lunarLander/lunarLander3Continuous.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Enable this cell if using Google Collab:\n",
    "\n",
    "#!pip install box2d-py\n",
    "#!pip install gym[Box_2D]\n",
    "\n",
    "# RESTART RUNTIME AFTER INSTALLING ON GOOGLE COLLAB"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **LunarLander-v2 Continuous (maximum 1000 steps)**\n",
    "#### Actor-Critic Architecture. The actor is in charge of the policy wheras the critic provides the actor with the state-action values.\n",
    "\n",
    "#### The input to the network consists of 8 float values that specify the coordinates of the agent, the speed in each direction, the angle...\n",
    "#### The outputs of the actor are two floats which control the intesity of the main engine (from -1 to 0: the engine is off and from 0 till 1: the power increases) and the relation between left and right engine (near -1: left engine is on and near 1: right engine is on)\n",
    "##### The output of the actor goes through a tanh activation function that returns values between -1 and 1. The critic gets the state and the action as the input and returns the Q-value without an activation function.\n",
    "##### **This environment is considered to be solved when getting an average of 200 over 100 consecutive episodes.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_pRiOxFrGenT"
   },
   "outputs": [],
   "source": [
    "# Imports the libraries:\n",
    "\n",
    "import random\n",
    "import numpy as np\n",
    "from tensorflow import keras\n",
    "import tensorflow as tf\n",
    "import gym\n",
    "env = gym.make('LunarLanderContinuous-v2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "TK2__aerF1VZ"
   },
   "outputs": [],
   "source": [
    "# Creating the actor:\n",
    "\n",
    "stateInput        = keras.Input(shape = (8,))    \n",
    "\n",
    "actorDense1       = keras.layers.Dense(units = 128, activation = \"relu\")   (stateInput)\n",
    "actorDense2       = keras.layers.Dense(units = 128, activation = \"relu\")   (actorDense1)\n",
    "actorDense3       = keras.layers.Dense(units = 32, activation = \"relu\")    (actorDense2)\n",
    "actorOutput       = keras.layers.Dense(units = 2, activation = \"tanh\")     (actorDense3)\n",
    "\n",
    "actorModel        = keras.Model (inputs = stateInput, outputs = actorOutput)\n",
    "actorModel.compile(optimizer=keras.optimizers.Adam(lr=0.001), loss =\"mse\")\n",
    "\n",
    "#print (actorModel.summary())\n",
    "\n",
    "# Creating the critic:\n",
    "\n",
    "stateDense1       = keras.layers.Dense(units = 128, activation=\"relu\")      (stateInput)\n",
    "stateDense2       = keras.layers.Dense(units = 64, activation=\"relu\")       (stateDense1)\n",
    "actionInput       = keras.Input(shape = (2,), name=\"actionInput\")                            \n",
    "actionDense1      = keras.layers.Dense(units = 64, activation=\"relu\")       (actionInput)       \n",
    "\n",
    "criticConcatenate = keras.layers.Concatenate()                              ([stateDense2, actionDense1])\n",
    "criticDense1      = keras.layers.Dense(units = 128, activation = \"relu\")    (criticConcatenate)\n",
    "criticDense2      = keras.layers.Dense(units = 128, activation = \"relu\")    (criticDense1)\n",
    "criticOutput      = keras.layers.Dense(units = 1)                           (criticDense2)\n",
    "\n",
    "criticModel = keras.Model(inputs = [stateInput, actionInput] , outputs = criticOutput )\n",
    "criticModel.compile(optimizer = keras.optimizers.Adam(lr=0.001), loss = \"mse\")\n",
    "\n",
    "\n",
    "# Creating the target networks for both the actor and the critic:\n",
    "\n",
    "staticActorModel  = keras.models.clone_model( actorModel, input_tensors=None, clone_function=None)\n",
    "staticCriticModel = keras.models.clone_model( criticModel, input_tensors=None, clone_function=None)\n",
    "\n",
    "\n",
    "\n",
    "#print (criticModel.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "V4Kqs94UFkGw"
   },
   "outputs": [],
   "source": [
    "# Function used to soft update the static Model weights for any given Tau:\n",
    "\n",
    "def updateStaticWeights(model, staticModel,tau):\n",
    "    modelWeights = model.get_weights()\n",
    "    staticWeights = staticModel.get_weights()\n",
    "    \n",
    "    #Python broadcasting:\n",
    "    \n",
    "    for i in range (len(staticWeights)):\n",
    "     \n",
    "        staticWeights[i] = (1-tau) * staticWeights[i] + tau * modelWeights[i]\n",
    "\n",
    "        \n",
    "    staticModel.set_weights(staticWeights)\n",
    "\n",
    "\n",
    "    return model, staticModel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_J8A_ghTFiXu"
   },
   "outputs": [],
   "source": [
    "# Memory Replay is carried out in this function.\n",
    "# 256 random previous experiences are \"remembered\" and processed in each step with a batch size of 32\n",
    "# This way we can prevent overfitting most recent states:\n",
    "\n",
    "def executeMemoryReplayStep(memoryPreviousStates, memoryStates, memoryActions, memoryRD, actorModel, staticActorModel, criticModel, staticCriticModel, itemsInMemory):\n",
    "    randomList = np.random.permutation(itemsInMemory) # After that number everything is 0\n",
    "    \n",
    "    memoryPreviousStates = memoryPreviousStates[randomList[:numberOfLearningFromReplaySteps]]\n",
    "    memoryStates = memoryStates[randomList[:numberOfLearningFromReplaySteps]]\n",
    "    memoryRD = memoryRD[randomList[:numberOfLearningFromReplaySteps]]\n",
    "    memoryActions = memoryActions[randomList[:numberOfLearningFromReplaySteps]]\n",
    "\n",
    "    \n",
    "    previousState = memoryPreviousStates\n",
    "    state = memoryStates\n",
    "    actions = memoryActions\n",
    "    reward, done = np.split(memoryRD,2,axis=1)\n",
    "    \n",
    "\n",
    "    \n",
    "    # Critic Model Training:\n",
    "    \n",
    "    futureActions       = staticActorModel.predict (state)\n",
    "\n",
    "    valueState          = staticCriticModel.predict ([state, futureActions])\n",
    "\n",
    "    valuePreviousState = reward + discount * valueState * (1-done) #for those rows in which done is False:\n",
    "\n",
    "    criticModel.fit ( x = [previousState, actions], y = valuePreviousState, epochs=1, verbose = 0, batch_size =32)\n",
    "\n",
    "\n",
    "    \n",
    "    # Actor Model Training:    \n",
    "    \n",
    "    with tf.GradientTape() as tape:\n",
    "        tape.watch(actorModel.trainable_weights)\n",
    "        \n",
    "        predictedActions = actorModel(previousState, training=False)\n",
    " \n",
    "        criticPreds = criticModel ([previousState, predictedActions],  training=False)\n",
    "        \n",
    "    fromCriticToActorGrads = tape.gradient(criticPreds, actorModel.trainable_weights)\n",
    "    fromCriticToActorGrads = -np.array(fromCriticToActorGrads) # We are trying to do gradient ascent\n",
    "    \n",
    "    gradsAndVariables = zip (fromCriticToActorGrads, actorModel.trainable_weights)\n",
    "    actorModel.optimizer.apply_gradients(gradsAndVariables)\n",
    "\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "WuT4-qWZFnCQ",
    "outputId": "e4b24f8d-8df6-4ffe-b4f4-9f28d23d8f1a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "STATIC WEIGHTS UPDATED\n",
      "0  Episode Reward:  -260.87831364163935 .Epsilon: 0.995\n",
      "1  Episode Reward:  -224.85470477814556 .Epsilon: 0.990025\n",
      "2  Episode Reward:  -325.16132679957786 .Epsilon: 0.985074875\n",
      "3  Episode Reward:  3.0691539368713308 .Epsilon: 0.9801495006250001\n",
      "4  Episode Reward:  -83.92501181423088 .Epsilon: 0.9752487531218751\n",
      "5  Episode Reward:  -113.99702657990771 .Epsilon: 0.9703725093562657\n",
      "6  Episode Reward:  -193.2320747910002 .Epsilon: 0.9655206468094844\n",
      "7  Episode Reward:  -217.52356495469718 .Epsilon: 0.960693043575437\n",
      "8  Episode Reward:  -389.7050604852948 .Epsilon: 0.9558895783575597\n",
      "9  Episode Reward:  -314.5162666885744 .Epsilon: 0.9511101304657719\n",
      "10  Episode Reward:  -396.6471164906467 .Epsilon: 0.946354579813443\n",
      "11  Episode Reward:  -326.9619693952766 .Epsilon: 0.9416228069143757\n",
      "12  Episode Reward:  -400.12902921328805 .Epsilon: 0.9369146928798039\n",
      "13  Episode Reward:  -168.36112452806907 .Epsilon: 0.9322301194154049\n",
      "14  Episode Reward:  -184.63011420199234 .Epsilon: 0.9275689688183278\n",
      "15  Episode Reward:  -307.0358703553545 .Epsilon: 0.9229311239742362\n",
      "16  Episode Reward:  -68.03336575062785 .Epsilon: 0.918316468354365\n",
      "17  Episode Reward:  -77.19470501461666 .Epsilon: 0.9137248860125932\n",
      "18  Episode Reward:  -200.8744635903006 .Epsilon: 0.9091562615825302\n",
      "19  Episode Reward:  -487.2261243724888 .Epsilon: 0.9046104802746175\n",
      "20  Episode Reward:  -76.26796466669242 .Epsilon: 0.9000874278732445\n",
      "STATIC WEIGHTS UPDATED\n",
      "21  Episode Reward:  -92.8026144965555 .Epsilon: 0.8955869907338783\n",
      "22  Episode Reward:  -92.75415837562448 .Epsilon: 0.8911090557802088\n",
      "23  Episode Reward:  -71.59662940986131 .Epsilon: 0.8866535105013078\n",
      "24  Episode Reward:  -115.89595603459344 .Epsilon: 0.8822202429488013\n",
      "25  Episode Reward:  -267.95534754605336 .Epsilon: 0.8778091417340573\n",
      "26  Episode Reward:  -137.44116306565365 .Epsilon: 0.8734200960253871\n",
      "27  Episode Reward:  -23.239056726407085 .Epsilon: 0.8690529955452602\n",
      "28  Episode Reward:  -124.34929072866996 .Epsilon: 0.8647077305675338\n",
      "29  Episode Reward:  -305.17497406289385 .Epsilon: 0.8603841919146962\n",
      "30  Episode Reward:  -70.5622095709242 .Epsilon: 0.8560822709551227\n",
      "31  Episode Reward:  -82.5088231090477 .Epsilon: 0.851801859600347\n",
      "32  Episode Reward:  -30.300419788608693 .Epsilon: 0.8475428503023453\n",
      "33  Episode Reward:  -128.40688799241883 .Epsilon: 0.8433051360508336\n",
      "34  Episode Reward:  -49.59872352263387 .Epsilon: 0.8390886103705794\n",
      "35  Episode Reward:  -13.335467193820634 .Epsilon: 0.8348931673187264\n",
      "36  Episode Reward:  -9.099018686614812 .Epsilon: 0.8307187014821328\n",
      "37  Episode Reward:  -138.99185507904252 .Epsilon: 0.8265651079747222\n",
      "38  Episode Reward:  -151.8248556429539 .Epsilon: 0.8224322824348486\n",
      "39  Episode Reward:  -206.9870471200481 .Epsilon: 0.8183201210226743\n",
      "40  Episode Reward:  -78.12302606428284 .Epsilon: 0.8142285204175609\n",
      "41  Episode Reward:  -44.54159264442617 .Epsilon: 0.810157377815473\n",
      "STATIC WEIGHTS UPDATED\n",
      "42  Episode Reward:  -44.67821983354475 .Epsilon: 0.8061065909263957\n",
      "43  Episode Reward:  -58.553150134763754 .Epsilon: 0.8020760579717637\n",
      "44  Episode Reward:  -39.77369667321442 .Epsilon: 0.798065677681905\n",
      "45  Episode Reward:  -52.392986443060124 .Epsilon: 0.7940753492934954\n",
      "46  Episode Reward:  -80.98238825265413 .Epsilon: 0.7901049725470279\n",
      "47  Episode Reward:  -27.76881510495184 .Epsilon: 0.7861544476842928\n",
      "48  Episode Reward:  33.72485564343677 .Epsilon: 0.7822236754458713\n",
      "49  Episode Reward:  -227.60714728822092 .Epsilon: 0.778312557068642\n",
      "50  Episode Reward:  -137.38490299226774 .Epsilon: 0.7744209942832988\n",
      "51  Episode Reward:  -72.46891625458727 .Epsilon: 0.7705488893118823\n",
      "52  Episode Reward:  -283.27690964431963 .Epsilon: 0.7666961448653229\n",
      "53  Episode Reward:  -46.767476665365095 .Epsilon: 0.7628626641409962\n",
      "54  Episode Reward:  -259.19028446010077 .Epsilon: 0.7590483508202912\n",
      "55  Episode Reward:  -79.4552845316657 .Epsilon: 0.7552531090661897\n",
      "56  Episode Reward:  -100.16872095742643 .Epsilon: 0.7514768435208588\n",
      "57  Episode Reward:  -327.31516444916787 .Epsilon: 0.7477194593032545\n",
      "58  Episode Reward:  -203.99389869897695 .Epsilon: 0.7439808620067382\n",
      "59  Episode Reward:  -210.1184225458585 .Epsilon: 0.7402609576967045\n",
      "60  Episode Reward:  5.60187669901218 .Epsilon: 0.736559652908221\n",
      "61  Episode Reward:  -33.84339368366955 .Epsilon: 0.7328768546436799\n",
      "STATIC WEIGHTS UPDATED\n",
      "62  Episode Reward:  -13.641586988567752 .Epsilon: 0.7292124703704616\n",
      "63  Episode Reward:  -38.04742854643992 .Epsilon: 0.7255664080186093\n",
      "64  Episode Reward:  -84.04686577089666 .Epsilon: 0.7219385759785162\n",
      "65  Episode Reward:  -315.2576026773696 .Epsilon: 0.7183288830986236\n",
      "66  Episode Reward:  21.169158501628235 .Epsilon: 0.7147372386831305\n",
      "67  Episode Reward:  -34.113529736355034 .Epsilon: 0.7111635524897149\n",
      "68  Episode Reward:  -244.44219969606104 .Epsilon: 0.7076077347272662\n",
      "69  Episode Reward:  -196.23702436037752 .Epsilon: 0.7040696960536299\n",
      "70  Episode Reward:  -255.6089929588509 .Epsilon: 0.7005493475733617\n",
      "71  Episode Reward:  -208.6602165695482 .Epsilon: 0.697046600835495\n",
      "72  Episode Reward:  -163.1560774920684 .Epsilon: 0.6935613678313175\n",
      "73  Episode Reward:  -40.09330593020548 .Epsilon: 0.6900935609921609\n",
      "74  Episode Reward:  -237.72489072220264 .Epsilon: 0.6866430931872001\n",
      "STATIC WEIGHTS UPDATED\n",
      "75  Episode Reward:  -42.90186815379459 .Epsilon: 0.6832098777212641\n",
      "76  Episode Reward:  -53.1406311932371 .Epsilon: 0.6797938283326578\n",
      "77  Episode Reward:  -631.697717806711 .Epsilon: 0.6763948591909945\n",
