1, 852), (1, 205), (1, 179)}}\n",
            "Episode:5\t Score:305.96 \t{'captured': 666, 'non-captured': 45, 'sensor 0': {(1, 780), (1, 226), (1, 766), (1, 612), (1, 377), (1, 726), (1, 996), (1, 912), (1, 709), (1, 503), (1, 782), (1, 338), (1, 312), (1, 521), (1, 881), (1, 803)}, 'sensor 1': {(1, 696), (1, 615)}, 'sensor 2': {(1, 870), (1, 528), (1, 424), (1, 497), (1, 88), (1, 759), (1, 710), (1, 948), (1, 594), (1, 536), (1, 603), (1, 545), (1, 676), (1, 255), (1, 279), (1, 404), (1, 947), (1, 245), (1, 837), (1, 254), (1, 269)}, 'sensor 3': set(), 'sensor 4': {(1, 870), (1, 528), (1, 424), (1, 497), (1, 88), (1, 759), (1, 710), (1, 948), (1, 594), (1, 536), (1, 603), (1, 545), (1, 676), (1, 255), (1, 279), (1, 404), (1, 947), (1, 245), (1, 837), (1, 254), (1, 269)}, 'sensor 5': {(1, 780), (1, 226), (1, 998), (1, 377), (1, 726), (1, 912), (1, 782), (1, 338), (1, 312), (1, 341), (1, 521), (1, 881), (1, 803)}, 'sensor 6': {(1, 696), (1, 615)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 998)}, 'sensor 10': set(), 'sensor 11': {(1, 603), (1, 870), (1, 545), (1, 676), (1, 255), (1, 528), (1, 424), (1, 279), (1, 404), (1, 497), (1, 947), (1, 837), (1, 948), (1, 594)}}\n",
            "Episode:6\t Score:318.12 \t{'captured': 662, 'non-captured': 55, 'sensor 0': {(1, 658), (1, 830), (1, 441), (1, 254), (1, 752), (1, 784), (1, 860), (1, 564), (1, 811), (1, 518), (1, 997), (1, 646), (1, 765), (1, 501), (1, 344)}, 'sensor 1': {(1, 379), (1, 718), (1, 162)}, 'sensor 2': {(1, 31), (1, 330), (1, 577), (1, 949), (1, 894), (1, 351), (1, 595), (1, 488), (1, 622), (1, 338), (1, 771), (1, 789), (1, 795), (1, 969), (1, 859), (1, 78), (1, 575), (1, 562), (1, 169), (1, 895), (1, 904), (1, 62), (1, 248), (1, 367), (1, 846), (1, 736), (1, 382)}, 'sensor 3': {(1, 997)}, 'sensor 4': {(1, 31), (1, 330), (1, 577), (1, 949), (1, 894), (1, 351), (1, 382), (1, 595), (1, 488), (1, 622), (1, 338), (1, 771), (1, 789), (1, 795), (1, 969), (1, 859), (1, 78), (1, 575), (1, 169), (1, 895), (1, 904), (1, 62), (1, 248), (1, 367), (1, 846), (1, 736), (1, 562)}, 'sensor 5': {(1, 501), (1, 830), (1, 441), (1, 784), (1, 518), (1, 1000), (1, 646), (1, 765), (1, 254), (1, 344)}, 'sensor 6': {(1, 379), (1, 718), (1, 162)}, 'sensor 7': set(), 'sensor 8': {(1, 1000)}, 'sensor 9': {(1, 1000)}, 'sensor 10': set(), 'sensor 11': {(1, 595), (1, 488), (1, 622), (1, 771), (1, 789), (1, 795), (1, 969), (1, 859), (1, 575), (1, 169), (1, 895), (1, 904), (1, 248), (1, 367), (1, 736), (1, 562)}}\n",
            "Episode:7\t Score:312.26 \t{'captured': 662, 'non-captured': 35, 'sensor 0': {(1, 760), (1, 392), (1, 546), (1, 944), (1, 410), (1, 654), (1, 239), (1, 405)}, 'sensor 1': {(1, 491), (1, 184), (1, 855), (1, 292)}, 'sensor 2': {(1, 336), (1, 702), (1, 552), (1, 802), (1, 631), (1, 634), (1, 512), (1, 167), (1, 655), (1, 89), (1, 667), (1, 844), (1, 978), (1, 273), (1, 459), (1, 880), (1, 468), (1, 922)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 667), (1, 336), (1, 844), (1, 702), (1, 922), (1, 978), (1, 273), (1, 459), (1, 880), (1, 552), (1, 468), (1, 802), (1, 631), (1, 634), (1, 512), (1, 167), (1, 655), (1, 89)}, 'sensor 5': {(1, 760), (1, 830), (1, 392), (1, 546), (1, 944), (1, 410), (1, 654), (1, 239), (1, 405)}, 'sensor 6': {(1, 491), (1, 184), (1, 855), (1, 292)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 667), (1, 336), (1, 844), (1, 978), (1, 880), (1, 459), (1, 552), (1, 468), (1, 631), (1, 167), (1, 1000), (1, 655)}}\n",
