Double_DQN_Keras(TensorFlow)_Solving_Gym_LunarLander.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c1f2274b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:23:38.261895Z",
     "iopub.status.busy": "2024-01-15T06:23:38.261235Z",
     "iopub.status.idle": "2024-01-15T06:23:53.289763Z",
     "shell.execute_reply": "2024-01-15T06:23:53.288409Z"
    },
    "id": "8x-y8DACYp0A",
    "papermill": {
     "duration": 15.038287,
     "end_time": "2024-01-15T06:23:53.292473",
     "exception": false,
     "start_time": "2024-01-15T06:23:38.254186",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import gym\n",
    "from keras.layers import Dense, Activation\n",
    "from keras.models import Sequential, load_model\n",
    "import keras\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e202efb6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:23:53.304736Z",
     "iopub.status.busy": "2024-01-15T06:23:53.303583Z",
     "iopub.status.idle": "2024-01-15T06:26:32.751539Z",
     "shell.execute_reply": "2024-01-15T06:26:32.749881Z"
    },
    "id": "NY4U5CiFZm_T",
    "outputId": "cd68818f-788f-4d1b-99c8-38cbb9164b14",
    "papermill": {
     "duration": 159.457329,
     "end_time": "2024-01-15T06:26:32.754822",
     "exception": false,
     "start_time": "2024-01-15T06:23:53.297493",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting swig\r\n",
      "  Obtaining dependency information for swig from https://files.pythonhosted.org/packages/89/57/aec312f259b202185481a4c688d517dd62c40723695b3d04b851e682aa4e/swig-4.1.1.post1-py2.py3-none-manylinux_2_5_x86_64.manylinux1_x86_64.whl.metadata\r\n",
      "  Downloading swig-4.1.1.post1-py2.py3-none-manylinux_2_5_x86_64.manylinux1_x86_64.whl.metadata (3.6 kB)\r\n",
      "Downloading swig-4.1.1.post1-py2.py3-none-manylinux_2_5_x86_64.manylinux1_x86_64.whl (1.8 MB)\r\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.8/1.8 MB\u001b[0m \u001b[31m34.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\r\n",
      "\u001b[?25hInstalling collected packages: swig\r\n",
      "Successfully installed swig-4.1.1.post1\r\n",
      "Collecting box2d-py\r\n",
      "  Downloading box2d-py-2.3.8.tar.gz (374 kB)\r\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m374.5/374.5 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\r\n",
      "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l-\b \bdone\r\n",
      "\u001b[?25hBuilding wheels for collected packages: box2d-py\r\n",
      "  Building wheel for box2d-py (setup.py) ... \u001b[?25l-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \bdone\r\n",
      "\u001b[?25h  Created wheel for box2d-py: filename=box2d_py-2.3.8-cp310-cp310-linux_x86_64.whl size=494637 sha256=96a9e02341c53cece44073c4c766e4508b4805f49ba9057eccf8167fb5876a51\r\n",
      "  Stored in directory: /root/.cache/pip/wheels/47/01/d2/6a780da77ccb98b1d2facdd520a8d10838a03b590f6f8d50c0\r\n",
      "Successfully built box2d-py\r\n",
      "Installing collected packages: box2d-py\r\n",
      "Successfully installed box2d-py-2.3.8\r\n",
      "Requirement already satisfied: gym[Box_2D] in /opt/conda/lib/python3.10/site-packages (0.26.2)\r\n",
      "\u001b[33mWARNING: gym 0.26.2 does not provide the extra 'box_2d'\u001b[0m\u001b[33m\r\n",
      "\u001b[0mRequirement already satisfied: numpy>=1.18.0 in /opt/conda/lib/python3.10/site-packages (from gym[Box_2D]) (1.24.3)\r\n",
      "Requirement already satisfied: cloudpickle>=1.2.0 in /opt/conda/lib/python3.10/site-packages (from gym[Box_2D]) (2.2.1)\r\n",
      "Requirement already satisfied: gym-notices>=0.0.4 in /opt/conda/lib/python3.10/site-packages (from gym[Box_2D]) (0.0.8)\r\n",
      "Requirement already satisfied: gym[box2d] in /opt/conda/lib/python3.10/site-packages (0.26.2)\r\n",
      "Requirement already satisfied: numpy>=1.18.0 in /opt/conda/lib/python3.10/site-packages (from gym[box2d]) (1.24.3)\r\n",
      "Requirement already satisfied: cloudpickle>=1.2.0 in /opt/conda/lib/python3.10/site-packages (from gym[box2d]) (2.2.1)\r\n",
      "Requirement already satisfied: gym-notices>=0.0.4 in /opt/conda/lib/python3.10/site-packages (from gym[box2d]) (0.0.8)\r\n",
      "Collecting box2d-py==2.3.5 (from gym[box2d])\r\n",
      "  Downloading box2d-py-2.3.5.tar.gz (374 kB)\r\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m374.4/374.4 kB\u001b[0m \u001b[31m14.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\r\n",
      "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l-\b \bdone\r\n",
      "\u001b[?25hCollecting pygame==2.1.0 (from gym[box2d])\r\n",
      "  Downloading pygame-2.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (18.3 MB)\r\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m18.3/18.3 MB\u001b[0m \u001b[31m52.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\r\n",
      "\u001b[?25hRequirement already satisfied: swig==4.* in /opt/conda/lib/python3.10/site-packages (from gym[box2d]) (4.1.1.post1)\r\n",