      "78  Episode Reward:  -106.99550839270152 .Epsilon: 0.6730128848950395\n",
      "79  Episode Reward:  -452.43318185765065 .Epsilon: 0.6696478204705644\n",
      "80  Episode Reward:  -226.72149351479197 .Epsilon: 0.6662995813682115\n",
      "81  Episode Reward:  -46.92525973716788 .Epsilon: 0.6629680834613705\n",
      "82  Episode Reward:  20.713494964783965 .Epsilon: 0.6596532430440636\n",
      "83  Episode Reward:  -122.81668511544761 .Epsilon: 0.6563549768288433\n",
      "STATIC WEIGHTS UPDATED\n",
      "84  Episode Reward:  -60.949379624840475 .Epsilon: 0.653073201944699\n",
      "85  Episode Reward:  -83.57214793750003 .Epsilon: 0.6498078359349755\n",
      "86  Episode Reward:  -130.95696980525167 .Epsilon: 0.6465587967553006\n",
      "87  Episode Reward:  10.975123421289666 .Epsilon: 0.6433260027715241\n",
      "88  Episode Reward:  -86.98917086041573 .Epsilon: 0.6401093727576664\n",
      "89  Episode Reward:  -58.816251174919245 .Epsilon: 0.6369088258938781\n",
      "90  Episode Reward:  -287.26620339383896 .Epsilon: 0.6337242817644086\n",
      "91  Episode Reward:  -217.77150727232015 .Epsilon: 0.6305556603555866\n",
      "92  Episode Reward:  -106.47523999549374 .Epsilon: 0.6274028820538087\n",
      "93  Episode Reward:  -148.71026136980305 .Epsilon: 0.6242658676435396\n",
      "94  Episode Reward:  -270.09923669353833 .Epsilon: 0.6211445383053219\n",
      "STATIC WEIGHTS UPDATED\n",
      "95  Episode Reward:  -240.9334458557036 .Epsilon: 0.6180388156137953\n",
      "96  Episode Reward:  44.31422741785471 .Epsilon: 0.6149486215357263\n",
      "97  Episode Reward:  -50.911381946063955 .Epsilon: 0.6118738784280476\n",
      "98  Episode Reward:  -86.70591366789802 .Epsilon: 0.6088145090359074\n",
      "99  Episode Reward:  8.823706246735952 .Epsilon: 0.6057704364907278\n",
      "100  Episode Reward:  -231.52211067275775 .Epsilon: 0.6027415843082742\n",
      "101  Episode Reward:  -327.5006381046454 .Epsilon: 0.5997278763867329\n",
      "102  Episode Reward:  -286.3795866431086 .Epsilon: 0.5967292370047992\n",
      "103  Episode Reward:  -61.35058898754446 .Epsilon: 0.5937455908197752\n",
      "104  Episode Reward:  -89.47274077936297 .Epsilon: 0.5907768628656763\n",
      "105  Episode Reward:  -213.61231615503408 .Epsilon: 0.5878229785513479\n",
      "106  Episode Reward:  -276.4575682504849 .Epsilon: 0.5848838636585911\n",
      "107  Episode Reward:  -15.657102811600254 .Epsilon: 0.5819594443402982\n",
      "108  Episode Reward:  -151.80211601063252 .Epsilon: 0.5790496471185967\n",
      "109  Episode Reward:  -171.95340473468173 .Epsilon: 0.5761543988830038\n",
      "110  Episode Reward:  -72.9715454340341 .Epsilon: 0.5732736268885887\n",
      "111  Episode Reward:  -67.91908408519342 .Epsilon: 0.5704072587541458\n",
      "112  Episode Reward:  -296.7238824910954 .Epsilon: 0.567555222460375\n",
      "STATIC WEIGHTS UPDATED\n",
      "113  Episode Reward:  -118.07622946433268 .Epsilon: 0.5647174463480732\n",
      "114  Episode Reward:  -195.29777179304432 .Epsilon: 0.5618938591163328\n",
      "115  Episode Reward:  -68.4962792864344 .Epsilon: 0.5590843898207511\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "116  Episode Reward:  -285.5348117689465 .Epsilon: 0.5562889678716474\n",
      "117  Episode Reward:  -177.27503616397703 .Epsilon: 0.5535075230322891\n",
      "118  Episode Reward:  18.71257058921165 .Epsilon: 0.5507399854171277\n",
      "119  Episode Reward:  -128.56100826503018 .Epsilon: 0.547986285490042\n",
      "120  Episode Reward:  -242.1141848818013 .Epsilon: 0.5452463540625918\n",
      "121  Episode Reward:  -99.30872372371662 .Epsilon: 0.5425201222922789\n",
      "122  Episode Reward:  -137.20796934332594 .Epsilon: 0.5398075216808175\n",
      "123  Episode Reward:  -54.939166862473485 .Epsilon: 0.5371084840724134\n",
      "124  Episode Reward:  -180.43408866972334 .Epsilon: 0.5344229416520513\n",
      "125  Episode Reward:  -111.47785776323435 .Epsilon: 0.531750826943791\n",
      "126  Episode Reward:  -252.00305759571256 .Epsilon: 0.5290920728090721\n",
      "STATIC WEIGHTS UPDATED\n",
      "127  Episode Reward:  -218.46072154390419 .Epsilon: 0.5264466124450268\n",
      "128  Episode Reward:  154.4974554561002 .Epsilon: 0.5238143793828016\n",
      "129  Episode Reward:  -94.10962377898358 .Epsilon: 0.5211953074858876\n",
      "130  Episode Reward:  -286.9175253454859 .Epsilon: 0.5185893309484582\n",
      "131  Episode Reward:  -187.9850568676173 .Epsilon: 0.5159963842937159\n",
      "132  Episode Reward:  -237.5480086923061 .Epsilon: 0.5134164023722473\n",
      "133  Episode Reward:  -29.906887794340435 .Epsilon: 0.510849320360386\n",
      "STATIC WEIGHTS UPDATED\n",
      "134  Episode Reward:  121.4883197565215 .Epsilon: 0.5082950737585841\n",
      "135  Episode Reward:  -85.24286041292001 .Epsilon: 0.5057535983897912\n",
      "136  Episode Reward:  -62.805004840109675 .Epsilon: 0.5032248303978422\n",
      "137  Episode Reward:  -223.67790613711261 .Epsilon: 0.500708706245853\n",
      "138  Episode Reward:  -223.29580901992557 .Epsilon: 0.4982051627146237\n",
      "139  Episode Reward:  -136.32437299586255 .Epsilon: 0.49571413690105054\n",
      "140  Episode Reward:  -263.1899058499711 .Epsilon: 0.4932355662165453\n",
      "141  Episode Reward:  -269.87520535212167 .Epsilon: 0.4907693883854626\n",
      "142  Episode Reward:  -92.05071285542725 .Epsilon: 0.4883155414435353\n",
      "143  Episode Reward:  -105.97495030132804 .Epsilon: 0.4858739637363176\n",
      "144  Episode Reward:  -248.68385536357204 .Epsilon: 0.483444593917636\n",
      "145  Episode Reward:  -82.30827320460172 .Epsilon: 0.4810273709480478\n",
      "146  Episode Reward:  -35.41771961271918 .Epsilon: 0.47862223409330756\n",
      "STATIC WEIGHTS UPDATED\n",
      "147  Episode Reward:  20.09680094009329 .Epsilon: 0.47622912292284103\n",
      "148  Episode Reward:  -270.87981980320444 .Epsilon: 0.4738479773082268\n",
      "149  Episode Reward:  7.3892759533392365 .Epsilon: 0.47147873742168567\n",
      "150  Episode Reward:  -391.40017376714735 .Epsilon: 0.46912134373457726\n",
      "151  Episode Reward:  -221.5767364877756 .Epsilon: 0.46677573701590436\n",
      "152  Episode Reward:  27.241772119986923 .Epsilon: 0.46444185833082485\n",
      "STATIC WEIGHTS UPDATED\n",
      "153  Episode Reward:  -229.22959762865705 .Epsilon: 0.46211964903917074\n",
      "154  Episode Reward:  -146.0284850199148 .Epsilon: 0.4598090507939749\n",
      "155  Episode Reward:  -275.2325244490201 .Epsilon: 0.457510005540005\n",
      "156  Episode Reward:  -199.94555204065517 .Epsilon: 0.45522245551230495\n",
      "STATIC WEIGHTS UPDATED\n",
      "157  Episode Reward:  30.65534015967789 .Epsilon: 0.4529463432347434\n",
      "158  Episode Reward:  -152.91172097324753 .Epsilon: 0.4506816115185697\n",
      "159  Episode Reward:  -13.975664082567484 .Epsilon: 0.4484282034609769\n",
      "STATIC WEIGHTS UPDATED\n",
      "160  Episode Reward:  -50.63411642529157 .Epsilon: 0.446186062443672\n",
      "161  Episode Reward:  20.701551622043702 .Epsilon: 0.4439551321314536\n",
      "162  Episode Reward:  -21.345545851729064 .Epsilon: 0.4417353564707963\n",
      "STATIC WEIGHTS UPDATED\n",
      "163  Episode Reward:  -39.289372767535625 .Epsilon: 0.43952667968844233\n",
      "164  Episode Reward:  -23.306988336579014 .Epsilon: 0.43732904629000013\n",
      "165  Episode Reward:  -59.51326762908621 .Epsilon: 0.4351424010585501\n",
      "STATIC WEIGHTS UPDATED\n",
      "166  Episode Reward:  -44.24715096278166 .Epsilon: 0.43296668905325736\n",
      "167  Episode Reward:  -72.65777103909774 .Epsilon: 0.43080185560799106\n",
      "168  Episode Reward:  -93.45960834849251 .Epsilon: 0.4286478463299511\n",
      "STATIC WEIGHTS UPDATED\n",
      "169  Episode Reward:  -107.46934797885247 .Epsilon: 0.42650460709830135\n",
      "170  Episode Reward:  -8.876844245133944 .Epsilon: 0.42437208406280985\n",
      "171  Episode Reward:  -41.630067044505964 .Epsilon: 0.4222502236424958\n",
      "STATIC WEIGHTS UPDATED\n",
      "172  Episode Reward:  -47.99160654319961 .Epsilon: 0.42013897252428334\n",
      "173  Episode Reward:  7.555127794466693 .Epsilon: 0.4180382776616619\n",
      "174  Episode Reward:  -50.005280261145295 .Epsilon: 0.4159480862733536\n",
      "STATIC WEIGHTS UPDATED\n",
      "175  Episode Reward:  -55.487618487746246 .Epsilon: 0.41386834584198684\n",
      "176  Episode Reward:  -128.12971640615257 .Epsilon: 0.4117990041127769\n",
      "177  Episode Reward:  -160.56887461589108 .Epsilon: 0.40974000909221303\n",
      "178  Episode Reward:  -150.56357297637553 .Epsilon: 0.40769130904675194\n",
      "STATIC WEIGHTS UPDATED\n",
      "179  Episode Reward:  -73.08911243679667 .Epsilon: 0.40565285250151817\n",
      "180  Episode Reward:  -42.41846874589787 .Epsilon: 0.4036245882390106\n",
      "181  Episode Reward:  -23.40233247495537 .Epsilon: 0.4016064652978155\n",
      "STATIC WEIGHTS UPDATED\n",
      "182  Episode Reward:  62.250819252301234 .Epsilon: 0.3995984329713264\n",
      "183  Episode Reward:  -199.66629168147733 .Epsilon: 0.3976004408064698\n",
      "184  Episode Reward:  -49.89640997006194 .Epsilon: 0.39561243860243744\n",
      "185  Episode Reward:  -42.522534248365034 .Epsilon: 0.3936343764094253\n",
      "STATIC WEIGHTS UPDATED\n",
      "186  Episode Reward:  -51.327882950516056 .Epsilon: 0.39166620452737816\n",
      "187  Episode Reward:  -222.61589965359283 .Epsilon: 0.3897078735047413\n",
      "188  Episode Reward:  -48.84850571722832 .Epsilon: 0.3877593341372176\n",
      "STATIC WEIGHTS UPDATED\n",
      "189  Episode Reward:  -173.40690601856937 .Epsilon: 0.3858205374665315\n",
      "190  Episode Reward:  -51.71988159182223 .Epsilon: 0.38389143477919885\n",
      "191  Episode Reward:  -13.586012889676635 .Epsilon: 0.3819719776053028\n",
      "STATIC WEIGHTS UPDATED\n",
      "192  Episode Reward:  -45.28338344323548 .Epsilon: 0.3800621177172763\n",
      "193  Episode Reward:  -300.55292031139203 .Epsilon: 0.37816180712868996\n",
      "194  Episode Reward:  -135.64546907298183 .Epsilon: 0.37627099809304654\n",
      "195  Episode Reward:  -199.58616175023667 .Epsilon: 0.3743896431025813\n",
      "196  Episode Reward:  -206.84110302495847 .Epsilon: 0.37251769488706843\n",
      "197  Episode Reward:  -220.44878545580417 .Epsilon: 0.3706551064126331\n",
      "STATIC WEIGHTS UPDATED\n",
      "198  Episode Reward:  -73.80029034737686 .Epsilon: 0.36880183088056995\n",
      "199  Episode Reward:  -264.2739215100321 .Epsilon: 0.3669578217261671\n",
      "200  Episode Reward:  -218.38671283357655 .Epsilon: 0.36512303261753626\n",
      "201  Episode Reward:  -131.4856070803432 .Epsilon: 0.3632974174544486\n",
      "202  Episode Reward:  -21.832459827856184 .Epsilon: 0.3614809303671764\n",
      "STATIC WEIGHTS UPDATED\n",
      "203  Episode Reward:  -361.94290968611864 .Epsilon: 0.3596735257153405\n",
      "204  Episode Reward:  -126.16056677250359 .Epsilon: 0.3578751580867638\n",
      "205  Episode Reward:  -199.22525960584045 .Epsilon: 0.35608578229633\n",
      "206  Episode Reward:  -226.7569556915567 .Epsilon: 0.3543053533848483\n",
      "207  Episode Reward:  -125.733661347656 .Epsilon: 0.35253382661792404\n",
      "208  Episode Reward:  -176.28114192019615 .Epsilon: 0.3507711574848344\n",
      "209  Episode Reward:  -364.7078662696705 .Epsilon: 0.34901730169741024\n",
      "STATIC WEIGHTS UPDATED\n",
      "210  Episode Reward:  52.98758039664425 .Epsilon: 0.3472722151889232\n",
      "211  Episode Reward:  -24.829565187881133 .Epsilon: 0.3455358541129786\n",
      "212  Episode Reward:  -283.49260738985333 .Epsilon: 0.3438081748424137\n",
      "213  Episode Reward:  -124.31236056379677 .Epsilon: 0.3420891339682016\n",
      "214  Episode Reward:  -60.230637005067955 .Epsilon: 0.3403786882983606\n",
      "STATIC WEIGHTS UPDATED\n",
      "215  Episode Reward:  32.997310157062564 .Epsilon: 0.3386767948568688\n",
      "216  Episode Reward:  -73.8920909289153 .Epsilon: 0.33698341088258443\n",
      "217  Episode Reward:  -100.67051415722966 .Epsilon: 0.3352984938281715\n",
      "STATIC WEIGHTS UPDATED\n",
      "218  Episode Reward:  -158.41869705177785 .Epsilon: 0.33362200135903064\n",
      "219  Episode Reward:  -113.19386830363813 .Epsilon: 0.33195389135223546\n",
      "220  Episode Reward:  42.91500967670372 .Epsilon: 0.3302941218954743\n",
      "221  Episode Reward:  -80.13292639333616 .Epsilon: 0.32864265128599696\n",
      "222  Episode Reward:  -18.96912642078769 .Epsilon: 0.326999438029567\n",
      "STATIC WEIGHTS UPDATED\n",
      "223  Episode Reward:  -107.52403749351947 .Epsilon: 0.3253644408394192\n",
      "224  Episode Reward:  99.30707503705389 .Epsilon: 0.3237376186352221\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "225  Episode Reward:  -36.452373386302774 .Epsilon: 0.322118930542046\n",
      "226  Episode Reward:  -20.773861218455167 .Epsilon: 0.32050833588933575\n",
      "STATIC WEIGHTS UPDATED\n",
      "227  Episode Reward:  -86.66863334684936 .Epsilon: 0.31890579420988907\n",
      "228  Episode Reward:  -34.217850758979594 .Epsilon: 0.3173112652388396\n",
      "229  Episode Reward:  179.29113417430457 .Epsilon: 0.3157247089126454\n",
      "STATIC WEIGHTS UPDATED\n",
      "230  Episode Reward:  -218.4304918135378 .Epsilon: 0.3141460853680822\n",
      "231  Episode Reward:  -97.67125777174276 .Epsilon: 0.3125753549412418\n",
      "232  Episode Reward:  -80.26333771080536 .Epsilon: 0.31101247816653554\n",
      "233  Episode Reward:  -165.04732126896914 .Epsilon: 0.30945741577570285\n",
      "234  Episode Reward:  -212.8489882847768 .Epsilon: 0.3079101286968243\n",