            "Episode:8\t Score:313.96 \t{'captured': 632, 'non-captured': 57, 'sensor 0': {(1, 211), (1, 995), (1, 287), (1, 203), (1, 595), (1, 828), (1, 657), (1, 489), (1, 579), (1, 463), (1, 312), (1, 762), (1, 730), (1, 440), (1, 324), (1, 809)}, 'sensor 1': {(1, 356), (1, 513), (1, 905), (1, 236)}, 'sensor 2': {(1, 226), (1, 345), (1, 458), (1, 403), (1, 958), (1, 842), (1, 665), (1, 903), (1, 787), (1, 613), (1, 671), (1, 973), (1, 735), (1, 335), (1, 207), (1, 161), (1, 521), (1, 771), (1, 533), (1, 179), (1, 725), (1, 859), (1, 209), (1, 874), (1, 819), (1, 877), (1, 352), (1, 416), (1, 175), (1, 541), (1, 593), (1, 544), (1, 666)}, 'sensor 3': set(), 'sensor 4': {(1, 226), (1, 345), (1, 458), (1, 403), (1, 958), (1, 842), (1, 665), (1, 903), (1, 787), (1, 613), (1, 973), (1, 671), (1, 735), (1, 335), (1, 207), (1, 161), (1, 521), (1, 771), (1, 533), (1, 179), (1, 725), (1, 859), (1, 209), (1, 874), (1, 819), (1, 877), (1, 352), (1, 416), (1, 175), (1, 541), (1, 593), (1, 544), (1, 666)}, 'sensor 5': {(1, 211), (1, 995), (1, 287), (1, 203), (1, 595), (1, 996), (1, 828), (1, 657), (1, 489), (1, 579), (1, 312), (1, 762), (1, 730), (1, 809)}, 'sensor 6': {(1, 356), (1, 513), (1, 905), (1, 236)}, 'sensor 7': set(), 'sensor 8': {(1, 996)}, 'sensor 9': {(1, 996)}, 'sensor 10': set(), 'sensor 11': {(1, 403), (1, 613), (1, 973), (1, 671), (1, 161), (1, 521), (1, 771), (1, 179), (1, 725), (1, 859), (1, 209), (1, 874), (1, 352), (1, 593), (1, 544), (1, 666)}}\n",
            "Episode:9\t Score:302.74 \t{'captured': 640, 'non-captured': 46, 'sensor 0': {(1, 380), (1, 804), (1, 331), (1, 784), (1, 369), (1, 584), (1, 227), (1, 735), (1, 979), (1, 239), (1, 695), (1, 504)}, 'sensor 1': {(1, 815), (1, 984), (1, 457), (1, 891), (1, 550), (1, 993)}, 'sensor 2': {(1, 928), (1, 217), (1, 394), (1, 766), (1, 830), (1, 168), (1, 467), (1, 357), (1, 180), (1, 790), (1, 192), (1, 634), (1, 881), (1, 719), (1, 777), (1, 249), (1, 432), (1, 264), (1, 206), (1, 544), (1, 123)}, 'sensor 3': set(), 'sensor 4': {(1, 928), (1, 217), (1, 394), (1, 766), (1, 830), (1, 168), (1, 467), (1, 357), (1, 180), (1, 790), (1, 192), (1, 634), (1, 881), (1, 719), (1, 777), (1, 249), (1, 432), (1, 264), (1, 206), (1, 544), (1, 123)}, 'sensor 5': {(1, 380), (1, 804), (1, 784), (1, 369), (1, 227), (1, 735), (1, 979), (1, 504)}, 'sensor 6': {(1, 815), (1, 984), (1, 457), (1, 891), (1, 550), (1, 993)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 928), (1, 777), (1, 249), (1, 432), (1, 264), (1, 467), (1, 357), (1, 206), (1, 180), (1, 192), (1, 544), (1, 881)}}\n",