      "Building wheels for collected packages: box2d-py\r\n",
      "  Building wheel for box2d-py (setup.py) ... \u001b[?25l-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \bdone\r\n",
      "\u001b[?25h  Created wheel for box2d-py: filename=box2d_py-2.3.5-cp310-cp310-linux_x86_64.whl size=494635 sha256=bab1f336f30556544bc917cc11b0c76a33259a4f372a0c2cb04b63ef994b68d2\r\n",
      "  Stored in directory: /root/.cache/pip/wheels/db/8f/6a/eaaadf056fba10a98d986f6dce954e6201ba3126926fc5ad9e\r\n",
      "Successfully built box2d-py\r\n",
      "Installing collected packages: box2d-py, pygame\r\n",
      "  Attempting uninstall: box2d-py\r\n",
      "    Found existing installation: box2d-py 2.3.8\r\n",
      "    Uninstalling box2d-py-2.3.8:\r\n",
      "      Successfully uninstalled box2d-py-2.3.8\r\n",
      "Successfully installed box2d-py-2.3.5 pygame-2.1.0\r\n"
     ]
    }
   ],
   "source": [
    "!pip3 install swig\n",
    "!pip3 install box2d-py\n",
    "!pip3 install gym[Box_2D]\n",
    "!pip install gym[box2d]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "db409021",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:26:32.776104Z",
     "iopub.status.busy": "2024-01-15T06:26:32.775657Z",
     "iopub.status.idle": "2024-01-15T06:26:32.868228Z",
     "shell.execute_reply": "2024-01-15T06:26:32.867037Z"
    },
    "id": "6CbiEYt9YyTr",
    "outputId": "51d25ef3-10ef-42a9-b235-9c9273f56d4d",
    "papermill": {
     "duration": 0.106802,
     "end_time": "2024-01-15T06:26:32.871173",
     "exception": false,
     "start_time": "2024-01-15T06:26:32.764371",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "env = gym.make('LunarLander-v2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ac67af20",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:26:32.893358Z",
     "iopub.status.busy": "2024-01-15T06:26:32.892915Z",
     "iopub.status.idle": "2024-01-15T06:26:32.904379Z",
     "shell.execute_reply": "2024-01-15T06:26:32.903318Z"
    },
    "id": "dLT6a5HGJZEe",
    "papermill": {
     "duration": 0.024929,
     "end_time": "2024-01-15T06:26:32.906842",
     "exception": false,
     "start_time": "2024-01-15T06:26:32.881913",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "class replayBuffer:\n",
    "  def __init__(self,maxSize,stateDim):\n",
    "    self.state=np.zeros((maxSize,stateDim))\n",
    "    self.action=np.zeros(maxSize,dtype= np.int8)\n",
    "    self.reward=np.zeros(maxSize)\n",
    "    self.done=np.zeros(maxSize,)\n",
    "    self.nextState=np.zeros((maxSize,stateDim))\n",
    "    self.maxSize=maxSize\n",
    "    self.curser=0\n",
    "    self.size=0\n",
    "\n",
    "  def save(self,state,action,reward,nextState,done):\n",
    "    self.state[self.curser]=state\n",
    "    self.action[self.curser]=action\n",
    "    self.reward[self.curser]=reward\n",
    "    self.nextState[self.curser]=nextState\n",
    "    self.done[self.curser]=done\n",
    "    self.curser=(self.curser+1)%self.maxSize\n",
    "    if self.size<self.maxSize:\n",
    "      self.size+=1\n",
    "\n",
    "  def sample(self,batchSize):\n",
    "    batchSize=min(self.size,batchSize-1)\n",
    "    indexes=np.random.choice([i for i in range(self.size-1)],batchSize)\n",
    "    return self.state[indexes],self.action[indexes],self.reward[indexes],self.nextState[indexes],self.done[indexes]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "952bd115",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:26:32.927911Z",
     "iopub.status.busy": "2024-01-15T06:26:32.927513Z",
     "iopub.status.idle": "2024-01-15T06:26:33.218542Z",
     "shell.execute_reply": "2024-01-15T06:26:33.217470Z"
    },
    "id": "Y37LHqC2RQEs",
    "papermill": {
     "duration": 0.30491,
     "end_time": "2024-01-15T06:26:33.221326",
     "exception": false,
     "start_time": "2024-01-15T06:26:32.916416",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "class Agent:\n",
    "  def __init__(self,stateShape,actionShape,exploreRate,exploreRateDecay,minimumExploreRate,gamma,copyNetsCycle):\n",
    "      self.gamma=gamma\n",
    "      self.exploreRate=exploreRate\n",
    "      self.exploreRateDecay=exploreRateDecay\n",
    "      self.minimumExploreRate=minimumExploreRate\n",
    "      self.actionShape=actionShape\n",
    "      self.memory=replayBuffer(1000000,stateShape)\n",
    "      self.model=self.buildModel(stateShape,actionShape)\n",
    "      self.model.compile(optimizer='Adam',loss='mse')\n",
    "      self.tModel=self.buildModel(stateShape,actionShape)\n",
    "      self.tModel.compile(optimizer='Adam',loss='mse')\n",
    "      self.learnThreshold=0\n",
    "      self.copyNetsCycle=copyNetsCycle\n",
    "\n",
    "  def buildModel(self,input,output):\n",
    "    inputLayer=keras.Input(shape=(input,))\n",
    "    layer=Dense(256,activation='relu')(inputLayer)\n",
    "    layer=Dense(256,activation='relu')(layer)\n",
    "    outputLayer=Dense(output)(layer)\n",