      "235  Episode Reward:  -22.214462922350847 .Epsilon: 0.3063705780533402\n",
      "236  Episode Reward:  93.7201808708315 .Epsilon: 0.30483872516307353\n",
      "237  Episode Reward:  -150.7237992174229 .Epsilon: 0.3033145315372582\n",
      "STATIC WEIGHTS UPDATED\n",
      "238  Episode Reward:  -158.84755176494787 .Epsilon: 0.3017979588795719\n",
      "239  Episode Reward:  -119.54357053173608 .Epsilon: 0.30028896908517405\n",
      "240  Episode Reward:  -15.267598886362464 .Epsilon: 0.2987875242397482\n",
      "241  Episode Reward:  -119.54833343735085 .Epsilon: 0.29729358661854943\n",
      "242  Episode Reward:  -56.701330963507026 .Epsilon: 0.29580711868545667\n",
      "STATIC WEIGHTS UPDATED\n",
      "243  Episode Reward:  -55.73207453361968 .Epsilon: 0.2943280830920294\n",
      "STATIC WEIGHTS UPDATED\n",
      "244  Episode Reward:  -14.905552470352134 .Epsilon: 0.29285644267656924\n",
      "245  Episode Reward:  -66.43353431123734 .Epsilon: 0.2913921604631864\n",
      "246  Episode Reward:  -85.99062124357977 .Epsilon: 0.28993519966087045\n",
      "247  Episode Reward:  -125.0826715760553 .Epsilon: 0.2884855236625661\n",
      "248  Episode Reward:  -215.74584178561278 .Epsilon: 0.28704309604425327\n",
      "STATIC WEIGHTS UPDATED\n",
      "249  Episode Reward:  -19.422240069502198 .Epsilon: 0.285607880564032\n",
      "250  Episode Reward:  -271.6753529709321 .Epsilon: 0.28417984116121187\n",
      "251  Episode Reward:  -93.33344229661556 .Epsilon: 0.2827589419554058\n",
      "252  Episode Reward:  -122.33501596454633 .Epsilon: 0.28134514724562876\n",
      "253  Episode Reward:  -115.23938977711514 .Epsilon: 0.2799384215094006\n",
      "STATIC WEIGHTS UPDATED\n",
      "254  Episode Reward:  -69.85930462287384 .Epsilon: 0.27853872940185365\n",
      "255  Episode Reward:  -87.75750496149274 .Epsilon: 0.27714603575484437\n",
      "256  Episode Reward:  -24.39207257237547 .Epsilon: 0.2757603055760701\n",
      "257  Episode Reward:  3.7567382605374746 .Epsilon: 0.2743815040481898\n",
      "258  Episode Reward:  -6.984449380445483 .Epsilon: 0.2730095965279488\n",
      "259  Episode Reward:  1.212705361866412 .Epsilon: 0.27164454854530906\n",
      "260  Episode Reward:  -8.921143849893838 .Epsilon: 0.2702863258025825\n",
      "STATIC WEIGHTS UPDATED\n",
      "261  Episode Reward:  -60.19472796082227 .Epsilon: 0.2689348941735696\n",
      "262  Episode Reward:  -31.75610393851167 .Epsilon: 0.26759021970270175\n",
      "263  Episode Reward:  -54.93573366412008 .Epsilon: 0.2662522686041882\n",
      "264  Episode Reward:  -105.52482406217217 .Epsilon: 0.2649210072611673\n",
      "STATIC WEIGHTS UPDATED\n",
      "265  Episode Reward:  -117.90067003011947 .Epsilon: 0.26359640222486147\n",
      "266  Episode Reward:  -35.8936645777014 .Epsilon: 0.26227842021373715\n",
      "267  Episode Reward:  -71.72857085179965 .Epsilon: 0.2609670281126685\n",
      "268  Episode Reward:  -26.713939687258247 .Epsilon: 0.25966219297210513\n",
      "269  Episode Reward:  -78.20120890539596 .Epsilon: 0.2583638820072446\n",
      "270  Episode Reward:  -17.52340122037066 .Epsilon: 0.2570720625972084\n",
      "271  Episode Reward:  -191.58526790739052 .Epsilon: 0.25578670228422234\n",
      "STATIC WEIGHTS UPDATED\n",
      "272  Episode Reward:  -102.85077327661972 .Epsilon: 0.25450776877280124\n",
      "273  Episode Reward:  -326.221485726453 .Epsilon: 0.2532352299289372\n",
      "274  Episode Reward:  -47.09768590881296 .Epsilon: 0.2519690537792925\n",
      "275  Episode Reward:  6.335078195101346 .Epsilon: 0.2507092085103961\n",
      "276  Episode Reward:  132.39230041722664 .Epsilon: 0.2494556624678441\n",
      "277  Episode Reward:  -126.98260899861394 .Epsilon: 0.24820838415550486\n",
      "STATIC WEIGHTS UPDATED\n",
      "278  Episode Reward:  -116.6461533698487 .Epsilon: 0.24696734223472733\n",
      "279  Episode Reward:  -20.42178223061717 .Epsilon: 0.2457325055235537\n",
      "280  Episode Reward:  -117.20384153997745 .Epsilon: 0.24450384299593592\n",
      "281  Episode Reward:  -117.21740933829574 .Epsilon: 0.24328132378095624\n",
      "282  Episode Reward:  -130.61333786554144 .Epsilon: 0.24206491716205145\n",
      "283  Episode Reward:  -49.390811908473445 .Epsilon: 0.2408545925762412\n",
      "284  Episode Reward:  -32.93330905708157 .Epsilon: 0.23965031961336\n",
      "285  Episode Reward:  -157.79875681904798 .Epsilon: 0.2384520680152932\n",
      "286  Episode Reward:  -208.1999415582176 .Epsilon: 0.23725980767521673\n",
      "STATIC WEIGHTS UPDATED\n",
      "287  Episode Reward:  -171.8007850285019 .Epsilon: 0.23607350863684065\n",
      "288  Episode Reward:  -130.30931597167847 .Epsilon: 0.23489314109365644\n",
      "289  Episode Reward:  -199.1731806551261 .Epsilon: 0.23371867538818816\n",
      "290  Episode Reward:  -80.49392587805148 .Epsilon: 0.23255008201124722\n",
      "291  Episode Reward:  -203.4275338729811 .Epsilon: 0.231387331601191\n",
      "292  Episode Reward:  -6.377047837438539 .Epsilon: 0.23023039494318503\n",
      "STATIC WEIGHTS UPDATED\n",
      "293  Episode Reward:  78.56449330431322 .Epsilon: 0.2290792429684691\n",
      "294  Episode Reward:  -104.19368831084995 .Epsilon: 0.22793384675362674\n",
      "295  Episode Reward:  -114.0297531550897 .Epsilon: 0.22679417751985861\n",
      "296  Episode Reward:  -94.57384349638848 .Epsilon: 0.22566020663225933\n",
      "297  Episode Reward:  -86.74213555909064 .Epsilon: 0.22453190559909803\n",
      "298  Episode Reward:  -43.26952049999754 .Epsilon: 0.22340924607110255\n",
      "299  Episode Reward:  -76.32072474924422 .Epsilon: 0.22229219984074702\n",
      "300  Episode Reward:  -157.93189090481997 .Epsilon: 0.2211807388415433\n",
      "301  Episode Reward:  -53.85918968303475 .Epsilon: 0.22007483514733558\n",
      "STATIC WEIGHTS UPDATED\n",
      "302  Episode Reward:  -225.23157987420302 .Epsilon: 0.2189744609715989\n",
      "303  Episode Reward:  109.78827269378135 .Epsilon: 0.2178795886667409\n",
      "304  Episode Reward:  -115.2303968334211 .Epsilon: 0.2167901907234072\n",
      "305  Episode Reward:  -121.20667813341949 .Epsilon: 0.21570623976979014\n",
      "306  Episode Reward:  -109.37547531724691 .Epsilon: 0.21462770857094118\n",
      "307  Episode Reward:  -143.83833716437113 .Epsilon: 0.21355457002808648\n",
      "308  Episode Reward:  -70.72836478201084 .Epsilon: 0.21248679717794605\n",
      "309  Episode Reward:  -167.14570340933562 .Epsilon: 0.21142436319205632\n",
      "310  Episode Reward:  -77.21976969688262 .Epsilon: 0.21036724137609603\n",
      "311  Episode Reward:  -164.76635581175157 .Epsilon: 0.20931540516921554\n",
      "312  Episode Reward:  -383.03650024160027 .Epsilon: 0.20826882814336947\n",
      "STATIC WEIGHTS UPDATED\n",
      "313  Episode Reward:  -114.05631625676898 .Epsilon: 0.20722748400265262\n",
      "314  Episode Reward:  1.4434182263489959 .Epsilon: 0.20619134658263935\n",
      "315  Episode Reward:  -85.8655957350017 .Epsilon: 0.20516038984972615\n",
      "316  Episode Reward:  -74.50786817965421 .Epsilon: 0.2041345879004775\n",
      "317  Episode Reward:  -56.904528841194484 .Epsilon: 0.2031139149609751\n",
      "318  Episode Reward:  -141.6027719536715 .Epsilon: 0.20209834538617025\n",
      "319  Episode Reward:  -130.85837038242954 .Epsilon: 0.2010878536592394\n",
      "320  Episode Reward:  -378.7877422790986 .Epsilon: 0.2000824143909432\n",
      "STATIC WEIGHTS UPDATED\n",
      "321  Episode Reward:  104.89162588635492 .Epsilon: 0.19908200231898848\n",
      "322  Episode Reward:  -107.96037253453112 .Epsilon: 0.19808659230739353\n",
      "323  Episode Reward:  23.87490574805426 .Epsilon: 0.19709615934585656\n",
      "324  Episode Reward:  -85.86156024409658 .Epsilon: 0.19611067854912728\n",
      "325  Episode Reward:  182.66928917376515 .Epsilon: 0.19513012515638165\n",
      "STATIC WEIGHTS UPDATED\n",
      "326  Episode Reward:  228.71442787570334 .Epsilon: 0.19415447453059972\n",
      "327  Episode Reward:  -151.53655069104138 .Epsilon: 0.19318370215794672\n",
      "328  Episode Reward:  -147.57076642182028 .Epsilon: 0.192217783647157\n",
      "329  Episode Reward:  -84.82265792019618 .Epsilon: 0.1912566947289212\n",
      "330  Episode Reward:  -111.75042418456808 .Epsilon: 0.1903004112552766\n",
      "331  Episode Reward:  -6.22734004029158 .Epsilon: 0.18934890919900021\n",
      "332  Episode Reward:  -40.428860767910955 .Epsilon: 0.18840216465300522\n",
      "STATIC WEIGHTS UPDATED\n",
      "333  Episode Reward:  222.3945408948528 .Epsilon: 0.18746015382974018\n",
      "334  Episode Reward:  107.33343417881234 .Epsilon: 0.1865228530605915\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "335  Episode Reward:  -395.62010805443805 .Epsilon: 0.18559023879528855\n",
      "336  Episode Reward:  -156.1179480017826 .Epsilon: 0.1846622876013121\n",
      "337  Episode Reward:  -46.58373287371969 .Epsilon: 0.18373897616330553\n",
      "STATIC WEIGHTS UPDATED\n",
      "338  Episode Reward:  -3.822359025795265 .Epsilon: 0.182820281282489\n",
      "339  Episode Reward:  -38.16660430832654 .Epsilon: 0.18190617987607657\n",
      "340  Episode Reward:  -28.271950921118503 .Epsilon: 0.18099664897669618\n",
      "341  Episode Reward:  233.36467429765045 .Epsilon: 0.1800916657318127\n",
      "342  Episode Reward:  233.46574337419958 .Epsilon: 0.17919120740315364\n",
      "343  Episode Reward:  -129.13200218874795 .Epsilon: 0.17829525136613786\n",
      "344  Episode Reward:  -94.93327492018348 .Epsilon: 0.17740377510930716\n",
      "STATIC WEIGHTS UPDATED\n",
      "345  Episode Reward:  231.2478863443314 .Epsilon: 0.17651675623376062\n",
      "346  Episode Reward:  -104.54851190797027 .Epsilon: 0.1756341724525918\n",
      "347  Episode Reward:  -99.14193070206741 .Epsilon: 0.17475600159032884\n",
      "348  Episode Reward:  -98.91004187542285 .Epsilon: 0.17388222158237718\n",
      "STATIC WEIGHTS UPDATED\n",
      "349  Episode Reward:  -72.85687934085723 .Epsilon: 0.1730128104744653\n",
      "350  Episode Reward:  -108.109332535519 .Epsilon: 0.17214774642209296\n",
      "351  Episode Reward:  -311.5614807637561 .Epsilon: 0.1712870076899825\n",
      "352  Episode Reward:  -102.70257761882077 .Epsilon: 0.17043057265153258\n",
      "353  Episode Reward:  -76.42394952728112 .Epsilon: 0.16957841978827493\n",
      "354  Episode Reward:  -68.5603531840572 .Epsilon: 0.16873052768933355\n",
      "355  Episode Reward:  18.952367900293638 .Epsilon: 0.1678868750508869\n",
      "356  Episode Reward:  273.6026642240687 .Epsilon: 0.16704744067563246\n",
      "357  Episode Reward:  40.09040591480763 .Epsilon: 0.1662122034722543\n",
      "358  Episode Reward:  -37.353685435436645 .Epsilon: 0.16538114245489302\n",
      "359  Episode Reward:  -73.5516850445089 .Epsilon: 0.16455423674261854\n",
      "360  Episode Reward:  -2.4749597088365363 .Epsilon: 0.16373146555890544\n",
      "361  Episode Reward:  -83.0688158227332 .Epsilon: 0.16291280823111093\n",
      "362  Episode Reward:  -61.06821522956512 .Epsilon: 0.16209824418995536\n",
      "363  Episode Reward:  -90.141673660223 .Epsilon: 0.16128775296900558\n",
      "STATIC WEIGHTS UPDATED\n",
      "364  Episode Reward:  -82.81902912033082 .Epsilon: 0.16048131420416054\n",
      "365  Episode Reward:  -43.11587129637064 .Epsilon: 0.15967890763313974\n",
      "366  Episode Reward:  -43.60357291849153 .Epsilon: 0.15888051309497406\n",
      "367  Episode Reward:  231.7810516819149 .Epsilon: 0.1580861105294992\n",
      "368  Episode Reward:  249.11879717892128 .Epsilon: 0.1572956799768517\n",
      "369  Episode Reward:  -22.258809056228586 .Epsilon: 0.15650920157696743\n",
      "370  Episode Reward:  -73.95121521899524 .Epsilon: 0.1557266555690826\n",
      "371  Episode Reward:  -205.32258407750626 .Epsilon: 0.1549480222912372\n",
      "372  Episode Reward:  263.2846665973316 .Epsilon: 0.15417328217978102\n",
      "373  Episode Reward:  -47.485736405498415 .Epsilon: 0.1534024157688821\n",
      "374  Episode Reward:  -191.68877831891552 .Epsilon: 0.1526354036900377\n",
      "375  Episode Reward:  252.05484464048064 .Epsilon: 0.1518722266715875\n",
      "STATIC WEIGHTS UPDATED\n",
      "376  Episode Reward:  249.9116098423467 .Epsilon: 0.15111286553822956\n",
      "377  Episode Reward:  -54.16591464715976 .Epsilon: 0.15035730121053842\n",
      "378  Episode Reward:  -141.9327003260694 .Epsilon: 0.14960551470448571\n",
      "379  Episode Reward:  -57.560065295226686 .Epsilon: 0.14885748713096328\n",
      "380  Episode Reward:  -28.213809420416027 .Epsilon: 0.14811319969530845\n",
      "381  Episode Reward:  -95.92921583481649 .Epsilon: 0.1473726336968319\n",
      "382  Episode Reward:  -312.75860791767775 .Epsilon: 0.14663577052834775\n",
      "383  Episode Reward:  -21.332267306128387 .Epsilon: 0.14590259167570602\n",
      "384  Episode Reward:  59.43888036790722 .Epsilon: 0.1451730787173275\n",
      "385  Episode Reward:  -49.563318993154816 .Epsilon: 0.14444721332374086\n",
      "386  Episode Reward:  -29.766194936458618 .Epsilon: 0.14372497725712216\n",
      "387  Episode Reward:  50.763903544278065 .Epsilon: 0.14300635237083656\n",
      "STATIC WEIGHTS UPDATED\n",
      "388  Episode Reward:  -284.0474963957686 .Epsilon: 0.14229132060898236\n",
      "389  Episode Reward:  222.78105960564622 .Epsilon: 0.14157986400593744\n",
      "390  Episode Reward:  -298.78752432121826 .Epsilon: 0.14087196468590776\n",
      "391  Episode Reward:  -202.5223903485742 .Epsilon: 0.14016760486247823\n",
      "392  Episode Reward:  -147.33598851904182 .Epsilon: 0.13946676683816583\n",
      "393  Episode Reward:  -71.4682294019595 .Epsilon: 0.138769433003975\n",