            "Episode:10\t Score:294.63 \t{'captured': 651, 'non-captured': 46, 'sensor 0': {(1, 191), (1, 998), (1, 409), (1, 412), (1, 209), (1, 459), (1, 854), (1, 346), (1, 262), (1, 713)}, 'sensor 1': {(1, 208), (1, 619)}, 'sensor 2': {(1, 705), (1, 964), (1, 430), (1, 442), (1, 271), (1, 930), (1, 869), (1, 448), (1, 167), (1, 1000), (1, 48), (1, 173), (1, 783), (1, 996), (1, 99), (1, 639), (1, 648), (1, 230), (1, 175), (1, 233), (1, 300), (1, 480), (1, 660), (1, 773), (1, 489), (1, 428), (1, 367), (1, 925), (1, 324)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 705), (1, 964), (1, 430), (1, 442), (1, 271), (1, 930), (1, 869), (1, 448), (1, 167), (1, 1000), (1, 48), (1, 173), (1, 783), (1, 996), (1, 99), (1, 639), (1, 648), (1, 230), (1, 175), (1, 233), (1, 300), (1, 480), (1, 773), (1, 660), (1, 489), (1, 428), (1, 367), (1, 925), (1, 324)}, 'sensor 5': {(1, 346), (1, 191), (1, 409), (1, 459), (1, 713), (1, 854)}, 'sensor 6': {(1, 208), (1, 619)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 996), (1, 639), (1, 648), (1, 442), (1, 175), (1, 480), (1, 773), (1, 448), (1, 869), (1, 930), (1, 367), (1, 1000), (1, 925)}}\n",
            "Episode:11\t Score:316.70 \t{'captured': 677, 'non-captured': 49, 'sensor 0': {(1, 763), (1, 882), (1, 827), (1, 833), (1, 711), (1, 981), (1, 206), (1, 709), (1, 535), (1, 553), (1, 1000), (1, 925), (1, 719)}, 'sensor 1': {(1, 391), (1, 455), (1, 43), (1, 244)}, 'sensor 2': {(1, 266), (1, 693), (1, 754), (1, 525), (1, 235), (1, 186), (1, 906), (1, 445), (1, 628), (1, 695), (1, 33), (1, 393), (1, 817), (1, 167), (1, 460), (1, 2), (1, 121), (1, 304), (1, 249), (1, 374), (1, 691), (1, 209), (1, 273), (1, 163), (1, 282), (1, 895), (1, 541), (1, 248), (1, 797), (1, 977)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 266), (1, 693), (1, 754), (1, 525), (1, 235), (1, 186), (1, 906), (1, 445), (1, 628), (1, 695), (1, 33), (1, 393), (1, 817), (1, 167), (1, 460), (1, 2), (1, 121), (1, 304), (1, 249), (1, 374), (1, 691), (1, 209), (1, 273), (1, 163), (1, 282), (1, 895), (1, 541), (1, 248), (1, 797), (1, 977)}, 'sensor 5': {(1, 882), (1, 827), (1, 833), (1, 711), (1, 981), (1, 206), (1, 535), (1, 553), (1, 925)}, 'sensor 6': {(1, 391), (1, 455), (1, 43), (1, 244)}, 'sensor 7': {(1, 1000)}, 'sensor 8': {(1, 1000)}, 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 525), (1, 374), (1, 235), (1, 186), (1, 209), (1, 273), (1, 163), (1, 282), (1, 445), (1, 895), (1, 695), (1, 817), (1, 167), (1, 1000)}}\n",
            "Episode:12\t Score:292.39 \t{'captured': 659, 'non-captured': 40, 'sensor 0': {(1, 928), (1, 391), (1, 284), (1, 470), (1, 322), (1, 868), (1, 720), (1, 279), (1, 697), (1, 564), (1, 390), (1, 794)}, 'sensor 1': {(1, 410)}, 'sensor 2': {(1, 40), (1, 589), (1, 894), (1, 595), (1, 662), (1, 549), (1, 680), (1, 155), (1, 39), (1, 997), (1, 713), (1, 472), (1, 731), (1, 310), (1, 374), (1, 569), (1, 648), (1, 651)}, 'sensor 3': set(), 'sensor 4': {(1, 40), (1, 589), (1, 894), (1, 595), (1, 662), (1, 549), (1, 680), (1, 155), (1, 39), (1, 997), (1, 713), (1, 472), (1, 731), (1, 310), (1, 374), (1, 569), (1, 648), (1, 651)}, 'sensor 5': {(1, 928), (1, 992), (1, 391), (1, 284), (1, 322), (1, 868), (1, 720), (1, 279), (1, 697), (1, 564), (1, 390), (1, 794)}, 'sensor 6': {(1, 410)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 992)}, 'sensor 10': set(), 'sensor 11': {(1, 310), (1, 374), (1, 589), (1, 894), (1, 595), (1, 662), (1, 651), (1, 680), (1, 155)}}\n",