    "    model=keras.Model(inputs=inputLayer,outputs=outputLayer)\n",
    "    model.compile(optimizer='Adam',loss='mse')\n",
    "    return model\n",
    "\n",
    "  def getAction(self,state,evaluation_mode=False):\n",
    "    q=self.model(np.expand_dims(state,axis=0))[0].numpy()\n",
    "    if np.random.random()<=self.exploreRate and (not(evaluation_mode)):\n",
    "      return np.random.choice([i for i in range(env.action_space.n)])\n",
    "    else:\n",
    "      return np.argmax(q)\n",
    "\n",
    "  def exploreDecay(self):\n",
    "      self.exploreRate=max(self.exploreRate*self.exploreRateDecay,self.minimumExploreRate)\n",
    "\n",
    "  def saveModel(self,modelName=\"DoubleDQN_LunarLanderV2.h\"):\n",
    "      self.model.save_weights(f\"{modelName}\")\n",
    "\n",
    "  def loadModel(self,modelName=\"DoubleDQN_LunarLanderV2.h\"):\n",
    "      self.model.load_weights(f\"{modelName}\")\n",
    "      self.tModel.set_weights(self.model.get_weights())\n",
    "\n",
    "  def learn(self,batchSize=64):\n",
    "    if self.memory.size>batchSize:\n",
    "      states,actions,rewards,nextStates,done=self.memory.sample(batchSize)\n",
    "      qState=self.model(states).numpy()\n",
    "      qNextState=self.model(nextStates).numpy()\n",
    "      qNextStateTarget=self.tModel(nextStates).numpy()\n",
    "      maxActions=np.argmax(qNextState,axis=1)\n",
    "      batchIndex = np.arange(batchSize-1, dtype=np.int32)\n",
    "      qState[batchIndex,actions]=(rewards+(self.gamma*qNextStateTarget[batchIndex,maxActions.astype(int)]*(1-done)))\n",
    "      _=self.model.fit(x=states,y=qState,verbose=0)\n",
    "      self.learnThreshold+=1\n",
    "      self.exploreDecay()\n",
    "      if(self.learnThreshold%self.copyNetsCycle)==0:\n",
    "        self.tModel.set_weights(self.model.get_weights())\n",
    "        self.saveModel()\n",
    "        self.learnThreshold=0\n",
    "agent=Agent(stateShape=env.observation_space.shape[0],actionShape=env.action_space.n\\\n",
    "            ,exploreRate=0.01,exploreRateDecay=0.99,minimumExploreRate=0.01,gamma=0.99,copyNetsCycle=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b287c10a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:26:33.242895Z",
     "iopub.status.busy": "2024-01-15T06:26:33.242507Z",
     "iopub.status.idle": "2024-01-15T06:26:33.351742Z",
     "shell.execute_reply": "2024-01-15T06:26:33.350564Z"
    },
    "id": "M3fwkBPoSXHb",
    "papermill": {
     "duration": 0.12366,
     "end_time": "2024-01-15T06:26:33.354535",
     "exception": false,
     "start_time": "2024-01-15T06:26:33.230875",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "agent=Agent(stateShape=env.observation_space.shape[0],actionShape=env.action_space.n\\\n",
    "            ,exploreRate=1.0,exploreRateDecay=0.9995,minimumExploreRate=0.01,gamma=0.99,copyNetsCycle=100)\n",
    "# agent.loadModel(\"DoubleDQN_LunarLanderV2.h\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ef2c9fae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T06:26:33.375820Z",
     "iopub.status.busy": "2024-01-15T06:26:33.375362Z",
     "iopub.status.idle": "2024-01-15T08:54:05.123529Z",
     "shell.execute_reply": "2024-01-15T08:54:05.122300Z"
    },
    "id": "Ld9e0nIucmo_",
    "outputId": "49864d7f-efe5-43e7-e857-48165b58a2b9",
    "papermill": {
     "duration": 8851.875638,
     "end_time": "2024-01-15T08:54:05.239965",
     "exception": false,
     "start_time": "2024-01-15T06:26:33.364327",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.10/site-packages/gym/utils/passive_env_checker.py:233: DeprecationWarning: `np.bool8` is a deprecated alias for `np.bool_`.  (Deprecated NumPy 1.24)\n",
      "  if not isinstance(terminated, (bool, np.bool8)):\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode: 2   reward: -123.4838284048032  avg so far:-123.4838284048032 exploreRate:0.9860940917766235\n",
      "episode: 3   reward: -120.2480775877887  avg so far:-121.86595299629596 exploreRate:0.9333105749632893\n",
      "episode: 4   reward: -124.38843377713812  avg so far:-122.70677992324335 exploreRate:0.8985016297353923\n",
      "episode: 5   reward: -172.6867510698475  avg so far:-135.20177270989439 exploreRate:0.8542431659050345\n",
      "episode: 6   reward: -185.0293336390434  avg so far:-145.1672848957242 exploreRate:0.821971943634938\n",
      "episode: 7   reward: -37.67185613154821  avg so far:-127.25138010169486 exploreRate:0.7948853455740398\n",
      "episode: 8   reward: -268.06423176214395  avg so far:-147.36750176747333 exploreRate:0.7561088752427627\n",
      "episode: 9   reward: -425.7797816503208  avg so far:-182.16903675282924 exploreRate:0.717786647345383\n",
      "episode: 10   reward: -11.2537898917019  avg so far:-163.17845376825954 exploreRate:0.6834545255907595\n",
      "episode: 11   reward: -238.72555770217554  avg so far:-170.73316416165113 exploreRate:0.6484902655231009\n",
      "episode: 12   reward: -301.5535697860296  avg so far:-182.62592830932192 exploreRate:0.6025246707069989\n",