      "394  Episode Reward:  -548.2447473739378 .Epsilon: 0.13807558583895513\n",
      "395  Episode Reward:  -36.178843262848275 .Epsilon: 0.13738520790976036\n",
      "396  Episode Reward:  0.6687632755829043 .Epsilon: 0.13669828187021155\n",
      "397  Episode Reward:  16.915870345494227 .Epsilon: 0.13601479046086049\n",
      "STATIC WEIGHTS UPDATED\n",
      "398  Episode Reward:  -23.616588840970195 .Epsilon: 0.1353347165085562\n",
      "399  Episode Reward:  247.0394850364896 .Epsilon: 0.1346580429260134\n",
      "400  Episode Reward:  -33.96484261958551 .Epsilon: 0.13398475271138335\n",
      "401  Episode Reward:  -69.0781838211318 .Epsilon: 0.13331482894782642\n",
      "402  Episode Reward:  -197.40055151584227 .Epsilon: 0.13264825480308728\n",
      "403  Episode Reward:  -30.79587464860333 .Epsilon: 0.13198501352907185\n",
      "404  Episode Reward:  -32.806887876980284 .Epsilon: 0.1313250884614265\n",
      "405  Episode Reward:  -173.61938752141242 .Epsilon: 0.13066846301911936\n",
      "406  Episode Reward:  -75.71175404268794 .Epsilon: 0.13001512070402377\n",
      "407  Episode Reward:  8.933268329921447 .Epsilon: 0.12936504510050365\n",
      "408  Episode Reward:  -69.94267815959016 .Epsilon: 0.12871821987500112\n",
      "409  Episode Reward:  -141.338031887713 .Epsilon: 0.12807462877562611\n",
      "410  Episode Reward:  -114.58613158474444 .Epsilon: 0.12743425563174798\n",
      "411  Episode Reward:  -54.94461504822858 .Epsilon: 0.12679708435358925\n",
      "412  Episode Reward:  -40.276005898023946 .Epsilon: 0.1261630989318213\n",
      "413  Episode Reward:  -9.031020629173554 .Epsilon: 0.1255322834371622\n",
      "414  Episode Reward:  -257.5058904519346 .Epsilon: 0.12490462201997637\n",
      "STATIC WEIGHTS UPDATED\n",
      "415  Episode Reward:  -168.93038220745962 .Epsilon: 0.1242800989098765\n",
      "416  Episode Reward:  -62.38240272989135 .Epsilon: 0.12365869841532712\n",
      "417  Episode Reward:  -34.15210840709942 .Epsilon: 0.12304040492325048\n",
      "418  Episode Reward:  -4.504320883113579 .Epsilon: 0.12242520289863423\n",
      "419  Episode Reward:  -48.39928005257603 .Epsilon: 0.12181307688414106\n",
      "420  Episode Reward:  -62.03088348533279 .Epsilon: 0.12120401149972035\n",
      "421  Episode Reward:  210.79450315960403 .Epsilon: 0.12059799144222175\n",
      "422  Episode Reward:  160.79350610319216 .Epsilon: 0.11999500148501063\n",
      "423  Episode Reward:  -19.548711059079267 .Epsilon: 0.11939502647758558\n",
      "424  Episode Reward:  -43.28886816951346 .Epsilon: 0.11879805134519765\n",
      "425  Episode Reward:  -81.48263175112999 .Epsilon: 0.11820406108847166\n",
      "426  Episode Reward:  -351.18536242069763 .Epsilon: 0.1176130407830293\n",
      "STATIC WEIGHTS UPDATED\n",
      "427  Episode Reward:  -394.62971653587095 .Epsilon: 0.11702497557911415\n",
      "428  Episode Reward:  -220.14797929633463 .Epsilon: 0.11643985070121858\n",
      "429  Episode Reward:  244.10177886111492 .Epsilon: 0.11585765144771248\n",
      "430  Episode Reward:  -97.86864955023294 .Epsilon: 0.11527836319047392\n",
      "431  Episode Reward:  -18.335867297970992 .Epsilon: 0.11470197137452155\n",
      "432  Episode Reward:  -212.7463854869557 .Epsilon: 0.11412846151764894\n",
      "433  Episode Reward:  -32.575009436874325 .Epsilon: 0.1135578192100607\n",
      "434  Episode Reward:  229.25302713958803 .Epsilon: 0.11299003011401039\n",
      "435  Episode Reward:  -433.2276382417929 .Epsilon: 0.11242507996344034\n",
      "436  Episode Reward:  -216.8113328861801 .Epsilon: 0.11186295456362313\n",
      "437  Episode Reward:  -81.6003548160572 .Epsilon: 0.11130363979080501\n",
      "438  Episode Reward:  -51.61391860577531 .Epsilon: 0.11074712159185099\n",
      "439  Episode Reward:  -267.73481054978697 .Epsilon: 0.11019338598389174\n",
      "STATIC WEIGHTS UPDATED\n",
      "440  Episode Reward:  23.21392993406809 .Epsilon: 0.10964241905397228\n",
      "441  Episode Reward:  -335.3160585459506 .Epsilon: 0.10909420695870241\n",
      "442  Episode Reward:  -270.88049587631224 .Epsilon: 0.1085487359239089\n",
      "443  Episode Reward:  257.9320932032652 .Epsilon: 0.10800599224428936\n",
      "444  Episode Reward:  206.37413236587778 .Epsilon: 0.10746596228306791\n",
      "445  Episode Reward:  82.28036746594026 .Epsilon: 0.10692863247165257\n",
      "446  Episode Reward:  -49.86867047432126 .Epsilon: 0.1063939893092943\n",
      "447  Episode Reward:  -52.893210471885794 .Epsilon: 0.10586201936274783\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "448  Episode Reward:  -335.97683809106354 .Epsilon: 0.10533270926593409\n",
      "449  Episode Reward:  -66.07767724201851 .Epsilon: 0.10480604571960442\n",
      "450  Episode Reward:  12.600412507384817 .Epsilon: 0.1042820154910064\n",
      "STATIC WEIGHTS UPDATED\n",
      "451  Episode Reward:  172.88432325342782 .Epsilon: 0.10376060541355137\n",
      "452  Episode Reward:  71.5435538300623 .Epsilon: 0.1032418023864836\n",
      "453  Episode Reward:  217.16681652084887 .Epsilon: 0.10272559337455119\n",
      "454  Episode Reward:  234.54335113379054 .Epsilon: 0.10221196540767843\n",
      "455  Episode Reward:  228.49083568265658 .Epsilon: 0.10170090558064004\n",
      "456  Episode Reward:  -11.281745423731849 .Epsilon: 0.10119240105273684\n",
      "STATIC WEIGHTS UPDATED\n",
      "457  Episode Reward:  -289.96157167424036 .Epsilon: 0.10068643904747315\n",
      "458  Episode Reward:  212.13926345042933 .Epsilon: 0.10018300685223579\n",
      "459  Episode Reward:  -15.764346572906476 .Epsilon: 0.1\n",
      "460  Episode Reward:  -79.48731913191435 .Epsilon: 0.1\n",
      "461  Episode Reward:  -19.053441261225743 .Epsilon: 0.1\n",
      "462  Episode Reward:  -20.439635485494705 .Epsilon: 0.1\n",
      "463  Episode Reward:  -51.11902848334275 .Epsilon: 0.1\n",
      "464  Episode Reward:  -60.27161335298118 .Epsilon: 0.1\n",
      "465  Episode Reward:  213.991774127814 .Epsilon: 0.1\n",
      "466  Episode Reward:  -84.9948136732992 .Epsilon: 0.1\n",
      "467  Episode Reward:  -125.63275564612614 .Epsilon: 0.1\n",
      "468  Episode Reward:  235.94914960662084 .Epsilon: 0.1\n",
      "469  Episode Reward:  -72.94221516283751 .Epsilon: 0.1\n",
      "470  Episode Reward:  -35.11643071152787 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "471  Episode Reward:  236.05491602488198 .Epsilon: 0.1\n",
      "472  Episode Reward:  3.688792365924826 .Epsilon: 0.1\n",
      "473  Episode Reward:  21.07704812849441 .Epsilon: 0.1\n",
      "474  Episode Reward:  4.388056063662489 .Epsilon: 0.1\n",
      "475  Episode Reward:  191.3510598260236 .Epsilon: 0.1\n",
      "476  Episode Reward:  24.67288481660603 .Epsilon: 0.1\n",
      "477  Episode Reward:  -12.30152019990112 .Epsilon: 0.1\n",
      "478  Episode Reward:  -22.160419741266125 .Epsilon: 0.1\n",
      "479  Episode Reward:  275.00639063481736 .Epsilon: 0.1\n",
      "480  Episode Reward:  -39.84230605629542 .Epsilon: 0.1\n",
      "481  Episode Reward:  -77.61877762497093 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "482  Episode Reward:  185.3820874683054 .Epsilon: 0.1\n",
      "483  Episode Reward:  -68.60280981856485 .Epsilon: 0.1\n",
      "484  Episode Reward:  11.407191072921407 .Epsilon: 0.1\n",
      "485  Episode Reward:  -44.08040145396378 .Epsilon: 0.1\n",
      "486  Episode Reward:  -1.4017946483012764 .Epsilon: 0.1\n",
      "487  Episode Reward:  -6.223130979353542 .Epsilon: 0.1\n",
      "488  Episode Reward:  115.84620251980797 .Epsilon: 0.1\n",
      "489  Episode Reward:  -67.94030843703318 .Epsilon: 0.1\n",
      "490  Episode Reward:  196.45851035710314 .Epsilon: 0.1\n",
      "491  Episode Reward:  -34.56440576997606 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "492  Episode Reward:  -162.22015591439754 .Epsilon: 0.1\n",
      "493  Episode Reward:  207.63674205276442 .Epsilon: 0.1\n",
      "494  Episode Reward:  36.670773166461856 .Epsilon: 0.1\n",
      "495  Episode Reward:  -2.1625586388466047 .Epsilon: 0.1\n",
      "496  Episode Reward:  -35.66588283015972 .Epsilon: 0.1\n",
      "497  Episode Reward:  46.45185796225368 .Epsilon: 0.1\n",
      "498  Episode Reward:  165.79554453654097 .Epsilon: 0.1\n",
      "499  Episode Reward:  7.415487035306569 .Epsilon: 0.1\n"
     ]
    }
   ],
   "source": [
    "numberEpisodes = 500\n",
    "discount = 0.99\n",
    "counter = 0\n",
    "epsilon = 1\n",
    "epsilonDecay = 0.995\n",
    "rewards = []\n",
    "tau = 0.5\n",
    "memorySize = 100000\n",
    "memoryPreviousStates = np.zeros((memorySize,8))\n",
    "memoryStates = np.zeros((memorySize,8))\n",
    "memoryActions = np.zeros((memorySize,2))\n",
    "memoryRD = np.zeros((memorySize,2))\n",
    "numberOfLearningFromReplaySteps = 256\n",
    "itemsInMemory = 0\n",
    "initialiseReplayAfter = 1000 \n",
    "\n",
    "for episode in range (numberEpisodes):\n",
    "\n",
    "    epsilon = max (epsilon*epsilonDecay, 0.1)\n",
    "    done = False\n",
    "    prevObs = env.reset()\n",
    "    episodeReward = 0\n",
    "\n",
    "    for step in range (2000):\n",
    "        if done:\n",
    "            break\n",
    "        #env.render()\n",
    "        \n",
    "\n",
    "        if (random.random()<epsilon):\n",
    "            actions = [random.random()*2-1, random.random()*2-1]\n",
    "        else:\n",
    "            actions = actorModel.predict(np.expand_dims(prevObs, axis = 0))[0] # Shape (2,)\n",
    "\n",
    "        \n",
    "        obs, reward, done, _ = env.step (actions)\n",
    "\n",
    "\n",
    "        memoryPreviousStates[counter] = prevObs\n",
    "        memoryStates[counter] = obs\n",
    "        memoryActions[counter] = actions\n",
    "        memoryRD[counter] = [reward, float (done)]\n",
    "\n",
    "        \n",
    "        if (counter % 3000 == 0):\n",
    "            actorModel.save_weights(\"ACWeights.h5\")\n",
    "            actorModel, staticActorModel   = updateStaticWeights(actorModel, staticActorModel, tau)\n",
    "            criticModel, staticCriticModel = updateStaticWeights(criticModel, staticCriticModel, tau)\n",
    "            print (\"STATIC WEIGHTS UPDATED\")\n",
    "        \n",
    "        if (itemsInMemory < (counter+1)):\n",
    "            itemsInMemory = min ((itemsInMemory+1), memorySize)\n",
    "        if (itemsInMemory >= initialiseReplayAfter):\n",
    "            executeMemoryReplayStep(memoryPreviousStates, memoryStates, memoryActions, memoryRD, actorModel, staticActorModel, criticModel, staticCriticModel, itemsInMemory)\n",
    "\n",
    "        prevObs = obs\n",
    "        episodeReward += reward\n",
    "        counter = (counter + 1) % memorySize\n",
    "    print (episode, \" Episode Reward: \", episodeReward, \".Epsilon:\", epsilon)\n",
    "    rewards.append(episodeReward)\n",
    "\n",
    "env.close()\n",
    "actorModel.save_weights(\"ACWeights.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot (list(range(len(rewards))), rewards)\n",
    "plt.title(\"Rewards obtained along the learning process\")\n",
    "plt.xlabel(\"Episode number\")\n",
    "plt.ylabel(\"Reward obtained\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Let's train the agent for longer:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "500  Episode Reward:  220.82212084785942 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "501  Episode Reward:  23.433244226053873 .Epsilon: 0.1\n",
      "502  Episode Reward:  15.19106441873096 .Epsilon: 0.1\n",
      "503  Episode Reward:  9.272180270211422 .Epsilon: 0.1\n",
      "504  Episode Reward:  263.82156464350896 .Epsilon: 0.1\n",
      "505  Episode Reward:  199.4385284896282 .Epsilon: 0.1\n",
      "506  Episode Reward:  -2.815508908460785 .Epsilon: 0.1\n",
      "507  Episode Reward:  -147.63398892994434 .Epsilon: 0.1\n",
      "508  Episode Reward:  -38.606133024529946 .Epsilon: 0.1\n",
      "509  Episode Reward:  -38.23941677929537 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "510  Episode Reward:  4.3450222458047705 .Epsilon: 0.1\n",
      "511  Episode Reward:  13.494123868959164 .Epsilon: 0.1\n",
      "512  Episode Reward:  -54.65489627153236 .Epsilon: 0.1\n",
      "513  Episode Reward:  -56.930707321852964 .Epsilon: 0.1\n",
      "514  Episode Reward:  253.72946377426908 .Epsilon: 0.1\n",
      "515  Episode Reward:  -18.532616382574844 .Epsilon: 0.1\n",
      "516  Episode Reward:  278.7822976942335 .Epsilon: 0.1\n",
      "517  Episode Reward:  14.840986501349093 .Epsilon: 0.1\n",
      "518  Episode Reward:  -43.351544781220355 .Epsilon: 0.1\n",
      "519  Episode Reward:  -133.1790960850672 .Epsilon: 0.1\n",
      "520  Episode Reward:  -9.762908153859854 .Epsilon: 0.1\n",
      "521  Episode Reward:  -65.39639144246024 .Epsilon: 0.1\n",
      "522  Episode Reward:  232.14447721227938 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "523  Episode Reward:  19.592467904658008 .Epsilon: 0.1\n",
      "524  Episode Reward:  208.0133893075611 .Epsilon: 0.1\n",
      "525  Episode Reward:  262.18073777392493 .Epsilon: 0.1\n",
      "526  Episode Reward:  22.726950281460603 .Epsilon: 0.1\n",
      "527  Episode Reward:  -19.112566127905055 .Epsilon: 0.1\n",
      "528  Episode Reward:  -93.02861397358772 .Epsilon: 0.1\n",
      "529  Episode Reward:  -8.608385490936783 .Epsilon: 0.1\n",
      "530  Episode Reward:  -43.42610290593912 .Epsilon: 0.1\n",
      "531  Episode Reward:  274.40089831251885 .Epsilon: 0.1\n",
      "532  Episode Reward:  -66.1032559224735 .Epsilon: 0.1\n",
      "533  Episode Reward:  25.02974538359156 .Epsilon: 0.1\n",