            "Episode:13\t Score:336.98 \t{'captured': 664, 'non-captured': 51, 'sensor 0': {(1, 304), (1, 673), (1, 740), (1, 258), (1, 203), (1, 238), (1, 601), (1, 999), (1, 616), (1, 207), (1, 939), (1, 701), (1, 318), (1, 414)}, 'sensor 1': {(1, 535), (1, 943), (1, 810), (1, 494)}, 'sensor 2': {(1, 937), (1, 156), (1, 214), (1, 461), (1, 824), (1, 290), (1, 162), (1, 403), (1, 232), (1, 366), (1, 436), (1, 744), (1, 698), (1, 643), (1, 704), (1, 469), (1, 832), (1, 371), (1, 978), (1, 804), (1, 383), (1, 932), (1, 395), (1, 697), (1, 288), (1, 230), (1, 477), (1, 910), (1, 989)}, 'sensor 3': {(1, 999)}, 'sensor 4': {(1, 937), (1, 156), (1, 214), (1, 461), (1, 824), (1, 290), (1, 162), (1, 403), (1, 232), (1, 366), (1, 436), (1, 744), (1, 698), (1, 643), (1, 704), (1, 469), (1, 832), (1, 371), (1, 978), (1, 804), (1, 383), (1, 932), (1, 395), (1, 697), (1, 288), (1, 230), (1, 989), (1, 477), (1, 910)}, 'sensor 5': {(1, 5), (1, 304), (1, 673), (1, 740), (1, 258), (1, 203), (1, 238), (1, 601), (1, 616), (1, 207), (1, 701), (1, 318), (1, 414)}, 'sensor 6': {(1, 535), (1, 943), (1, 810), (1, 494)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 937), (1, 156), (1, 214), (1, 403), (1, 162), (1, 804), (1, 232), (1, 932), (1, 999), (1, 436), (1, 832), (1, 910), (1, 704), (1, 469), (1, 989)}}\n",
            "Episode:14\t Score:293.01 \t{'captured': 640, 'non-captured': 54, 'sensor 0': {(1, 760), (1, 415), (1, 726), (1, 970), (1, 857), (1, 503), (1, 878), (1, 576), (1, 1000), (1, 948), (1, 658), (1, 371), (1, 563), (1, 456), (1, 282), (1, 221), (1, 532), (1, 724), (1, 248)}, 'sensor 1': {(1, 771), (1, 919), (1, 628)}, 'sensor 2': {(1, 275), (1, 458), (1, 775), (1, 961), (1, 604), (1, 787), (1, 552), (1, 506), (1, 210), (1, 826), (1, 524), (1, 164), (1, 832), (1, 5), (1, 914), (1, 862), (1, 627), (1, 807), (1, 999), (1, 947), (1, 178), (1, 117), (1, 962), (1, 190), (1, 196), (1, 260), (1, 318)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 275), (1, 458), (1, 775), (1, 961), (1, 604), (1, 787), (1, 552), (1, 506), (1, 210), (1, 826), (1, 524), (1, 164), (1, 832), (1, 5), (1, 914), (1, 862), (1, 627), (1, 807), (1, 999), (1, 947), (1, 178), (1, 117), (1, 962), (1, 190), (1, 196), (1, 260), (1, 318)}, 'sensor 5': {(1, 995), (1, 371), (1, 760), (1, 726), (1, 282), (1, 532), (1, 724), (1, 248), (1, 878), (1, 948)}, 'sensor 6': {(1, 628), (1, 771), (1, 919)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 995)}, 'sensor 10': set(), 'sensor 11': {(1, 458), (1, 914), (1, 627), (1, 775), (1, 604), (1, 999), (1, 787), (1, 552), (1, 210), (1, 178), (1, 962), (1, 190), (1, 164), (1, 260), (1, 318)}}\n",