      "episode: 13   reward: -120.77281332623825  avg so far:-177.4715020607316 exploreRate:0.582378225659823\n",
      "episode: 14   reward: -119.89580741154116  avg so far:-173.04260247233236 exploreRate:0.5470851615600282\n",
      "episode: 15   reward: -372.0865336932594  avg so far:-187.26002613096998 exploreRate:0.510089893972876\n",
      "episode: 16   reward: -93.45911570357939  avg so far:-181.00663210247728 exploreRate:0.48496384592188013\n",
      "episode: 17   reward: -35.09151153467246  avg so far:-171.88693706698947 exploreRate:0.45762946982477415\n",
      "episode: 18   reward: -80.14800461689539  avg so far:-166.49052927580746 exploreRate:0.42668341995630865\n",
      "episode: 19   reward: -314.54025339957724  avg so far:-174.71551394935022 exploreRate:0.39406759337370495\n",
      "episode: 20   reward: -118.19994869634078  avg so far:-171.74101051498133 exploreRate:0.37353401519915735\n",
      "episode: 21   reward: -116.24634455781487  avg so far:-168.966277217123 exploreRate:0.35054643911827926\n",
      "episode: 22   reward: -82.73810805575839  avg so far:-164.8601739237247 exploreRate:0.3274961128836704\n",
      "episode: 23   reward: -64.89910345705493  avg so far:-160.31648890251242 exploreRate:0.30626766148829976\n",
      "episode: 24   reward: -92.52296798330724  avg so far:-157.3689445147209 exploreRate:0.28299795565300107\n",
      "episode: 25   reward: -64.2411211811071  avg so far:-153.488618542487 exploreRate:0.26544924862125363\n",
      "episode: 26   reward: -49.69368793715303  avg so far:-149.33682131827362 exploreRate:0.24824270402236964\n",
      "episode: 27   reward: -81.74180480107557  avg so far:-146.7370129906891 exploreRate:0.23331545165791032\n",
      "episode: 28   reward: 19.978629408119318  avg so far:-140.562359568511 exploreRate:0.2093200192862765\n",
      "episode: 29   reward: -105.79809762224113  avg so far:-139.32077878471566 exploreRate:0.19565392171799403\n",
      "episode: 30   reward: -52.19845779691747  avg so far:-136.31656081961916 exploreRate:0.16772217330870298\n",
      "episode: 31   reward: -74.80626993435051  avg so far:-134.2662177901102 exploreRate:0.14464346723516094\n",
      "episode: 32   reward: -49.99014180543845  avg so far:-131.54763469383047 exploreRate:0.12612047569284585\n",
      "episode: 33   reward: -74.09642389466256  avg so far:-129.75228435635645 exploreRate:0.10581515136579583\n",
      "episode: 34   reward: -63.25459392612083  avg so far:-127.7372028281675 exploreRate:0.0968023682492183\n",
      "episode: 35   reward: -111.5082942302834  avg so far:-127.25988198705326 exploreRate:0.08715125746578377\n",
      "episode: 36   reward: -21.55929648606306  avg so far:-124.23986525845355 exploreRate:0.08142059044946617\n",
      "episode: 37   reward: 17.415706373455663  avg so far:-120.3049882686783 exploreRate:0.0663256650993032\n",
      "episode: 38   reward: -13.633420152562465  avg so far:-117.42197291418867 exploreRate:0.05882382252580734\n",
      "episode: 39   reward: -53.95772668203037  avg so far:-115.75186117123715 exploreRate:0.05277406008946195\n",
      "episode: 40   reward: -18.807254372609236  avg so far:-113.26610202255438 exploreRate:0.04775075008886719\n",
      "episode: 41   reward: -90.1435829441037  avg so far:-112.68803904559311 exploreRate:0.044056576776103695\n",
      "episode: 42   reward: -344.58048658440794  avg so far:-118.34395240019836 exploreRate:0.0371488198889033\n",
      "episode: 43   reward: -143.4658351600295  avg so far:-118.94209246590862 exploreRate:0.035090054747437864\n",
      "episode: 44   reward: -34.86553444391481  avg so far:-116.98682367469947 exploreRate:0.03250515260355337\n",
      "episode: 45   reward: -33.83159474433364  avg so far:-115.09693210810023 exploreRate:0.03068838482755771\n",
      "episode: 46   reward: -15.02435242730786  avg so far:-112.87309700408264 exploreRate:0.028130727419394437\n",
      "episode: 47   reward: -63.5600690831217  avg so far:-111.80107465797478 exploreRate:0.026518635876378482\n",
      "episode: 48   reward: 15.586223248135681  avg so far:-109.09070661741924 exploreRate:0.02477489199529718\n",
      "episode: 49   reward: -43.31236160001143  avg so far:-107.72032442955657 exploreRate:0.023507454683176096\n",
      "episode: 50   reward: 5.090893789615464  avg so far:-105.41805466998163 exploreRate:0.022071823407932747\n",
      "episode: 51   reward: 2.6476389620282106  avg so far:-103.25674079734142 exploreRate:0.020703149341900956\n",
      "episode: 52   reward: 15.230328233738462  avg so far:-100.93346493398691 exploreRate:0.018920827951510052\n",
      "episode: 53   reward: -30.042970973491222  avg so far:-99.57018620397737 exploreRate:0.01747451339040307\n",
      "episode: 54   reward: -40.21244703906416  avg so far:-98.45022886124316 exploreRate:0.010597502922393745\n",
      "episode: 55   reward: -60.26020356540108  avg so far:-97.74300617057942 exploreRate:0.01\n",