      "534  Episode Reward:  -164.28232010908596 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "535  Episode Reward:  226.1236901296453 .Epsilon: 0.1\n",
      "536  Episode Reward:  10.617766752208539 .Epsilon: 0.1\n",
      "537  Episode Reward:  293.99248273485705 .Epsilon: 0.1\n",
      "538  Episode Reward:  12.973986169441346 .Epsilon: 0.1\n",
      "539  Episode Reward:  65.64872494591202 .Epsilon: 0.1\n",
      "540  Episode Reward:  34.64390446561026 .Epsilon: 0.1\n",
      "541  Episode Reward:  -70.56872185008844 .Epsilon: 0.1\n",
      "542  Episode Reward:  15.687644476798567 .Epsilon: 0.1\n",
      "543  Episode Reward:  -54.43282865386995 .Epsilon: 0.1\n",
      "544  Episode Reward:  -30.888698466232228 .Epsilon: 0.1\n",
      "545  Episode Reward:  -62.7653227664515 .Epsilon: 0.1\n",
      "546  Episode Reward:  -89.2500939272951 .Epsilon: 0.1\n",
      "547  Episode Reward:  -9.577738163764892 .Epsilon: 0.1\n",
      "548  Episode Reward:  220.30245020268794 .Epsilon: 0.1\n",
      "549  Episode Reward:  265.49069266632125 .Epsilon: 0.1\n",
      "550  Episode Reward:  -30.17518545741902 .Epsilon: 0.1\n",
      "551  Episode Reward:  7.184610956014026 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "552  Episode Reward:  10.204365725410128 .Epsilon: 0.1\n",
      "553  Episode Reward:  -50.183357442479 .Epsilon: 0.1\n",
      "554  Episode Reward:  49.67079129592986 .Epsilon: 0.1\n",
      "555  Episode Reward:  18.850501289282448 .Epsilon: 0.1\n",
      "556  Episode Reward:  275.39469501477276 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "557  Episode Reward:  -58.603791245113584 .Epsilon: 0.1\n",
      "558  Episode Reward:  24.452231743634 .Epsilon: 0.1\n",
      "559  Episode Reward:  -13.115075630582751 .Epsilon: 0.1\n",
      "560  Episode Reward:  18.746655887871086 .Epsilon: 0.1\n",
      "561  Episode Reward:  3.721165016738155 .Epsilon: 0.1\n",
      "562  Episode Reward:  -49.784474047691624 .Epsilon: 0.1\n",
      "563  Episode Reward:  -61.91537935523069 .Epsilon: 0.1\n",
      "564  Episode Reward:  -37.998712466600104 .Epsilon: 0.1\n",
      "565  Episode Reward:  -46.55223813474693 .Epsilon: 0.1\n",
      "566  Episode Reward:  -19.013327371005545 .Epsilon: 0.1\n",
      "567  Episode Reward:  -156.0580517921835 .Epsilon: 0.1\n",
      "568  Episode Reward:  273.19701914283235 .Epsilon: 0.1\n",
      "569  Episode Reward:  -66.59067463522396 .Epsilon: 0.1\n",
      "570  Episode Reward:  36.7177123900656 .Epsilon: 0.1\n",
      "571  Episode Reward:  -84.7594667934239 .Epsilon: 0.1\n",
      "572  Episode Reward:  14.09441394116314 .Epsilon: 0.1\n",
      "573  Episode Reward:  -13.810078917263127 .Epsilon: 0.1\n",
      "574  Episode Reward:  16.509902359483846 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "575  Episode Reward:  247.8532966118118 .Epsilon: 0.1\n",
      "576  Episode Reward:  -41.51453835820517 .Epsilon: 0.1\n",
      "577  Episode Reward:  -49.96219778359108 .Epsilon: 0.1\n",
      "578  Episode Reward:  -63.629250929279834 .Epsilon: 0.1\n",
      "579  Episode Reward:  -32.6397902500309 .Epsilon: 0.1\n",
      "580  Episode Reward:  222.23158692855992 .Epsilon: 0.1\n",
      "581  Episode Reward:  -54.223365587672966 .Epsilon: 0.1\n",
      "582  Episode Reward:  -68.4651084431021 .Epsilon: 0.1\n",
      "583  Episode Reward:  -31.051216227363 .Epsilon: 0.1\n",
      "584  Episode Reward:  -39.86273069422643 .Epsilon: 0.1\n",
      "585  Episode Reward:  -28.492156627721712 .Epsilon: 0.1\n",
      "586  Episode Reward:  -48.483619648678314 .Epsilon: 0.1\n",
      "587  Episode Reward:  -105.22228128670751 .Epsilon: 0.1\n",
      "588  Episode Reward:  -52.13216267758301 .Epsilon: 0.1\n",
      "589  Episode Reward:  -167.75657415126685 .Epsilon: 0.1\n",
      "590  Episode Reward:  -28.887830917482233 .Epsilon: 0.1\n",
      "591  Episode Reward:  245.55995401842662 .Epsilon: 0.1\n",
      "592  Episode Reward:  3.691818326599872 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "593  Episode Reward:  34.57743555486633 .Epsilon: 0.1\n",
      "594  Episode Reward:  -85.64988438215498 .Epsilon: 0.1\n",
      "595  Episode Reward:  3.0966006641100137 .Epsilon: 0.1\n",
      "596  Episode Reward:  -3.189802853604249 .Epsilon: 0.1\n",
      "597  Episode Reward:  -53.285564808868955 .Epsilon: 0.1\n",
      "598  Episode Reward:  -61.5608427517613 .Epsilon: 0.1\n",
      "599  Episode Reward:  8.283095047149743 .Epsilon: 0.1\n",
      "600  Episode Reward:  251.36819159537768 .Epsilon: 0.1\n",
      "601  Episode Reward:  -45.64462786066685 .Epsilon: 0.1\n",
      "602  Episode Reward:  221.29884672177786 .Epsilon: 0.1\n",
      "603  Episode Reward:  241.16148018891914 .Epsilon: 0.1\n",
      "604  Episode Reward:  -8.03499375479889 .Epsilon: 0.1\n",
      "605  Episode Reward:  -23.74410873086329 .Epsilon: 0.1\n",
      "606  Episode Reward:  -21.946835017378817 .Epsilon: 0.1\n",
      "607  Episode Reward:  -23.572123010639274 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "608  Episode Reward:  3.431527662388703 .Epsilon: 0.1\n",
      "609  Episode Reward:  -9.580261946782244 .Epsilon: 0.1\n",
      "610  Episode Reward:  -13.4110503309264 .Epsilon: 0.1\n",
      "611  Episode Reward:  -71.59227099141692 .Epsilon: 0.1\n",
      "612  Episode Reward:  -0.8468763120112044 .Epsilon: 0.1\n",
      "613  Episode Reward:  282.8491280656939 .Epsilon: 0.1\n",
      "614  Episode Reward:  39.96061961271636 .Epsilon: 0.1\n",
      "615  Episode Reward:  28.085575897167615 .Epsilon: 0.1\n",
      "616  Episode Reward:  273.17317166380406 .Epsilon: 0.1\n",
      "617  Episode Reward:  -156.31306689415464 .Epsilon: 0.1\n",
      "618  Episode Reward:  -22.53915120357314 .Epsilon: 0.1\n",
      "619  Episode Reward:  24.38211995587541 .Epsilon: 0.1\n",
      "620  Episode Reward:  -142.69788229341006 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "621  Episode Reward:  104.69631286969246 .Epsilon: 0.1\n",
      "622  Episode Reward:  187.8359789706667 .Epsilon: 0.1\n",
      "623  Episode Reward:  228.22685306174415 .Epsilon: 0.1\n",
      "624  Episode Reward:  -72.71784736189258 .Epsilon: 0.1\n",
      "625  Episode Reward:  267.3665111861625 .Epsilon: 0.1\n",
      "626  Episode Reward:  281.6106994386807 .Epsilon: 0.1\n",
      "627  Episode Reward:  257.24264320695477 .Epsilon: 0.1\n",
      "628  Episode Reward:  46.15222630179181 .Epsilon: 0.1\n",
      "629  Episode Reward:  -11.312612303791965 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "630  Episode Reward:  213.9672064577971 .Epsilon: 0.1\n",
      "631  Episode Reward:  278.72361149078654 .Epsilon: 0.1\n",
      "632  Episode Reward:  27.50579104573312 .Epsilon: 0.1\n",
      "633  Episode Reward:  220.50071119688607 .Epsilon: 0.1\n",
      "634  Episode Reward:  -2.549854985734413 .Epsilon: 0.1\n",
      "635  Episode Reward:  19.073441905183046 .Epsilon: 0.1\n",
      "636  Episode Reward:  23.418837588401246 .Epsilon: 0.1\n",
      "637  Episode Reward:  -0.4857818066456474 .Epsilon: 0.1\n",
      "638  Episode Reward:  31.15849242591267 .Epsilon: 0.1\n",
      "639  Episode Reward:  249.05939281001957 .Epsilon: 0.1\n",
      "640  Episode Reward:  9.665114714567054 .Epsilon: 0.1\n",
      "641  Episode Reward:  312.61568923066824 .Epsilon: 0.1\n",
      "642  Episode Reward:  232.2925662888911 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "643  Episode Reward:  207.06051753854703 .Epsilon: 0.1\n",
      "644  Episode Reward:  -32.01717135965821 .Epsilon: 0.1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "645  Episode Reward:  84.98654563242562 .Epsilon: 0.1\n",
      "646  Episode Reward:  250.25204627433502 .Epsilon: 0.1\n",
      "647  Episode Reward:  228.36640311492965 .Epsilon: 0.1\n",
      "648  Episode Reward:  8.726860884097078 .Epsilon: 0.1\n",
      "649  Episode Reward:  -36.82865504137557 .Epsilon: 0.1\n",
      "650  Episode Reward:  4.252689864397112 .Epsilon: 0.1\n",
      "651  Episode Reward:  277.8296928345976 .Epsilon: 0.1\n",
      "652  Episode Reward:  -1.9809180688664014 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "653  Episode Reward:  -58.10499494032992 .Epsilon: 0.1\n",
      "654  Episode Reward:  191.00899380189514 .Epsilon: 0.1\n",
      "655  Episode Reward:  267.2109789320141 .Epsilon: 0.1\n",
      "656  Episode Reward:  -21.854322244087626 .Epsilon: 0.1\n",
      "657  Episode Reward:  -19.387618779964683 .Epsilon: 0.1\n",
      "658  Episode Reward:  14.20253379091234 .Epsilon: 0.1\n",
      "659  Episode Reward:  238.5744548510841 .Epsilon: 0.1\n",
      "660  Episode Reward:  -2.2612308459861197 .Epsilon: 0.1\n",
      "661  Episode Reward:  239.4680894541447 .Epsilon: 0.1\n",
      "662  Episode Reward:  302.2096048867712 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "663  Episode Reward:  163.01368785987867 .Epsilon: 0.1\n",
      "664  Episode Reward:  264.7779177660825 .Epsilon: 0.1\n",
      "665  Episode Reward:  -82.89176524797061 .Epsilon: 0.1\n",
      "666  Episode Reward:  -367.89859095912067 .Epsilon: 0.1\n",
      "667  Episode Reward:  45.28593578169355 .Epsilon: 0.1\n",
      "668  Episode Reward:  232.96023481596606 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "669  Episode Reward:  180.58493158967772 .Epsilon: 0.1\n",
      "670  Episode Reward:  17.657114527292265 .Epsilon: 0.1\n",
      "671  Episode Reward:  -131.09853037094146 .Epsilon: 0.1\n",
      "672  Episode Reward:  -21.822980365330537 .Epsilon: 0.1\n",
      "673  Episode Reward:  7.71371525658266 .Epsilon: 0.1\n",
      "674  Episode Reward:  -108.04178151699875 .Epsilon: 0.1\n",
      "675  Episode Reward:  -71.6280676075743 .Epsilon: 0.1\n",
      "676  Episode Reward:  274.5051990963333 .Epsilon: 0.1\n",
      "677  Episode Reward:  19.044147408540525 .Epsilon: 0.1\n",
      "678  Episode Reward:  270.20935580478607 .Epsilon: 0.1\n",
      "679  Episode Reward:  235.86339035270407 .Epsilon: 0.1\n",
      "680  Episode Reward:  257.24098014062906 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "681  Episode Reward:  264.310679763522 .Epsilon: 0.1\n",
      "682  Episode Reward:  275.43846008977135 .Epsilon: 0.1\n",
      "683  Episode Reward:  138.03859185005356 .Epsilon: 0.1\n",
      "684  Episode Reward:  260.35932108210295 .Epsilon: 0.1\n",
      "685  Episode Reward:  282.7243253993073 .Epsilon: 0.1\n",
      "686  Episode Reward:  290.7404570442243 .Epsilon: 0.1\n",
      "687  Episode Reward:  -73.99591935780026 .Epsilon: 0.1\n",
      "688  Episode Reward:  271.08651118245666 .Epsilon: 0.1\n",
      "689  Episode Reward:  281.03616768746366 .Epsilon: 0.1\n",
      "690  Episode Reward:  277.7461735574746 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "691  Episode Reward:  250.27776303692588 .Epsilon: 0.1\n",
      "692  Episode Reward:  279.80515350419716 .Epsilon: 0.1\n",
      "693  Episode Reward:  243.04762399849187 .Epsilon: 0.1\n",
      "694  Episode Reward:  284.13807000969297 .Epsilon: 0.1\n",
      "695  Episode Reward:  44.01603125451638 .Epsilon: 0.1\n",
      "696  Episode Reward:  -37.12941148135093 .Epsilon: 0.1\n",
      "697  Episode Reward:  -62.879036095426684 .Epsilon: 0.1\n",
      "698  Episode Reward:  31.47664473393266 .Epsilon: 0.1\n",
      "699  Episode Reward:  259.02252419796946 .Epsilon: 0.1\n",
      "700  Episode Reward:  296.45572461397484 .Epsilon: 0.1\n",
      "701  Episode Reward:  -9.15142678339673 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "702  Episode Reward:  147.12364878914622 .Epsilon: 0.1\n",
      "703  Episode Reward:  226.86641581632375 .Epsilon: 0.1\n",
      "704  Episode Reward:  68.16232212099717 .Epsilon: 0.1\n",
      "705  Episode Reward:  273.70238338722294 .Epsilon: 0.1\n",
      "706  Episode Reward:  9.695959426797728 .Epsilon: 0.1\n",
      "707  Episode Reward:  13.194847793197141 .Epsilon: 0.1\n",
      "708  Episode Reward:  103.94697893210896 .Epsilon: 0.1\n",
      "709  Episode Reward:  251.06961190346848 .Epsilon: 0.1\n",
      "710  Episode Reward:  67.03324340471482 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "711  Episode Reward:  260.59183604918064 .Epsilon: 0.1\n",
      "712  Episode Reward:  0.06899219065331863 .Epsilon: 0.1\n",
      "713  Episode Reward:  -123.18028869776634 .Epsilon: 0.1\n",
      "714  Episode Reward:  263.1559152340932 .Epsilon: 0.1\n",
      "715  Episode Reward:  242.8962631172202 .Epsilon: 0.1\n",
      "716  Episode Reward:  249.7270374136625 .Epsilon: 0.1\n",
      "717  Episode Reward:  -111.98675465322734 .Epsilon: 0.1\n",
      "718  Episode Reward:  276.80027251004765 .Epsilon: 0.1\n",
      "719  Episode Reward:  251.27725198254532 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "720  Episode Reward:  187.71651336390616 .Epsilon: 0.1\n",
      "721  Episode Reward:  33.74867092942509 .Epsilon: 0.1\n",
      "722  Episode Reward:  197.43561919526542 .Epsilon: 0.1\n",
      "723  Episode Reward:  -61.990763246675 .Epsilon: 0.1\n",
      "724  Episode Reward:  273.68329967202544 .Epsilon: 0.1\n",
      "725  Episode Reward:  159.7770693655526 .Epsilon: 0.1\n",
      "726  Episode Reward:  73.7494526291681 .Epsilon: 0.1\n",
      "727  Episode Reward:  268.85771676969466 .Epsilon: 0.1\n",
      "728  Episode Reward:  289.7169813132278 .Epsilon: 0.1\n",
      "729  Episode Reward:  264.9134606391964 .Epsilon: 0.1\n",
      "730  Episode Reward:  273.25851062447964 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "731  Episode Reward:  249.94884792017834 .Epsilon: 0.1\n",
      "732  Episode Reward:  63.06809840054731 .Epsilon: 0.1\n",
      "733  Episode Reward:  239.60096324584296 .Epsilon: 0.1\n",
      "734  Episode Reward:  236.03265978557104 .Epsilon: 0.1\n",
      "735  Episode Reward:  295.4336409450649 .Epsilon: 0.1\n",