            "Episode:15\t Score:280.86 \t{'captured': 643, 'non-captured': 56, 'sensor 0': {(1, 458), (1, 284), (1, 644), (1, 232), (1, 256), (1, 262), (1, 570), (1, 753), (1, 692), (1, 576), (1, 286), (1, 740), (1, 859), (1, 459), (1, 639), (1, 419), (1, 489), (1, 916)}, 'sensor 1': {(1, 203), (1, 836), (1, 247), (1, 613), (1, 997), (1, 527), (1, 974), (1, 295), (1, 719)}, 'sensor 2': {(1, 147), (1, 37), (1, 339), (1, 882), (1, 171), (1, 549), (1, 552), (1, 326), (1, 927), (1, 1000), (1, 945), (1, 350), (1, 890), (1, 655), (1, 548), (1, 264), (1, 404), (1, 584), (1, 877), (1, 532), (1, 538), (1, 843), (1, 321)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 147), (1, 37), (1, 339), (1, 882), (1, 171), (1, 549), (1, 552), (1, 326), (1, 927), (1, 1000), (1, 945), (1, 890), (1, 350), (1, 655), (1, 548), (1, 264), (1, 404), (1, 584), (1, 877), (1, 532), (1, 538), (1, 843), (1, 321)}, 'sensor 5': {(1, 458), (1, 284), (1, 644), (1, 232), (1, 256), (1, 262), (1, 753), (1, 692), (1, 576), (1, 286), (1, 740), (1, 859), (1, 459), (1, 419), (1, 489), (1, 916)}, 'sensor 6': {(1, 613), (1, 719), (1, 203), (1, 997), (1, 836), (1, 527), (1, 974), (1, 295), (1, 247)}, 'sensor 7': {(1, 1000)}, 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 339), (1, 548), (1, 882), (1, 584), (1, 552), (1, 890), (1, 532), (1, 326), (1, 843), (1, 1000), (1, 945), (1, 321), (1, 350), (1, 655)}}\n",
            "Episode:16\t Score:289.67 \t{'captured': 648, 'non-captured': 48, 'sensor 0': {(1, 783), (1, 963), (1, 496), (1, 714), (1, 891), (1, 235), (1, 270), (1, 755), (1, 903), (1, 761), (1, 785), (1, 393), (1, 422), (1, 518), (1, 1000)}, 'sensor 1': {(1, 444), (1, 980), (1, 998), (1, 268)}, 'sensor 2': {(1, 940), (1, 467), (1, 171), (1, 238), (1, 372), (1, 271), (1, 384), (1, 930), (1, 957), (1, 301), (1, 60), (1, 377), (1, 41), (1, 468), (1, 291), (1, 68), (1, 788), (1, 553), (1, 681), (1, 498)}, 'sensor 3': set(), 'sensor 4': {(1, 940), (1, 467), (1, 171), (1, 238), (1, 372), (1, 271), (1, 384), (1, 930), (1, 957), (1, 301), (1, 60), (1, 377), (1, 41), (1, 468), (1, 291), (1, 68), (1, 788), (1, 553), (1, 681), (1, 498)}, 'sensor 5': {(1, 783), (1, 963), (1, 496), (1, 714), (1, 891), (1, 235), (1, 903), (1, 761), (1, 393), (1, 518), (1, 1000), (1, 713)}, 'sensor 6': {(1, 444), (1, 980), (1, 998), (1, 268)}, 'sensor 7': {(1, 998)}, 'sensor 8': {(1, 998)}, 'sensor 9': {(1, 1000)}, 'sensor 10': {(1, 998)}, 'sensor 11': {(1, 957), (1, 301), (1, 940), (1, 377), (1, 467), (1, 171), (1, 238), (1, 468), (1, 271), (1, 384), (1, 930), (1, 553), (1, 498)}}\n",
            "Episode:17\t Score:267.74 \t{'captured': 646, 'non-captured': 59, 'sensor 0': {(1, 670), (1, 879), (1, 287), (1, 987), (1, 241), (1, 845), (1, 700), (1, 721), (1, 239), (1, 753), (1, 878), (1, 817), (1, 1000), (1, 225), (1, 379)}, 'sensor 1': {(1, 47), (1, 107), (1, 142), (1, 643), (1, 682), (1, 993)}, 'sensor 2': {(1, 217), (1, 455), (1, 284), (1, 897), (1, 900), (1, 125), (1, 247), (1, 851), (1, 915), (1, 680), (1, 625), (1, 204), (1, 692), (1, 991), (1, 228), (1, 774), (1, 629), (1, 118), (1, 743), (1, 691), (1, 450), (1, 227), (1, 166), (1, 471), (1, 901), (1, 419), (1, 657), (1, 312), (1, 675), (1, 562)}, 'sensor 3': {(1, 1000)}, 'sensor 4': {(1, 217), (1, 455), (1, 284), (1, 897), (1, 900), (1, 125), (1, 247), (1, 851), (1, 915), (1, 680), (1, 625), (1, 204), (1, 692), (1, 991), (1, 228), (1, 774), (1, 118), (1, 743), (1, 691), (1, 