      "episode: 56   reward: -229.1775756189617  avg so far:-100.13272561509547 exploreRate:0.01\n",
      "episode: 57   reward: -124.18460070220631  avg so far:-100.56222338450814 exploreRate:0.01\n",
      "episode: 58   reward: -132.32633617418344  avg so far:-101.11948852116912 exploreRate:0.01\n",
      "episode: 59   reward: -2.6926712781100406  avg so far:-99.42247443077154 exploreRate:0.01\n",
      "episode: 60   reward: -127.66049807723057  avg so far:-99.90108500105052 exploreRate:0.01\n",
      "episode: 61   reward: -52.15468270479044  avg so far:-99.10531162944618 exploreRate:0.01\n",
      "episode: 62   reward: -247.4750980113139  avg so far:-101.5376032094768 exploreRate:0.01\n",
      "episode: 63   reward: -88.40214517307751  avg so far:-101.32574098308326 exploreRate:0.01\n",
      "episode: 64   reward: -144.72454798438207  avg so far:-102.01461093548484 exploreRate:0.01\n",
      "episode: 65   reward: -162.36282009341164  avg so far:-102.95755170357745 exploreRate:0.01\n",
      "episode: 66   reward: -241.54922095640123  avg so far:-105.08973123054396 exploreRate:0.01\n",
      "episode: 67   reward: -132.91530760243324  avg so far:-105.51133087254229 exploreRate:0.01\n",
      "episode: 68   reward: -117.53966627859563  avg so far:-105.69085826666247 exploreRate:0.01\n",
      "episode: 69   reward: -113.82848094858642  avg so far:-105.81052918845548 exploreRate:0.01\n",
      "episode: 70   reward: -75.45749534832099  avg so far:-105.37063014729411 exploreRate:0.01\n",
      "episode: 71   reward: -201.9707178576109  avg so far:-106.75063140029863 exploreRate:0.01\n",
      "episode: 72   reward: -125.84534751789522  avg so far:-107.01957106392675 exploreRate:0.01\n",
      "episode: 73   reward: 146.21845925656942  avg so far:-103.5023761983643 exploreRate:0.01\n",
      "episode: 74   reward: -72.3793011776562  avg so far:-103.07603270492994 exploreRate:0.01\n",
      "episode: 75   reward: -51.74926894528299  avg so far:-102.38242778925904 exploreRate:0.01\n",
      "episode: 76   reward: 135.06297173539136  avg so far:-99.21648912893038 exploreRate:0.01\n",
      "episode: 77   reward: 93.9429292345178  avg so far:-96.67491783467447 exploreRate:0.01\n",
      "episode: 78   reward: -26.27937796599307  avg so far:-95.76069004417212 exploreRate:0.01\n",
      "episode: 79   reward: 223.86527249263636  avg so far:-91.66292129370022 exploreRate:0.01\n",
      "episode: 80   reward: 2.7618457448884612  avg so far:-90.46767107802188 exploreRate:0.01\n",
      "episode: 81   reward: -114.37936354501174  avg so far:-90.76656723385926 exploreRate:0.01\n",
      "episode: 82   reward: -4.965413173904286  avg so far:-89.70729372694623 exploreRate:0.01\n",
      "episode: 83   reward: 158.71869069253864  avg so far:-86.67770855109886 exploreRate:0.01\n",
      "episode: 84   reward: 179.82488709775214  avg so far:-83.46683390472715 exploreRate:0.01\n",
      "episode: 85   reward: -98.27096885519394  avg so far:-83.64307360651843 exploreRate:0.01\n",
      "episode: 86   reward: -48.58030069210426  avg so far:-83.23057039576061 exploreRate:0.01\n",
      "episode: 87   reward: -93.11430788989612  avg so far:-83.34549757592498 exploreRate:0.01\n",
      "episode: 88   reward: -67.88553032133538  avg so far:-83.16779680288371 exploreRate:0.01\n",
      "episode: 89   reward: 226.97088341654666  avg so far:-79.643493618572 exploreRate:0.01\n",
      "episode: 90   reward: -38.37155376780055  avg so far:-79.17976395732738 exploreRate:0.01\n",
      "episode: 91   reward: 222.41985785632707  avg so far:-75.82865704828677 exploreRate:0.01\n",
      "episode: 92   reward: -114.47854830394733  avg so far:-76.25338112801931 exploreRate:0.01\n",
      "episode: 93   reward: -102.34688684383043  avg so far:-76.5370061901477 exploreRate:0.01\n",
      "episode: 94   reward: -78.43013325382164  avg so far:-76.55736239513344 exploreRate:0.01\n",
      "episode: 95   reward: -57.36324728630183  avg so far:-76.35316968120969 exploreRate:0.01\n",
      "episode: 96   reward: -108.2988989947929  avg so far:-76.68944051608952 exploreRate:0.01\n",
      "episode: 97   reward: 24.63757233901933  avg so far:-75.6339507988488 exploreRate:0.01\n",
      "episode: 98   reward: -56.51819178492001  avg so far:-75.43688111829283 exploreRate:0.01\n",
      "episode: 99   reward: -95.59611355176675  avg so far:-75.64258757169563 exploreRate:0.01\n",
      "episode: 100   reward: -48.1395787626832  avg so far:-75.36477940190763 exploreRate:0.01\n",
      "episode: 101   reward: -158.56379114280878  avg so far:-76.19676951931663 exploreRate:0.01\n",
      "episode: 102   reward: 3.8384166714230616  avg so far:-75.4043419332697 exploreRate:0.01\n",
      "episode: 103   reward: -73.48217446295344  avg so far:-75.38549715414895 exploreRate:0.01\n",
      "episode: 104   reward: 205.57720318751095  avg so far:-72.65770394694837 exploreRate:0.01\n",