      "736  Episode Reward:  289.9557047798735 .Epsilon: 0.1\n",
      "737  Episode Reward:  217.06958324284182 .Epsilon: 0.1\n",
      "738  Episode Reward:  279.20114470447845 .Epsilon: 0.1\n",
      "739  Episode Reward:  258.33324482423143 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "740  Episode Reward:  -3.1547363196987948 .Epsilon: 0.1\n",
      "741  Episode Reward:  270.02364456977637 .Epsilon: 0.1\n",
      "742  Episode Reward:  256.5185660815624 .Epsilon: 0.1\n",
      "743  Episode Reward:  -237.0927344368734 .Epsilon: 0.1\n",
      "744  Episode Reward:  276.90062063050954 .Epsilon: 0.1\n",
      "745  Episode Reward:  264.12473057902605 .Epsilon: 0.1\n",
      "746  Episode Reward:  -14.99693154813771 .Epsilon: 0.1\n",
      "747  Episode Reward:  275.5189182704081 .Epsilon: 0.1\n",
      "748  Episode Reward:  275.23108239382043 .Epsilon: 0.1\n",
      "749  Episode Reward:  278.2489822846761 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "750  Episode Reward:  229.18173122583192 .Epsilon: 0.1\n",
      "751  Episode Reward:  243.9968505333066 .Epsilon: 0.1\n",
      "752  Episode Reward:  266.45651588983657 .Epsilon: 0.1\n",
      "753  Episode Reward:  253.01850011487497 .Epsilon: 0.1\n",
      "754  Episode Reward:  35.13105503370869 .Epsilon: 0.1\n",
      "755  Episode Reward:  255.96022175487326 .Epsilon: 0.1\n",
      "756  Episode Reward:  246.84339764835818 .Epsilon: 0.1\n",
      "757  Episode Reward:  257.58341893037004 .Epsilon: 0.1\n",
      "758  Episode Reward:  262.88306393930816 .Epsilon: 0.1\n",
      "759  Episode Reward:  276.79905852040804 .Epsilon: 0.1\n",
      "760  Episode Reward:  275.5561841915488 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "761  Episode Reward:  260.8784336066803 .Epsilon: 0.1\n",
      "762  Episode Reward:  -10.131509707890913 .Epsilon: 0.1\n",
      "763  Episode Reward:  -96.37392498480354 .Epsilon: 0.1\n",
      "764  Episode Reward:  14.271951995265539 .Epsilon: 0.1\n",
      "765  Episode Reward:  252.5296236960445 .Epsilon: 0.1\n",
      "766  Episode Reward:  167.01400399174298 .Epsilon: 0.1\n",
      "767  Episode Reward:  209.5048981491137 .Epsilon: 0.1\n",
      "768  Episode Reward:  -60.56371601654604 .Epsilon: 0.1\n",
      "769  Episode Reward:  -54.33494529158097 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "770  Episode Reward:  -259.9120842194984 .Epsilon: 0.1\n",
      "771  Episode Reward:  -165.15817235954688 .Epsilon: 0.1\n",
      "772  Episode Reward:  -47.360110337458856 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "773  Episode Reward:  130.41519336492138 .Epsilon: 0.1\n",
      "774  Episode Reward:  -245.80869229410067 .Epsilon: 0.1\n",
      "775  Episode Reward:  124.79411520969171 .Epsilon: 0.1\n",
      "776  Episode Reward:  -164.9298728066977 .Epsilon: 0.1\n",
      "777  Episode Reward:  -118.54778354406119 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "778  Episode Reward:  69.73235724238548 .Epsilon: 0.1\n",
      "779  Episode Reward:  -3.346099248748544 .Epsilon: 0.1\n",
      "780  Episode Reward:  312.3042339045317 .Epsilon: 0.1\n",
      "781  Episode Reward:  -32.21630447439199 .Epsilon: 0.1\n",
      "782  Episode Reward:  3.2779357507540254 .Epsilon: 0.1\n",
      "783  Episode Reward:  52.389523158693834 .Epsilon: 0.1\n",
      "784  Episode Reward:  12.716752997825992 .Epsilon: 0.1\n",
      "785  Episode Reward:  182.42858175373044 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "786  Episode Reward:  -181.41496987654875 .Epsilon: 0.1\n",
      "787  Episode Reward:  -102.12162193659759 .Epsilon: 0.1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "788  Episode Reward:  -92.50678868819782 .Epsilon: 0.1\n",
      "789  Episode Reward:  61.25337162862664 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "790  Episode Reward:  -313.75560897014896 .Epsilon: 0.1\n",
      "791  Episode Reward:  -72.08371011120207 .Epsilon: 0.1\n",
      "792  Episode Reward:  224.7677260011943 .Epsilon: 0.1\n",
      "793  Episode Reward:  -17.537809687530597 .Epsilon: 0.1\n",
      "794  Episode Reward:  -104.9517205308236 .Epsilon: 0.1\n",
      "795  Episode Reward:  258.92380162617496 .Epsilon: 0.1\n",
      "796  Episode Reward:  48.045955145372034 .Epsilon: 0.1\n",
      "797  Episode Reward:  20.342048164019133 .Epsilon: 0.1\n",
      "798  Episode Reward:  17.3698996646704 .Epsilon: 0.1\n",
      "799  Episode Reward:  -234.10016948089873 .Epsilon: 0.1\n",
      "800  Episode Reward:  -169.70994039570073 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "801  Episode Reward:  1.3468104338724487 .Epsilon: 0.1\n",
      "802  Episode Reward:  275.1364135726262 .Epsilon: 0.1\n",
      "803  Episode Reward:  30.961531919747642 .Epsilon: 0.1\n",
      "804  Episode Reward:  201.1281715258213 .Epsilon: 0.1\n",
      "805  Episode Reward:  -198.68607877872452 .Epsilon: 0.1\n",
      "806  Episode Reward:  -280.05396425137565 .Epsilon: 0.1\n",
      "807  Episode Reward:  274.62457807249723 .Epsilon: 0.1\n",
      "808  Episode Reward:  -127.49295474934081 .Epsilon: 0.1\n",
      "809  Episode Reward:  -33.148434589703726 .Epsilon: 0.1\n",
      "810  Episode Reward:  -115.04237932960596 .Epsilon: 0.1\n",
      "811  Episode Reward:  226.84130948314908 .Epsilon: 0.1\n",
      "812  Episode Reward:  -187.00310509532906 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "813  Episode Reward:  242.15570284815934 .Epsilon: 0.1\n",
      "814  Episode Reward:  11.995069465948617 .Epsilon: 0.1\n",
      "815  Episode Reward:  272.350478122856 .Epsilon: 0.1\n",
      "816  Episode Reward:  227.83403668839168 .Epsilon: 0.1\n",
      "817  Episode Reward:  -35.50512584612461 .Epsilon: 0.1\n",
      "818  Episode Reward:  -205.95543240026686 .Epsilon: 0.1\n",
      "819  Episode Reward:  -33.88412067998651 .Epsilon: 0.1\n",
      "820  Episode Reward:  -1.2392463593999707 .Epsilon: 0.1\n",
      "821  Episode Reward:  -309.3090115995603 .Epsilon: 0.1\n",
      "822  Episode Reward:  -324.20385383651933 .Epsilon: 0.1\n",
      "823  Episode Reward:  -39.05777069088468 .Epsilon: 0.1\n",
      "824  Episode Reward:  -4.930142873912075 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "825  Episode Reward:  162.13833610237194 .Epsilon: 0.1\n",
      "826  Episode Reward:  -42.010596193151834 .Epsilon: 0.1\n",
      "827  Episode Reward:  -143.87719601198498 .Epsilon: 0.1\n",
      "828  Episode Reward:  41.118010998500154 .Epsilon: 0.1\n",
      "829  Episode Reward:  219.36936028006417 .Epsilon: 0.1\n",
      "830  Episode Reward:  6.085616451704908 .Epsilon: 0.1\n",
      "831  Episode Reward:  -294.24213808958166 .Epsilon: 0.1\n",
      "832  Episode Reward:  12.840694035991987 .Epsilon: 0.1\n",
      "833  Episode Reward:  -100.70926374078721 .Epsilon: 0.1\n",
      "834  Episode Reward:  227.8523133049773 .Epsilon: 0.1\n",
      "835  Episode Reward:  2.530306102430899 .Epsilon: 0.1\n",
      "836  Episode Reward:  -327.64118633047724 .Epsilon: 0.1\n",
      "837  Episode Reward:  -256.1457674076569 .Epsilon: 0.1\n",
      "838  Episode Reward:  -31.281163622770137 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "839  Episode Reward:  156.24753105191698 .Epsilon: 0.1\n",
      "840  Episode Reward:  9.657200546097826 .Epsilon: 0.1\n",
      "841  Episode Reward:  260.3512341169349 .Epsilon: 0.1\n",
      "842  Episode Reward:  -187.33837676002682 .Epsilon: 0.1\n",
      "843  Episode Reward:  -23.527811643959296 .Epsilon: 0.1\n",
      "844  Episode Reward:  -3.24202819963719 .Epsilon: 0.1\n",
      "845  Episode Reward:  -278.59613908007606 .Epsilon: 0.1\n",
      "846  Episode Reward:  271.60511032956896 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "847  Episode Reward:  -0.9980288279074282 .Epsilon: 0.1\n",
      "848  Episode Reward:  7.1797990368589595 .Epsilon: 0.1\n",
      "849  Episode Reward:  149.61524714224487 .Epsilon: 0.1\n",
      "850  Episode Reward:  -98.61032268079352 .Epsilon: 0.1\n",
      "851  Episode Reward:  -98.38232872350602 .Epsilon: 0.1\n",
      "852  Episode Reward:  -223.22144859757844 .Epsilon: 0.1\n",
      "853  Episode Reward:  248.5110934808245 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "854  Episode Reward:  -17.64831852176271 .Epsilon: 0.1\n",
      "855  Episode Reward:  -38.538423291346476 .Epsilon: 0.1\n",
      "856  Episode Reward:  -2.1377254455702825 .Epsilon: 0.1\n",
      "857  Episode Reward:  26.746338525212437 .Epsilon: 0.1\n",
      "858  Episode Reward:  -190.71710662790855 .Epsilon: 0.1\n",
      "859  Episode Reward:  39.109163017313364 .Epsilon: 0.1\n",
      "860  Episode Reward:  -168.4050135398731 .Epsilon: 0.1\n",
      "861  Episode Reward:  20.128854610526275 .Epsilon: 0.1\n",
      "862  Episode Reward:  5.3256854614665405 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "863  Episode Reward:  -234.98223375077853 .Epsilon: 0.1\n",
      "864  Episode Reward:  270.80483400481677 .Epsilon: 0.1\n",
      "865  Episode Reward:  -87.32522914266373 .Epsilon: 0.1\n",
      "866  Episode Reward:  4.581990322316969 .Epsilon: 0.1\n",
      "867  Episode Reward:  -6.2099443828682315 .Epsilon: 0.1\n",
      "868  Episode Reward:  -115.16781000240404 .Epsilon: 0.1\n",
      "869  Episode Reward:  0.737508678925721 .Epsilon: 0.1\n",
      "870  Episode Reward:  -29.28363165347696 .Epsilon: 0.1\n",
      "871  Episode Reward:  54.63140840279749 .Epsilon: 0.1\n",
      "872  Episode Reward:  -14.632138088293658 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "873  Episode Reward:  -378.54373755476394 .Epsilon: 0.1\n",
      "874  Episode Reward:  -273.3545020644379 .Epsilon: 0.1\n",
      "875  Episode Reward:  -48.608138116493215 .Epsilon: 0.1\n",
      "876  Episode Reward:  -527.5226868839429 .Epsilon: 0.1\n",
      "877  Episode Reward:  210.82807925493066 .Epsilon: 0.1\n",
      "878  Episode Reward:  -287.37316360224906 .Epsilon: 0.1\n",
      "879  Episode Reward:  -51.69858581286287 .Epsilon: 0.1\n",
      "880  Episode Reward:  -175.42194629368652 .Epsilon: 0.1\n",
      "881  Episode Reward:  -277.09574672922656 .Epsilon: 0.1\n",
      "882  Episode Reward:  280.2439520179148 .Epsilon: 0.1\n",
      "883  Episode Reward:  -20.495033257942822 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "884  Episode Reward:  194.89526693996785 .Epsilon: 0.1\n",
      "885  Episode Reward:  -86.58545903060985 .Epsilon: 0.1\n",
      "886  Episode Reward:  -234.70803122231638 .Epsilon: 0.1\n",
      "887  Episode Reward:  -25.24118402941353 .Epsilon: 0.1\n",
      "888  Episode Reward:  -16.84616980413115 .Epsilon: 0.1\n",
      "889  Episode Reward:  -8.371998013982036 .Epsilon: 0.1\n",
      "890  Episode Reward:  -247.24162120026014 .Epsilon: 0.1\n",
      "891  Episode Reward:  258.09721647626816 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "892  Episode Reward:  -241.20306375998604 .Epsilon: 0.1\n",
      "893  Episode Reward:  243.39966957573245 .Epsilon: 0.1\n",
      "894  Episode Reward:  52.33980447023578 .Epsilon: 0.1\n",
      "895  Episode Reward:  -264.73418926103557 .Epsilon: 0.1\n",
      "896  Episode Reward:  -468.45372929455704 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "897  Episode Reward:  291.1600829810165 .Epsilon: 0.1\n",
      "898  Episode Reward:  -109.8740802134183 .Epsilon: 0.1\n",
      "899  Episode Reward:  -207.4878402518499 .Epsilon: 0.1\n",
      "900  Episode Reward:  -3.0969439938031513 .Epsilon: 0.1\n",
      "901  Episode Reward:  -34.369454004245625 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "902  Episode Reward:  -49.36053474302817 .Epsilon: 0.1\n",
      "903  Episode Reward:  -4.775039484889234 .Epsilon: 0.1\n",
      "904  Episode Reward:  -490.62952323504703 .Epsilon: 0.1\n",
      "905  Episode Reward:  284.50181544122466 .Epsilon: 0.1\n",
      "906  Episode Reward:  -57.97699235779879 .Epsilon: 0.1\n",
      "907  Episode Reward:  -178.71289659259804 .Epsilon: 0.1\n",
      "908  Episode Reward:  -25.844040208686877 .Epsilon: 0.1\n",
      "909  Episode Reward:  -23.281594798226166 .Epsilon: 0.1\n",
      "910  Episode Reward:  -59.168984609296885 .Epsilon: 0.1\n",
      "911  Episode Reward:  234.81555740432728 .Epsilon: 0.1\n",
      "912  Episode Reward:  255.24444921587659 .Epsilon: 0.1\n",
      "913  Episode Reward:  14.277280246068912 .Epsilon: 0.1\n",
      "914  Episode Reward:  -4.107392255314522 .Epsilon: 0.1\n",
      "915  Episode Reward:  16.04176334329425 .Epsilon: 0.1\n",
      "916  Episode Reward:  35.22044291743106 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "917  Episode Reward:  -13.58446353026406 .Epsilon: 0.1\n",
      "918  Episode Reward:  25.02452531184126 .Epsilon: 0.1\n",
      "919  Episode Reward:  -8.286046743884498 .Epsilon: 0.1\n",
      "920  Episode Reward:  -110.6167855424193 .Epsilon: 0.1\n",
      "921  Episode Reward:  -67.97080051332975 .Epsilon: 0.1\n",
      "922  Episode Reward:  -172.96715342544667 .Epsilon: 0.1\n",
      "923  Episode Reward:  -38.52177052296176 .Epsilon: 0.1\n",
      "924  Episode Reward:  246.2849955388318 .Epsilon: 0.1\n",
      "925  Episode Reward:  270.945083671067 .Epsilon: 0.1\n",
      "926  Episode Reward:  5.058312685489426 .Epsilon: 0.1\n",
      "927  Episode Reward:  -29.007529052493666 .Epsilon: 0.1\n",
      "928  Episode Reward:  -93.74759822010644 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "929  Episode Reward:  -12.846804536463312 .Epsilon: 0.1\n",
      "930  Episode Reward:  -201.0190424081382 .Epsilon: 0.1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "931  Episode Reward:  -36.12114672571105 .Epsilon: 0.1\n",
      "932  Episode Reward:  -4.8915246220790465 .Epsilon: 0.1\n",