450), (1, 227), (1, 166), (1, 471), (1, 562), (1, 901), (1, 419), (1, 657), (1, 312), (1, 675), (1, 629)}, 'sensor 5': {(1, 670), (1, 879), (1, 287), (1, 845), (1, 239), (1, 753), (1, 878), (1, 817), (1, 225), (1, 379)}, 'sensor 6': {(1, 47), (1, 107), (1, 142), (1, 643), (1, 682), (1, 993)}, 'sensor 7': {(1, 1000)}, 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 284), (1, 897), (1, 247), (1, 851), (1, 915), (1, 680), (1, 625), (1, 204), (1, 692), (1, 991), (1, 1000), (1, 774), (1, 743), (1, 166), (1, 471), (1, 901), (1, 419), (1, 657), (1, 312), (1, 629)}}\n",
            "Episode:18\t Score:317.71 \t{'captured': 656, 'non-captured': 42, 'sensor 0': {(1, 510), (1, 786), (1, 911), (1, 580), (1, 746), (1, 421), (1, 221), (1, 851), (1, 323), (1, 599), (1, 207), (1, 219), (1, 899)}, 'sensor 1': {(1, 985), (1, 958), (1, 267), (1, 575)}, 'sensor 2': {(1, 644), (1, 290), (1, 952), (1, 717), (1, 259), (1, 579), (1, 222), (1, 701), (1, 167), (1, 783), (1, 740), (1, 502), (1, 352), (1, 834), (1, 840), (1, 672), (1, 675), (1, 919)}, 'sensor 3': set(), 'sensor 4': {(1, 783), (1, 644), (1, 290), (1, 740), (1, 502), (1, 952), (1, 717), (1, 352), (1, 259), (1, 834), (1, 840), (1, 579), (1, 222), (1, 672), (1, 167), (1, 701), (1, 675), (1, 919)}, 'sensor 5': {(1, 786), (1, 911), (1, 998), (1, 746), (1, 421), (1, 221), (1, 851), (1, 323), (1, 599), (1, 219), (1, 899)}, 'sensor 6': {(1, 985), (1, 958), (1, 267), (1, 575)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 998)}, 'sensor 10': set(), 'sensor 11': {(1, 783), (1, 644), (1, 740), (1, 717), (1, 352), (1, 834), (1, 840), (1, 579), (1, 222), (1, 672), (1, 701), (1, 675), (1, 919)}}\n",
            "Episode:19\t Score:321.45 \t{'captured': 671, 'non-captured': 49, 'sensor 0': {(1, 217), (1, 670), (1, 316), (1, 978), (1, 467), (1, 917), (1, 964), (1, 944), (1, 590), (1, 570), (1, 280)}, 'sensor 1': {(1, 751), (1, 606), (1, 261), (1, 756), (1, 874)}, 'sensor 2': {(1, 928), (1, 867), (1, 995), (1, 104), (1, 342), (1, 354), (1, 668), (1, 67), (1, 942), (1, 951), (1, 768), (1, 228), (1, 496), (1, 615), (1, 679), (1, 444), (1, 743), (1, 682), (1, 325), (1, 749), (1, 355), (1, 135), (1, 742)}, 'sensor 3': set(), 'sensor 4': {(1, 928), (1, 867), (1, 995), (1, 104), (1, 342), (1, 354), (1, 668), (1, 67), (1, 942), (1, 951), (1, 768), (1, 228), (1, 496), (1, 615), (1, 679), (1, 444), (1, 743), (1, 682), (1, 325), (1, 749), (1, 355), (1, 135), (1, 742)}, 'sensor 5': {(1, 281), (1, 670), (1, 316), (1, 978), (1, 467), (1, 964), (1, 944), (1, 590), (1, 570), (1, 280)}, 'sensor 6': {(1, 751), (1, 606), (1, 261), (1, 756), (1, 874)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': set(), 'sensor 10': set(), 'sensor 11': {(1, 867), (1, 995), (1, 342), (1, 354), (1, 668), (1, 942), (1, 228), (1, 496), (1, 615), (1, 679), (1, 444), (1, 743), (1, 682), (1, 325), (1, 749), (1, 355)}}\n",