      "episode: 105   reward: -336.93459489380353  avg so far:-75.19882789836043 exploreRate:0.01\n",
      "episode: 106   reward: 229.78209717153396  avg so far:-72.29424765959955 exploreRate:0.01\n",
      "episode: 107   reward: -77.51533604138265  avg so far:-72.34350321037108 exploreRate:0.01\n",
      "episode: 108   reward: -71.89027911755697  avg so far:-72.3392674711859 exploreRate:0.01\n",
      "episode: 109   reward: 9.393330383748093  avg so far:-71.58248415771429 exploreRate:0.01\n",
      "episode: 110   reward: 7.341496780068098  avg so far:-70.8584109381016 exploreRate:0.01\n",
      "episode: 111   reward: -157.64585323736137  avg so far:-71.6473876862767 exploreRate:0.01\n",
      "episode: 112   reward: 198.420590902853  avg so far:-69.21434283412238 exploreRate:0.01\n",
      "episode: 113   reward: 196.8990041089059  avg so far:-66.83833080784534 exploreRate:0.01\n",
      "episode: 114   reward: -77.43492946532764  avg so far:-66.93210601720358 exploreRate:0.01\n",
      "episode: 115   reward: 3.935061363606451  avg so far:-66.31046419807367 exploreRate:0.01\n",
      "episode: 116   reward: -29.80371069665881  avg so far:-65.99301416762658 exploreRate:0.01\n",
      "episode: 117   reward: -51.81737308409937  avg so far:-65.87081036518238 exploreRate:0.01\n",
      "episode: 118   reward: -335.40806568936665  avg so far:-68.17454759017541 exploreRate:0.01\n",
      "episode: 119   reward: -15.657469559774675  avg so far:-67.72948760686693 exploreRate:0.01\n",
      "episode: 120   reward: 275.90033920472126  avg so far:-64.8418420034082 exploreRate:0.01\n",
      "episode: 121   reward: -8.106385112121075  avg so far:-64.36904652931415 exploreRate:0.01\n",
      "episode: 122   reward: 216.76025006153472  avg so far:-62.045663912860846 exploreRate:0.01\n",
      "episode: 123   reward: -1.4695334393046693  avg so far:-61.54913825324153 exploreRate:0.01\n",
      "episode: 124   reward: -43.09710161369794  avg so far:-61.39912169519646 exploreRate:0.01\n",
      "episode: 125   reward: -37.324038191650274  avg so far:-61.20496779597432 exploreRate:0.01\n",
      "episode: 126   reward: -69.47906154805904  avg so far:-61.271160545990995 exploreRate:0.01\n",
      "episode: 127   reward: 267.52066787368915  avg so far:-58.66170159027924 exploreRate:0.01\n",
      "episode: 128   reward: -46.66153241536894  avg so far:-58.56721206921696 exploreRate:0.01\n",
      "episode: 129   reward: -47.41841036086162  avg so far:-58.48011205587043 exploreRate:0.01\n",
      "episode: 130   reward: 249.02664068538203  avg so far:-56.09633877880646 exploreRate:0.01\n",
      "episode: 131   reward: 34.847613492433624  avg so far:-55.39676991518153 exploreRate:0.01\n",
      "episode: 132   reward: 249.52656293018734  avg so far:-53.06911088582757 exploreRate:0.01\n",
      "episode: 133   reward: -404.07223866339547  avg so far:-55.72822549020308 exploreRate:0.01\n",
      "episode: 134   reward: -28.56523260464428  avg so far:-55.52399246098835 exploreRate:0.01\n",
      "episode: 135   reward: 218.48126621603146  avg so far:-53.47917709772701 exploreRate:0.01\n",
      "episode: 136   reward: -275.136384106125  avg so far:-55.121082334826255 exploreRate:0.01\n",
      "episode: 137   reward: -112.95062299301904  avg so far:-55.546299545548266 exploreRate:0.01\n",
      "episode: 138   reward: 7.831569389439409  avg so far:-55.08368736354105 exploreRate:0.01\n",
      "episode: 139   reward: -19.76979524588242  avg so far:-54.82778959457251 exploreRate:0.01\n",
      "episode: 140   reward: -596.9429072142426  avg so far:-58.72789835442625 exploreRate:0.01\n",
      "episode: 141   reward: -373.91560164250524  avg so far:-60.97923909219825 exploreRate:0.01\n",
      "episode: 142   reward: -561.6197271346612  avg so far:-64.52988085136465 exploreRate:0.01\n",
      "episode: 143   reward: -103.58509614997791  avg so far:-64.80491757881967 exploreRate:0.01\n",
      "episode: 144   reward: -222.8742601187646  avg so far:-65.91029759658153 exploreRate:0.01\n",
      "episode: 145   reward: -681.0385860006248  avg so far:-70.1820218216096 exploreRate:0.01\n",
      "episode: 146   reward: -72.1566940587703  avg so far:-70.1956402508314 exploreRate:0.01\n",
      "episode: 147   reward: -29.79129292294074  avg so far:-69.91889814584584 exploreRate:0.01\n",
      "episode: 148   reward: 201.13615130871852  avg so far:-68.07498624479439 exploreRate:0.01\n",
      "episode: 149   reward: -414.46013950073075  avg so far:-70.41542646949667 exploreRate:0.01\n",
      "episode: 150   reward: -320.3842190682818  avg so far:-72.09306937284423 exploreRate:0.01\n",
      "episode: 151   reward: 200.36070766421315  avg so far:-70.27671085926384 exploreRate:0.01\n",
      "episode: 152   reward: -65.94109425162843  avg so far:-70.24799816649804 exploreRate:0.01\n",
      "episode: 153   reward: -100.50807329279463  avg so far:-70.4470776081184 exploreRate:0.01\n",
      "episode: 154   reward: -52.71018235337198  avg so far:-70.33115018815275 exploreRate:0.01\n",