      "933  Episode Reward:  12.375639632267124 .Epsilon: 0.1\n",
      "934  Episode Reward:  9.336830815532338 .Epsilon: 0.1\n",
      "935  Episode Reward:  38.30957944977675 .Epsilon: 0.1\n",
      "936  Episode Reward:  261.39310477146626 .Epsilon: 0.1\n",
      "937  Episode Reward:  11.14152146083292 .Epsilon: 0.1\n",
      "938  Episode Reward:  -160.02410306858786 .Epsilon: 0.1\n",
      "939  Episode Reward:  -69.75553802289357 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "940  Episode Reward:  232.56362381196584 .Epsilon: 0.1\n",
      "941  Episode Reward:  -49.2244858353978 .Epsilon: 0.1\n",
      "942  Episode Reward:  -6.121159983346491 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "943  Episode Reward:  -21.102391302166712 .Epsilon: 0.1\n",
      "944  Episode Reward:  287.17752727887705 .Epsilon: 0.1\n",
      "945  Episode Reward:  14.932266486467313 .Epsilon: 0.1\n",
      "946  Episode Reward:  252.37150507047278 .Epsilon: 0.1\n",
      "947  Episode Reward:  -18.115017281896467 .Epsilon: 0.1\n",
      "948  Episode Reward:  -100.21400204910938 .Epsilon: 0.1\n",
      "949  Episode Reward:  229.47892728531224 .Epsilon: 0.1\n",
      "950  Episode Reward:  17.779804625196775 .Epsilon: 0.1\n",
      "951  Episode Reward:  -50.54646143026051 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "952  Episode Reward:  8.988159351738275 .Epsilon: 0.1\n",
      "953  Episode Reward:  -70.4217007357394 .Epsilon: 0.1\n",
      "954  Episode Reward:  232.19451028783567 .Epsilon: 0.1\n",
      "955  Episode Reward:  -38.02144791460913 .Epsilon: 0.1\n",
      "956  Episode Reward:  178.01619836945753 .Epsilon: 0.1\n",
      "957  Episode Reward:  282.5940258253514 .Epsilon: 0.1\n",
      "958  Episode Reward:  8.635031602764073 .Epsilon: 0.1\n",
      "959  Episode Reward:  -274.98877026454977 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "960  Episode Reward:  -236.6458332205568 .Epsilon: 0.1\n",
      "961  Episode Reward:  -8.922469552590641 .Epsilon: 0.1\n",
      "962  Episode Reward:  272.5555892642701 .Epsilon: 0.1\n",
      "963  Episode Reward:  -104.56553374188515 .Epsilon: 0.1\n",
      "964  Episode Reward:  279.66080893114656 .Epsilon: 0.1\n",
      "965  Episode Reward:  -11.222649121372413 .Epsilon: 0.1\n",
      "966  Episode Reward:  -109.53484043385635 .Epsilon: 0.1\n",
      "967  Episode Reward:  -27.011684316427548 .Epsilon: 0.1\n",
      "968  Episode Reward:  193.91766928062788 .Epsilon: 0.1\n",
      "969  Episode Reward:  -87.7609259663734 .Epsilon: 0.1\n",
      "970  Episode Reward:  -68.7417623057832 .Epsilon: 0.1\n",
      "971  Episode Reward:  -53.17674741856438 .Epsilon: 0.1\n",
      "972  Episode Reward:  20.125663608198522 .Epsilon: 0.1\n",
      "973  Episode Reward:  -161.72270836184384 .Epsilon: 0.1\n",
      "974  Episode Reward:  -43.11926203018084 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "975  Episode Reward:  -80.1186015181455 .Epsilon: 0.1\n",
      "976  Episode Reward:  -166.552519561578 .Epsilon: 0.1\n",
      "977  Episode Reward:  15.513265618297211 .Epsilon: 0.1\n",
      "978  Episode Reward:  -124.40113972451945 .Epsilon: 0.1\n",
      "979  Episode Reward:  -82.62619945613744 .Epsilon: 0.1\n",
      "980  Episode Reward:  -79.71186896745837 .Epsilon: 0.1\n",
      "981  Episode Reward:  -69.67362477326733 .Epsilon: 0.1\n",
      "982  Episode Reward:  -2.467557349999339 .Epsilon: 0.1\n",
      "983  Episode Reward:  -37.81665758848337 .Epsilon: 0.1\n",
      "984  Episode Reward:  24.086423913607277 .Epsilon: 0.1\n",
      "985  Episode Reward:  -239.5113329070062 .Epsilon: 0.1\n",
      "986  Episode Reward:  8.499113361524365 .Epsilon: 0.1\n",
      "987  Episode Reward:  20.534984827422345 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "988  Episode Reward:  261.99985952875795 .Epsilon: 0.1\n",
      "989  Episode Reward:  5.122143019308298 .Epsilon: 0.1\n",
      "990  Episode Reward:  43.832350839188564 .Epsilon: 0.1\n",
      "991  Episode Reward:  -49.07587503275956 .Epsilon: 0.1\n",
      "992  Episode Reward:  193.10752780456716 .Epsilon: 0.1\n",
      "STATIC WEIGHTS UPDATED\n",
      "993  Episode Reward:  45.4568922693814 .Epsilon: 0.1\n",
      "994  Episode Reward:  2.750466611145839 .Epsilon: 0.1\n",
      "995  Episode Reward:  -103.46131221699511 .Epsilon: 0.1\n",
      "996  Episode Reward:  -102.60048125859524 .Epsilon: 0.1\n",
      "997  Episode Reward:  -13.561244394467309 .Epsilon: 0.1\n",
      "998  Episode Reward:  -87.4261523042874 .Epsilon: 0.1\n",
      "999  Episode Reward:  -64.11973942433502 .Epsilon: 0.1\n"
     ]
    }
   ],
   "source": [
    "for episode in range (500, 1000):\n",
    "\n",
    "    epsilon = max (epsilon*epsilonDecay, 0.1)\n",
    "    done = False\n",
    "    prevObs = env.reset()\n",
    "    episodeReward = 0\n",
    "\n",
    "    for step in range (2000):\n",
    "        if done:\n",
    "            break\n",
    "        #env.render()\n",
    "        \n",
    "\n",
    "        if (random.random()<epsilon):\n",
    "            actions = [random.random()*2-1, random.random()*2-1]\n",
    "        else:\n",
    "            actions = actorModel.predict(np.expand_dims(prevObs, axis = 0))[0] # Shape (2,)\n",
    "\n",
    "        \n",
    "        obs, reward, done, _ = env.step (actions)\n",
    "\n",
    "\n",
    "        memoryPreviousStates[counter] = prevObs\n",
    "        memoryStates[counter] = obs\n",
    "        memoryActions[counter] = actions\n",
    "        memoryRD[counter] = [reward, float (done)]\n",
    "\n",
    "        \n",
    "        if (counter % 3000 == 0):\n",
    "            actorModel.save_weights(\"ACWeights.h5\")\n",
    "            actorModel, staticActorModel   = updateStaticWeights(actorModel, staticActorModel, tau)\n",
    "            criticModel, staticCriticModel = updateStaticWeights(criticModel, staticCriticModel, tau)\n",
    "            print (\"STATIC WEIGHTS UPDATED\")\n",
    "        \n",
    "        if (itemsInMemory < (counter+1)):\n",
    "            itemsInMemory = min ((itemsInMemory+1), memorySize)\n",
    "        if (itemsInMemory >= initialiseReplayAfter):\n",
    "            executeMemoryReplayStep(memoryPreviousStates, memoryStates, memoryActions, memoryRD, actorModel, staticActorModel, criticModel, staticCriticModel, itemsInMemory)\n",
    "\n",
    "        prevObs = obs\n",
    "        episodeReward += reward\n",
    "        counter = (counter + 1) % memorySize\n",
    "    print (episode, \" Episode Reward: \", episodeReward, \".Epsilon:\", epsilon)\n",
    "    rewards.append(episodeReward)\n",
    "\n",
    "env.close()\n",
    "actorModel.save_weights(\"ACWeights.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot (list(range(len(rewards))), rewards)\n",
    "plt.title(\"Rewards obtained along the learning process\")\n",
    "plt.xlabel(\"Episode number\")\n",
    "plt.ylabel(\"Reward obtained\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### It seems like the agent was learning till the episode number 700-750 after which it started being more unstable. Let's see how it behaves after 1000 episodes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0  Episode Reward:  -141.73831187858465\n",
      "1  Episode Reward:  16.444819945082827\n",
      "2  Episode Reward:  -99.00878881743233\n",
      "3  Episode Reward:  -79.5641804614252\n",
      "4  Episode Reward:  -116.87796199820517\n",
      "5  Episode Reward:  -111.48274261304185\n",
      "6  Episode Reward:  -107.19208756826605\n",
      "7  Episode Reward:  -104.31122730204464\n",
      "8  Episode Reward:  -96.21781593522245\n",
      "9  Episode Reward:  185.43544132550164\n",
      "10  Episode Reward:  -33.26735269952488\n",
      "11  Episode Reward:  -35.08828632123557\n",
      "12  Episode Reward:  -90.09567623089464\n",
      "13  Episode Reward:  -70.39016928375659\n",
      "14  Episode Reward:  229.78536873641127\n",
      "15  Episode Reward:  -80.80289663724385\n",
      "16  Episode Reward:  -171.26420070479548\n",
      "17  Episode Reward:  -125.03273556748276\n",
      "18  Episode Reward:  276.3827482785854\n",
      "19  Episode Reward:  -101.99332030594735\n",
      "20  Episode Reward:  -101.18251923685455\n",
      "21  Episode Reward:  266.2614461213449\n",
      "22  Episode Reward:  53.06938313652111\n",
      "23  Episode Reward:  -116.37385756593895\n",
      "24  Episode Reward:  42.329400075738505\n",
      "25  Episode Reward:  -18.335985110578108\n",
      "26  Episode Reward:  -185.1246135095506\n",
      "27  Episode Reward:  -56.481981588836746\n",
      "28  Episode Reward:  -2.532712655814052\n",
      "29  Episode Reward:  -86.78947392549294\n",
      "30  Episode Reward:  164.2986894401153\n",
      "31  Episode Reward:  -82.02102410694354\n",
      "32  Episode Reward:  -64.88674456674315\n",
      "33  Episode Reward:  -170.09414811754942\n",
      "34  Episode Reward:  -74.25798691190934\n",
      "35  Episode Reward:  -75.47297952493322\n",
      "36  Episode Reward:  -131.89871792085313\n",
      "37  Episode Reward:  10.804293054462988\n",
      "38  Episode Reward:  -98.24196991231209\n",
      "39  Episode Reward:  -71.85230173647636\n",
      "40  Episode Reward:  -109.40980272626508\n",
      "41  Episode Reward:  -72.42020234877072\n",
      "42  Episode Reward:  -20.434291608038905\n",
      "43  Episode Reward:  236.23632370306717\n",
      "44  Episode Reward:  -187.7874212987747\n",
      "45  Episode Reward:  20.73851251951919\n",
      "46  Episode Reward:  -87.29810205732855\n",
      "47  Episode Reward:  13.686882327983298\n",
      "48  Episode Reward:  164.30064685473872\n",
      "49  Episode Reward:  197.53020381094422\n",
      "50  Episode Reward:  -167.7092029614181\n",
      "51  Episode Reward:  293.079354624655\n",
      "52  Episode Reward:  11.59678831718341\n",
      "53  Episode Reward:  -27.14054446362732\n",
      "54  Episode Reward:  -34.30525919323367\n",
      "55  Episode Reward:  -20.447632088646515\n",
      "56  Episode Reward:  -166.47957402922134\n",
      "57  Episode Reward:  -78.15924709702955\n",
      "58  Episode Reward:  -84.35091550691672\n",
      "59  Episode Reward:  -76.75146574713393\n",
      "60  Episode Reward:  -130.1377310385896\n",
      "61  Episode Reward:  261.0951239262647\n",
      "62  Episode Reward:  241.71789914266893\n",
      "63  Episode Reward:  -89.36931178322884\n",
      "64  Episode Reward:  8.520710007316822\n",
      "65  Episode Reward:  272.44664927785163\n",
      "66  Episode Reward:  306.71902420331844\n",
      "67  Episode Reward:  -88.04969782444653\n",
      "68  Episode Reward:  255.4615117396569\n",
      "69  Episode Reward:  -15.683882354058852\n",
      "70  Episode Reward:  -101.7082510550365\n",
      "71  Episode Reward:  -79.94532436450737\n",
      "72  Episode Reward:  -150.91715722147688\n",
      "73  Episode Reward:  -33.59534159210938\n",
      "74  Episode Reward:  238.31399042899864\n",
      "75  Episode Reward:  -24.524455647633616\n",
      "76  Episode Reward:  -90.01207203980634\n",
      "77  Episode Reward:  -82.8172106982947\n",
      "78  Episode Reward:  231.99243926355416\n",
      "79  Episode Reward:  42.95735917201094\n",
      "80  Episode Reward:  -123.51469968357175\n",
      "81  Episode Reward:  -20.46295223666914\n",
      "82  Episode Reward:  -162.57898881796365\n",
      "83  Episode Reward:  -83.9792907606331\n",
      "84  Episode Reward:  -23.652122687283992\n",
      "85  Episode Reward:  -181.64333494615948\n",
      "86  Episode Reward:  -131.09669356357608\n",
      "87  Episode Reward:  -65.22287345591286\n",
      "88  Episode Reward:  196.23471156772848\n",
      "89  Episode Reward:  -41.958242142360525\n",
      "90  Episode Reward:  6.707036205667777\n",
      "91  Episode Reward:  -56.26743360679912\n",
      "92  Episode Reward:  26.760367115141378\n",
      "93  Episode Reward:  -39.8645767164961\n",
      "94  Episode Reward:  262.9295946540476\n",
      "95  Episode Reward:  -123.56017748377057\n",
      "96  Episode Reward:  7.434773739047088\n",
      "97  Episode Reward:  276.22178944401765\n",
      "98  Episode Reward:  -125.94915981001549\n",
      "99  Episode Reward:  -22.566409463972064\n"
     ]
    }
   ],
   "source": [
    "# Testing loop (Google colab cannot render the environment, only computations can be done. Render in Local Machine!)\n",
    "\n",
    "numberOfTestingEpisodes = 100\n",
    "testingRewards = []\n",
    "for episode in range (numberOfTestingEpisodes):\n",
    "    done = False\n",
    "    episodeReward = 0\n",
    "    prevObs = env.reset()\n",
    "    for step in range (1010):\n",
    "       \n",
    "        if done:\n",
    "            break\n",
    "        #env.render()\n",
    "        \n",
    "       \n",
    "\n",
    "        actions = actorModel.predict(np.expand_dims(prevObs, axis = 0))[0] # Shape (2,)\n",
    "        \n",
    "        \n",
    "        obs, reward, done, info = env.step (actions)\n",
    "        \n",
    "        prevObs = obs\n",
    "        episodeReward += reward\n",
    "      \n",
    "    print (episode, \" Episode Reward: \", episodeReward)\n",
    "    testingRewards.append(episodeReward)\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot (list(range(len(testingRewards))), testingRewards)\n",
    "plt.plot (list(range(len(testingRewards))), [np.average(testingRewards)]*len(testingRewards))\n",
    "plt.title(\"Rewards obtained during Testing\")\n",
    "plt.xlabel(\"Episode number\")\n",
    "plt.ylabel(\"Reward obtained\")\n",