            "Episode:20\t Score:337.87 \t{'captured': 668, 'non-captured': 55, 'sensor 0': {(1, 272), (1, 487), (1, 714), (1, 735), (1, 654), (1, 663), (1, 608), (1, 402), (1, 623)}, 'sensor 1': {(1, 690), (1, 966), (1, 169), (1, 951), (1, 890), (1, 938)}, 'sensor 2': {(1, 992), (1, 336), (1, 995), (1, 760), (1, 363), (1, 848), (1, 793), (1, 198), (1, 390), (1, 396), (1, 161), (1, 707), (1, 414), (1, 829), (1, 716), (1, 789), (1, 740), (1, 432), (1, 377), (1, 508), (1, 276), (1, 755), (1, 996), (1, 813), (1, 291), (1, 50), (1, 541), (1, 117), (1, 364), (1, 251), (1, 141), (1, 629)}, 'sensor 3': set(), 'sensor 4': {(1, 992), (1, 336), (1, 995), (1, 760), (1, 363), (1, 848), (1, 793), (1, 198), (1, 390), (1, 396), (1, 161), (1, 707), (1, 829), (1, 414), (1, 716), (1, 789), (1, 740), (1, 432), (1, 377), (1, 508), (1, 276), (1, 755), (1, 996), (1, 813), (1, 291), (1, 50), (1, 541), (1, 117), (1, 364), (1, 251), (1, 141), (1, 629)}, 'sensor 5': {(1, 272), (1, 487), (1, 438), (1, 714), (1, 735), (1, 663), (1, 608), (1, 402), (1, 1000), (1, 623)}, 'sensor 6': {(1, 690), (1, 966), (1, 169), (1, 951), (1, 890), (1, 938)}, 'sensor 7': set(), 'sensor 8': set(), 'sensor 9': {(1, 1000)}, 'sensor 10': set(), 'sensor 11': {(1, 716), (1, 992), (1, 336), (1, 995), (1, 760), (1, 789), (1, 377), (1, 508), (1, 276), (1, 755), (1, 363), (1, 198), (1, 541), (1, 251), (1, 707)}}\n",
            "Average Score: 305.27159746217393\n"
          ]
        }
      ],
      "source": [
        "episodes = 20\n",
        "scores = []\n",
        "\n",
        "for episode in range(1, episodes+1):\n",
        "    state, info = env.reset()\n",
        "    done = False\n",
        "    score = 0\n",
        "\n",
        "    while not done:\n",
        "        action = model.predict(state, deterministic=True)\n",
        "        n_state, reward, terminated, truncated, info = env.step(action[0])\n",
        "        score += reward\n",
        "        done = terminated or truncated\n",
        "        state = n_state\n",
        "\n",
        "    scores.append(score)\n",
        "    print('Episode:{}\\t Score:{:.2f} \\t{}'.format(episode, score, info))\n",
        "\n",
        "# Print average score\n",
        "print('Average Score:', np.mean(scores))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nrxI0m4DW5KU",
        "outputId": "e35d0d7a-1589-4c6a-cbe5-37082f867f1e"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([[7.3199391e-01, 5.9865850e-01, 2.1000000e+01],\n",
              "       [1.5601864e-01, 1.5599452e-01, 1.0000000e+00],\n",
              "       [5.8083612e-02, 8.6617613e-01, 0.0000000e+00],\n",
              "       [6.0111499e-01, 7.0807260e-01, 2.0400000e+02],\n",
              "       [2.0584494e-02, 9.6990985e-01, 0.0000000e+00],\n",
              "       [8.3244264e-01, 2.1233912e-01, 0.0000000e+00],\n",
              "       [1.8182497e-01, 1.8340451e-01, 0.0000000e+00],\n",
              "       [3.0424225e-01, 5.2475643e-01, 1.2500000e+02],\n",
              "       [4.3194503e-01, 2.9122913e-01, 0.0000000e+00],\n",
              "       [6.1185288e-01, 1.3949387e-01, 1.4300000e+02],\n",
              "       [2.9214466e-01, 3.6636186e-01, 2.0700000e+02],\n",
              "       [4.5606998e-01, 7.8517598e-01, 2.2000000e+01]], dtype=float32)"
            ]
          },