      "episode: 155   reward: -70.7745098032415  avg so far:-70.33402914669229 exploreRate:0.01\n",
      "episode: 156   reward: -53.956772930756216  avg so far:-70.22836942917012 exploreRate:0.01\n",
      "episode: 157   reward: 224.01605538334638  avg so far:-68.34218721883347 exploreRate:0.01\n",
      "episode: 158   reward: 208.2362241631032  avg so far:-66.58054128646444 exploreRate:0.01\n",
      "episode: 159   reward: 136.88240890826702  avg so far:-65.29280109535856 exploreRate:0.01\n",
      "episode: 160   reward: -5.739446117453227  avg so far:-64.9182516929818 exploreRate:0.01\n",
      "episode: 161   reward: -324.10801296854004  avg so far:-66.53818770095403 exploreRate:0.01\n",
      "episode: 162   reward: -81.8440701377666  avg so far:-66.63325529372926 exploreRate:0.01\n",
      "episode: 163   reward: 212.05408233346049  avg so far:-64.91296308615401 exploreRate:0.01\n",
      "episode: 164   reward: 267.9580520187974  avg so far:-62.87080961925246 exploreRate:0.01\n",
      "episode: 165   reward: -55.041190962239625  avg so far:-62.82306804207556 exploreRate:0.01\n",
      "episode: 166   reward: 13.237372789318613  avg so far:-62.36209567340044 exploreRate:0.01\n",
      "episode: 167   reward: 216.9954926433739  avg so far:-60.679218635347574 exploreRate:0.01\n",
      "episode: 168   reward: -55.65583621453549  avg so far:-60.649138501091215 exploreRate:0.01\n",
      "episode: 169   reward: 162.912944523963  avg so far:-59.31841181641827 exploreRate:0.01\n",
      "episode: 170   reward: 254.753032344092  avg so far:-57.46000090422591 exploreRate:0.01\n",
      "episode: 171   reward: -75.02932149243166  avg so far:-57.56334984886241 exploreRate:0.01\n",
      "episode: 172   reward: -282.9065350059629  avg so far:-58.8811462532899 exploreRate:0.01\n",
      "episode: 173   reward: 213.71724183625787  avg so far:-57.296271903932066 exploreRate:0.01\n",
      "episode: 174   reward: 266.55773808438295  avg so far:-55.424283406889785 exploreRate:0.01\n",
      "episode: 175   reward: 225.42922792967155  avg so far:-53.81018276702449 exploreRate:0.01\n",
      "episode: 176   reward: 192.76370385777662  avg so far:-52.401189129168486 exploreRate:0.01\n",
      "episode: 177   reward: 104.21345336244012  avg so far:-51.511333205920714 exploreRate:0.01\n",
      "episode: 178   reward: 29.91544104689573  avg so far:-51.05129493330593 exploreRate:0.01\n",
      "episode: 179   reward: 257.127358374307  avg so far:-49.31995418438675 exploreRate:0.01\n",
      "episode: 180   reward: 276.42144259947366  avg so far:-47.50016984481211 exploreRate:0.01\n",
      "episode: 181   reward: 257.76834759652206  avg so far:-45.804233636804696 exploreRate:0.01\n",
      "episode: 182   reward: 31.16572628903474  avg so far:-45.37898523942437 exploreRate:0.01\n",
      "episode: 183   reward: 225.32582594584682  avg so far:-43.89159616697782 exploreRate:0.01\n",
      "episode: 184   reward: -98.01051817014667  avg so far:-44.187327981202785 exploreRate:0.01\n",
      "episode: 185   reward: 274.5143616356618  avg so far:-42.45525358111113 exploreRate:0.01\n",
      "episode: 186   reward: -136.17144422171037  avg so far:-42.96182758457383 exploreRate:0.01\n",
      "episode: 187   reward: 276.21540464602595  avg so far:-41.245820959678134 exploreRate:0.01\n",
      "episode: 188   reward: 245.19628848502418  avg so far:-39.71404497334282 exploreRate:0.01\n",
      "episode: 189   reward: 243.07391947810862  avg so far:-38.20985367306914 exploreRate:0.01\n",
      "episode: 190   reward: 250.69775723264772  avg so far:-36.681241975155295 exploreRate:0.01\n",
      "episode: 191   reward: 169.66571897320824  avg so far:-35.59520533858496 exploreRate:0.01\n",
      "episode: 192   reward: -11.558316915126696  avg so far:-35.46935775521607 exploreRate:0.01\n",
      "episode: 193   reward: -38.487251198601015  avg so far:-35.4850759502337 exploreRate:0.01\n",
      "episode: 194   reward: 261.6131314367452  avg so far:-33.945707000042106 exploreRate:0.01\n",
      "episode: 195   reward: 88.68414597562632  avg so far:-33.31359435583763 exploreRate:0.01\n",
      "episode: 196   reward: 246.26444312937312  avg so far:-31.879860830272445 exploreRate:0.01\n",
      "episode: 197   reward: 251.92505409186438  avg so far:-30.431876570465622 exploreRate:0.01\n",
      "episode: 198   reward: 231.71571436577486  avg so far:-29.101178139317195 exploreRate:0.01\n",
      "episode: 199   reward: 272.2123657171065  avg so far:-27.57939256428475 exploreRate:0.01\n",
      "episode: 200   reward: 222.4710808984072  avg so far:-26.32285752175866 exploreRate:0.01\n"
     ]
    }
   ],
   "source": [
    "averageRewards=[]\n",
    "totalRewards=[]\n",
    "for i in range(1,200):\n",
    "  done = False\n",
    "  truncuated = False\n",
    "  state,_=env.reset()\n",
    "  rewards=0\n",
    "  while (not done) and (not truncuated):\n",