    "plt.legend([\"Rewards\", \"Average: %.2f\" % np.average(testingRewards)])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The agent is unstable and even though it gets more than 200 points quite frequently, it has forgotten how to avoid crashing by flying slower. Let's check some previously saved versions to see if it had previously learnt a better policy:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0  Episode Reward:  273.30187807370305\n",
      "1  Episode Reward:  254.67306635376053\n",
      "2  Episode Reward:  280.7085340431548\n",
      "3  Episode Reward:  223.2921579611775\n",
      "4  Episode Reward:  202.75277368891375\n",
      "5  Episode Reward:  233.7991162407378\n",
      "6  Episode Reward:  256.7787144919349\n",
      "7  Episode Reward:  254.23973308235858\n",
      "8  Episode Reward:  254.41627880257835\n",
      "9  Episode Reward:  286.58973654677203\n",
      "10  Episode Reward:  261.8912812160525\n",
      "11  Episode Reward:  287.1453423954157\n",
      "12  Episode Reward:  10.230226382820504\n",
      "13  Episode Reward:  228.15683166144134\n",
      "14  Episode Reward:  218.1877652345555\n",
      "15  Episode Reward:  230.99339214509615\n",
      "16  Episode Reward:  45.845812255061816\n",
      "17  Episode Reward:  -120.1616984214564\n",
      "18  Episode Reward:  286.87189579666943\n",
      "19  Episode Reward:  279.78275689807447\n",
      "20  Episode Reward:  274.41748114262725\n",
      "21  Episode Reward:  252.23241223609563\n",
      "22  Episode Reward:  33.54671024592798\n",
      "23  Episode Reward:  288.0247372887902\n",
      "24  Episode Reward:  268.05325554366675\n",
      "25  Episode Reward:  291.22199392191044\n",
      "26  Episode Reward:  280.24921100543656\n",
      "27  Episode Reward:  271.04526780568114\n",
      "28  Episode Reward:  233.3070662103408\n",
      "29  Episode Reward:  -273.3764393035086\n",
      "30  Episode Reward:  262.87366499801647\n",
      "31  Episode Reward:  -6.935711467922019\n",
      "32  Episode Reward:  238.8628517408861\n",
      "33  Episode Reward:  251.4483288244183\n",
      "34  Episode Reward:  273.6661611239475\n",
      "35  Episode Reward:  -5.40139263833899\n",
      "36  Episode Reward:  260.6562840301284\n",
      "37  Episode Reward:  255.9987368990999\n",
      "38  Episode Reward:  276.0905188024167\n",
      "39  Episode Reward:  255.46483230618168\n",
      "40  Episode Reward:  255.05162814786578\n",
      "41  Episode Reward:  264.372693965172\n",
      "42  Episode Reward:  225.05551529904454\n",
      "43  Episode Reward:  268.0214052319234\n",
      "44  Episode Reward:  283.2391721550179\n",
      "45  Episode Reward:  247.47454142340274\n",
      "46  Episode Reward:  -200.6856182016861\n",
      "47  Episode Reward:  279.4698886312847\n",
      "48  Episode Reward:  -167.56873993242505\n",
      "49  Episode Reward:  32.643940247799804\n",
      "50  Episode Reward:  186.54588743805562\n",
      "51  Episode Reward:  206.16617233769458\n",
      "52  Episode Reward:  252.75332470921495\n",
      "53  Episode Reward:  288.93806463962613\n",
      "54  Episode Reward:  250.19603325304053\n",
      "55  Episode Reward:  277.0098835210292\n",
      "56  Episode Reward:  232.97905528619222\n",
      "57  Episode Reward:  204.3837505977903\n",
      "58  Episode Reward:  232.67648193303174\n",
      "59  Episode Reward:  208.71283894935925\n",
      "60  Episode Reward:  255.5280008691253\n",
      "61  Episode Reward:  267.39858384565196\n",
      "62  Episode Reward:  -49.100999925497405\n",
      "63  Episode Reward:  -241.872724499322\n",
      "64  Episode Reward:  184.55535734309348\n",
      "65  Episode Reward:  -9.601339519747853\n",
      "66  Episode Reward:  293.3335288994532\n",
      "67  Episode Reward:  279.9844727427654\n",
      "68  Episode Reward:  254.65818629046814\n",
      "69  Episode Reward:  279.9270310685139\n",
      "70  Episode Reward:  245.64148468052974\n",
      "71  Episode Reward:  216.56481395367103\n",
      "72  Episode Reward:  285.3656285675513\n",
      "73  Episode Reward:  209.65745194008318\n",
      "74  Episode Reward:  246.27606199560392\n",
      "75  Episode Reward:  262.02547094447215\n",
      "76  Episode Reward:  -103.7243264568451\n",
      "77  Episode Reward:  -96.83892881993698\n",
      "78  Episode Reward:  255.12690688887798\n",
      "79  Episode Reward:  238.56086148414\n",
      "80  Episode Reward:  -288.9769217441428\n",
      "81  Episode Reward:  252.5524452252169\n",
      "82  Episode Reward:  255.9731903657367\n",
      "83  Episode Reward:  30.84273695652726\n",
      "84  Episode Reward:  250.50189274853824\n",
      "85  Episode Reward:  188.17553078928432\n",
      "86  Episode Reward:  -7.415677440909619\n",
      "87  Episode Reward:  214.09702074669934\n",
      "88  Episode Reward:  285.65009890959703\n",
      "89  Episode Reward:  279.9700865732453\n",
      "90  Episode Reward:  273.3817842642897\n",
      "91  Episode Reward:  270.5883271174532\n",
      "92  Episode Reward:  245.21558065837934\n",
      "93  Episode Reward:  252.41938484730858\n",
      "94  Episode Reward:  222.91212610161568\n",
      "95  Episode Reward:  196.47976976829753\n",
      "96  Episode Reward:  -67.56568384861349\n",
      "97  Episode Reward:  -1.3927177180548966\n",
      "98  Episode Reward:  263.43725569780025\n",
      "99  Episode Reward:  276.62955499828945\n"
     ]
    }
   ],
   "source": [
    "# Testing loop (Google colab cannot render the environment, only computations can be done. Render in Local Machine!)\n",
    "\n",
    "actorModel.load_weights(\"lunarLanderContinuous750episodes.h5\")\n",
    "numberOfTestingEpisodes = 100\n",
    "testingRewards = []\n",
    "for episode in range (numberOfTestingEpisodes):\n",
    "    done = False\n",
    "    episodeReward = 0\n",
    "    prevObs = env.reset()\n",
    "    for step in range (1010):\n",
    "       \n",
    "        if done:\n",
    "            break\n",
    "        env.render()\n",
    "        \n",
    "       \n",
    "\n",
    "        actions = actorModel.predict(np.expand_dims(prevObs, axis = 0))[0] # Shape (2,)\n",
    "        \n",
    "        \n",
    "        obs, reward, done, info = env.step (actions)\n",
    "        \n",
    "        prevObs = obs\n",
    "        episodeReward += reward\n",
    "      \n",
    "    print (episode, \" Episode Reward: \", episodeReward)\n",
    "    testingRewards.append(episodeReward)\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot (list(range(len(testingRewards))), testingRewards)\n",
    "plt.plot (list(range(len(testingRewards))), [np.average(testingRewards)]*len(testingRewards))\n",
    "plt.title(\"Rewards obtained during Testing\")\n",
    "plt.xlabel(\"Episode number\")\n",
    "plt.ylabel(\"Reward obtained\")\n",
    "plt.legend([\"Rewards\", \"Average: %.2f\" % np.average(testingRewards)])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0  Episode Reward:  245.24740378292339\n",
      "1  Episode Reward:  245.87810222551025\n",
      "2  Episode Reward:  223.15829970093694\n",
      "3  Episode Reward:  288.9930288933916\n",
      "4  Episode Reward:  263.3465753167518\n",
      "5  Episode Reward:  232.61219742090373\n",
      "6  Episode Reward:  244.62361292725333\n",
      "7  Episode Reward:  285.2174105880822\n",
      "8  Episode Reward:  232.10638155301945\n",
      "9  Episode Reward:  295.82746209231664\n",
      "10  Episode Reward:  263.6805047395683\n",
      "11  Episode Reward:  271.52296585203993\n",
      "12  Episode Reward:  -8.0663093325293\n",
      "13  Episode Reward:  280.69305573932047\n",
      "14  Episode Reward:  233.35048552375486\n",
      "15  Episode Reward:  -43.294733190986356\n",
      "16  Episode Reward:  273.1560924084511\n",
      "17  Episode Reward:  254.80690882043453\n",
      "18  Episode Reward:  -17.512308004339346\n",
      "19  Episode Reward:  202.76931849970447\n",
      "20  Episode Reward:  22.549821086133065\n",
      "21  Episode Reward:  275.29508191912646\n",
      "22  Episode Reward:  278.33426747186843\n",
      "23  Episode Reward:  275.1298326752816\n",
      "24  Episode Reward:  157.21157804186953\n",
      "25  Episode Reward:  -19.173961308345667\n",
      "26  Episode Reward:  -31.113214716756247\n",
      "27  Episode Reward:  261.7671554850557\n",
      "28  Episode Reward:  286.78699248430325\n",
      "29  Episode Reward:  -76.14557049225533\n",
      "30  Episode Reward:  244.5670773890865\n",
      "31  Episode Reward:  266.63948580908664\n",
      "32  Episode Reward:  272.12335460471377\n",
      "33  Episode Reward:  238.02566677889115\n",
      "34  Episode Reward:  11.905749774533135\n",
      "35  Episode Reward:  289.9328629814876\n",
      "36  Episode Reward:  244.44323740219517\n",
      "37  Episode Reward:  268.4704152870089\n",
      "38  Episode Reward:  291.2822623422829\n",
      "39  Episode Reward:  250.6596111677957\n",
      "40  Episode Reward:  282.39531257822637\n",
      "41  Episode Reward:  228.79497211494817\n",
      "42  Episode Reward:  291.63673918281114\n",
      "43  Episode Reward:  277.00906793775005\n",
      "44  Episode Reward:  276.7406858847097\n",
      "45  Episode Reward:  277.4300558157113\n",
      "46  Episode Reward:  26.994697154555865\n",
      "47  Episode Reward:  271.5928532643718\n",
      "48  Episode Reward:  184.73410122322935\n",
      "49  Episode Reward:  288.7496326886386\n",
      "50  Episode Reward:  235.02426864792304\n",
      "51  Episode Reward:  20.53569879429699\n",
      "52  Episode Reward:  306.4342188296966\n",
      "53  Episode Reward:  226.3275106760707\n",
      "54  Episode Reward:  64.64277793865452\n",
      "55  Episode Reward:  221.76982565492938\n",
      "56  Episode Reward:  -115.67277358561547\n",
      "57  Episode Reward:  274.6769353278444\n",
      "58  Episode Reward:  252.06791708949828\n",
      "59  Episode Reward:  -130.16587998079473\n",
      "60  Episode Reward:  214.91904344533557\n",
      "61  Episode Reward:  262.1821698670818\n",
      "62  Episode Reward:  260.33819201460994\n",
      "63  Episode Reward:  172.1120309452398\n",
      "64  Episode Reward:  282.69033509534495\n",
      "65  Episode Reward:  217.6522842596391\n",
      "66  Episode Reward:  279.9182370587026\n",
      "67  Episode Reward:  254.5991525578549\n",
      "68  Episode Reward:  253.01479568411864\n",
      "69  Episode Reward:  -120.33468594574578\n",
      "70  Episode Reward:  238.74482161585283\n",
      "71  Episode Reward:  206.01943057796723\n",
      "72  Episode Reward:  -192.6570562597642\n",
      "73  Episode Reward:  277.597935944334\n",
      "74  Episode Reward:  294.62541454745394\n",
      "75  Episode Reward:  -26.535593468765242\n",
      "76  Episode Reward:  292.88353674475786\n",
      "77  Episode Reward:  244.57935759632414\n",
      "78  Episode Reward:  41.62130882185474\n",
      "79  Episode Reward:  6.697057369247489\n",
      "80  Episode Reward:  252.77941946252017\n",
      "81  Episode Reward:  201.48067472455847\n",
      "82  Episode Reward:  288.04146303802213\n",
      "83  Episode Reward:  -199.35282458735335\n",
      "84  Episode Reward:  281.46486596450376\n",
      "85  Episode Reward:  -389.25995891891625\n",
      "86  Episode Reward:  59.875153290381746\n",
      "87  Episode Reward:  243.422836136603\n",
      "88  Episode Reward:  243.31194377624865\n",
      "89  Episode Reward:  270.22589420787364\n",
      "90  Episode Reward:  -73.3820388453157\n",
      "91  Episode Reward:  262.47771171717847\n",
      "92  Episode Reward:  261.5066836179252\n",
      "93  Episode Reward:  255.12215169015414\n",
      "94  Episode Reward:  298.84654127199644\n",
      "95  Episode Reward:  272.33204707991354\n",
      "96  Episode Reward:  22.070041232750256\n",
      "97  Episode Reward:  244.72586062914735\n",
      "98  Episode Reward:  247.32977085917315\n",
      "99  Episode Reward:  271.6432369495403\n"
     ]
    }
   ],
   "source": [
    "# Testing loop (Google colab cannot render the environment, only computations can be done. Render in Local Machine!)\n",
    "\n",
    "actorModel.load_weights(\"lunarLanderContinuous700episodes.h5\")\n",
    "numberOfTestingEpisodes = 100\n",
    "testingRewards = []\n",
    "for episode in range (numberOfTestingEpisodes):\n",
    "    done = False\n",
    "    episodeReward = 0\n",
    "    prevObs = env.reset()\n",
    "    for step in range (1010):\n",
    "       \n",
    "        if done:\n",
    "            break\n",
    "        env.render()\n",
    "        \n",
    "       \n",
    "\n",
    "        actions = actorModel.predict(np.expand_dims(prevObs, axis = 0))[0] # Shape (2,)\n",
    "        \n",
    "        \n",
    "        obs, reward, done, info = env.step (actions)\n",
    "        \n",
    "        prevObs = obs\n",
    "        episodeReward += reward\n",
    "      \n",
    "    print (episode, \" Episode Reward: \", episodeReward)\n",
    "    testingRewards.append(episodeReward)\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot (list(range(len(testingRewards))), testingRewards)\n",
    "plt.plot (list(range(len(testingRewards))), [np.average(testingRewards)]*len(testingRewards))\n",
    "plt.title(\"Rewards obtained during Testing\")\n",
    "plt.xlabel(\"Episode number\")\n",
    "plt.ylabel(\"Reward obtained\")\n",
    "plt.legend([\"Rewards\", \"Average: %.2f\" % np.average(testingRewards)])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Both these versions achieve an average really close to 200.0 points over 100 consecutive episodes and therefore, this environment can be considered solved!"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "LunarLanderContinuous.ipynb",
   "provenance": []
  },
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}