          "execution_count": 11,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "env.sensor_positions"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Time Latency Analysis for DQN Scheduler\n",
        "\n",
        "This section of the code analyzes and visualizes the time latency for events captured by the trained Deep Q-Network (DQN) scheduler in the Wireless Sensor Network (WSN) environment. The process includes:\n",
        "\n",
        "- **Environment reset and initialization** to prepare for the latency analysis.\n",
        "- **Latency data collection**:\n",
        "  - For each step, the DQN model predicts an action, and the latency for each event captured by the selected sensor is recorded.\n",
        "  - Latency is calculated as the difference between the current step count and the event occurrence time.\n",
        "- **Calculation of average, minimum, and maximum latency values** to provide insights into the DQN scheduler's performance in minimizing latency.\n",
        "- **Plotting the time latency**:\n",
        "  - A line plot is generated to visualize the evolution of time latency for captured events across the episode.\n",
        "  - The plot includes labeled axes and a legend for clarity, and the plot is displayed with a tight layout for better visualization.\n",
        "\n",
        "This analysis helps to assess the efficiency of the DQN scheduler in minimizing the latency of captured events, providing a basis for comparison with other scheduling strategies.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 599
        },
        "id": "bMrm6GFSuBK2",
        "outputId": "64ff3929-6e82-43f6-b564-e28bb00b4546"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Average Time latency: 7.49 min: 1.00 max: 133.00\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 0 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "obs, info = env.reset()\n",
        "\n",
        "captured_latency = []  # List to store time latency for captured events\n",
        "\n",
        "for step in range(env.max_steps):\n",
        "    action = model.predict(state, deterministic=True)\n",
        "\n",
        "    if (info['sensor '+ str(action[0])] != set()):\n",
        "      for event in info['sensor '+ str(action[0])]:\n",
        "        time_latency = env.step_count - event[-1]\n",
        "        captured_latency.append(time_latency)\n",
        "\n",
        "    n_obs, reward, terminated, truncated, info = env.step(action[0])\n",
        "    state = n_obs\n",
        "\n",
        "print('Average Time latency: {:.2f} min: {:.2f} max: {:.2f}'.format(np.mean(captured_latency), min(captured_latency), max(captured_latency)))\n",
        "\n",
        "# Plot the evolution of time latency\n",
        "plt.figure(figsize=(10, 6))\n",
        "plt.plot(captured_latency, label='Time Latency for Captured Events')\n",
        "plt.title('SB3 Scheduler')\n",
        "plt.xlabel('Captured Event Index')\n",
        "plt.ylabel('Time Latency (in timesteps)')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "plt.tight_layout()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}