    "    action=agent.getAction(state)\n",
    "    nextState,reward,done,truncuated,_=env.step(action)\n",
    "    agent.memory.save(state,action,reward,nextState,int(done))\n",
    "    rewards+=reward\n",
    "    state=nextState\n",
    "    agent.learn(batchSize=64)\n",
    "  totalRewards.append(rewards)\n",
    "  averageRewards.append(sum(totalRewards)/len(totalRewards))\n",
    "  print(f\"episode: {i+1}   reward: {rewards}  avg so far:{averageRewards[-1]} exploreRate:{agent.exploreRate}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c484b686",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T08:54:05.302356Z",
     "iopub.status.busy": "2024-01-15T08:54:05.301870Z",
     "iopub.status.idle": "2024-01-15T08:54:05.953590Z",
     "shell.execute_reply": "2024-01-15T08:54:05.952182Z"
    },
    "id": "rFBZ-TCb8m48",
    "papermill": {
     "duration": 0.687277,
     "end_time": "2024-01-15T08:54:05.956635",
     "exception": false,
     "start_time": "2024-01-15T08:54:05.269358",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(f'Total Rewards')\n",
    "# plt.yscale('symlog')\n",
    "plt.plot(totalRewards)\n",
    "plt.savefig(\"Total Rewards\",dpi=200)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "cf22c9c8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T08:54:06.017757Z",
     "iopub.status.busy": "2024-01-15T08:54:06.017341Z",
     "iopub.status.idle": "2024-01-15T08:54:06.623485Z",
     "shell.execute_reply": "2024-01-15T08:54:06.622583Z"
    },
    "papermill": {
     "duration": 0.639388,
     "end_time": "2024-01-15T08:54:06.625964",
     "exception": false,
     "start_time": "2024-01-15T08:54:05.986576",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.clf()\n",
    "plt.title(f'Average Rewards')\n",
    "# plt.yscale('symlog')\n",
    "plt.plot(averageRewards)\n",
    "plt.savefig(\"Average Rewards\",dpi=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6fc1de48",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-01-15T08:54:06.687231Z",
     "iopub.status.busy": "2024-01-15T08:54:06.686746Z",
     "iopub.status.idle": "2024-01-15T08:54:44.383064Z",
     "shell.execute_reply": "2024-01-15T08:54:44.381961Z"
    },
    "papermill": {
     "duration": 37.730569,
     "end_time": "2024-01-15T08:54:44.386130",
     "exception": false,
     "start_time": "2024-01-15T08:54:06.655561",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Test episode: 1   gained reward: 256.3287448846416\n",
      " Test episode: 2   gained reward: -266.5255463715048\n",
      " Test episode: 3   gained reward: -83.38799896225918\n",
      " Test episode: 4   gained reward: -209.25251056503032\n",
      " Test episode: 5   gained reward: 248.78689041237715\n",
      " Test episode: 6   gained reward: 237.58942444329242\n",
      " Test episode: 7   gained reward: 205.72654972013942\n",
      " Test episode: 8   gained reward: 131.55612561734586\n",
      " Test episode: 9   gained reward: 177.84647690540885\n",
      " Test episode: 10   gained reward: 266.44505495238855\n",
      " Test episode: 11   gained reward: 199.5909098097983\n",
      " Test episode: 12   gained reward: 262.34798168296544\n",
      " Test episode: 13   gained reward: 247.7073683864242\n",
      " Test episode: 14   gained reward: 227.58019588952845\n",
      " Test episode: 15   gained reward: -269.1541025698533\n",
      " Test episode: 16   gained reward: 261.7617897090562\n",
      " Test episode: 17   gained reward: -10.381268442657055\n",
      " Test episode: 18   gained reward: 283.84203770769\n",
      " Test episode: 19   gained reward: 210.27934990423626\n",
      " Test episode: 20   gained reward: 261.44985496030625\n",
      "average of evaluation rewards: 132.0068664037147\n"
     ]
    }
   ],
   "source": [
    "# test loop for 10 diffrent episodeds\n",
    "sum_evaluation_rewards=0\n",
    "for i in range(0,20):\n",
    "  done = False\n",
    "  truncuated = False\n",
    "  state,_=env.reset()\n",
    "  rewards=0\n",
    "  while (not done) and (not truncuated):\n",
    "    action=agent.getAction(state,True)\n",
    "    nextState,reward,done,truncuated,_=env.step(action)\n",
    "    rewards+=reward\n",
    "    state=nextState\n",
    "    # no learning we want to test the model\n",
    "  print(f\" Test episode: {i+1}   gained reward: {rewards}\")\n",
    "  sum_evaluation_rewards+=rewards\n",
    "print(f\"average of evaluation rewards: {sum_evaluation_rewards/20}\")"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "kaggle": {
   "accelerator": "none",
   "dataSources": [],
   "dockerImageVersionId": 30635,
   "isGpuEnabled": false,
   "isInternetEnabled": true,
   "language": "python",
   "sourceType": "notebook"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 9074.860975,
   "end_time": "2024-01-15T08:54:49.639046",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2024-01-15T06:23:34.778071",
   "version": "2.4.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}