RL_BipedalWalker.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reinforcement Learning for Bipedal Walker game"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import libraries:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python36\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import gym\n",
    "import matplotlib.pyplot as plt\n",
    "% matplotlib notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Defining hyperparameters:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we define all the hyperparameters for the model, this will allow as to easily iterate and find the values that give best results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n",
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "num_episodes = 500      # Number of episodes for training\n",
    "learning_rate = 0.01\n",
    "gamma = 0.99           # Discount factor for reward\n",
    "num_Hidden = 200       # number of nodes in the hidden layer\n",
    "\n",
    "env = gym.make('BipedalWalker-v2')    # Choose a game and create an environment\n",
    "env._max_episode_steps = 5000\n",
    "# env = env.unwrapped              # The wrapper limits the number of steps in an episode, let's get rid of it\n",
    "obs_dim = env.reset().shape      # obervation dimension\n",
    "num_actions = len(env.action_space.high) # number of actions (this works only for continuous action space, which is the case here)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define functions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Discount reward function**: We we'll train our agent based on the reward gained for his actions. For each action we'll define the episode reward as the total reward gained in all the next steps of the current episode. Since a reward gained further away in the future has less correlation to the present action, we will give it less weight by discounting future rewards.\n",
    "\n",
    "The formula for the discounted rewards is given by:\n",
    "\n",
    "$$ R_t = \\sum_k \\gamma^k r_{t+k} $$\n",
    "\n",
    "Where $r_t$ is the reward gained in the step $t$ and $\\gamma \\in [0,1]$ is a hyperparameter called the discount factor.\n",
    "Here we define a function that takes a vector of rewards in consequent steps and returns the discounted reward vector:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def discount_rewards(r):\n",
    "    '''Takes a 1D rewards (one episode) and discounts it and also standardize\n",
    "    the rewards to be unit normal (helps control the gradient estimator variance)'''\n",
    "    \n",
    "    # Discounting\n",
    "    dis_r = np.zeros_like(r)\n",
    "    running_sum = 0\n",
    "    for t in reversed(range(len(r))):\n",
    "        running_sum = gamma*running_sum + r[t]\n",
    "        dis_r[t] = running_sum\n",
    "    \n",
    "    # Normailzing\n",
    "    dis_r = (dis_r - np.mean(dis_r))/np.std(dis_r)\n",
    "        \n",
    "    return dis_r"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This game takes a continuous range of values for each action. In the Gym library language, the action space is a Box instead of Discrete"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Box(4,)\n",
      "[1. 1. 1. 1.]\n",
      "[-1. -1. -1. -1.]\n"
     ]
    }
   ],
   "source": [
    "print(env.action_space)\n",
    "print(env.action_space.high)\n",
    "print(env.action_space.low)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case, each action is a vector of length 4 where every element value is in the [-1,1] domain. We will use Gaussian policy to train this model. Our NN would output the 4 mean values and 4 std values for each element of the action vector, and then use them to generate actions in each step."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since the mean values must be contained in the region [-1,1], we would like to use some activation function that reflects that. We'll use a custom activation which is basically a sigmoid scaled to our desired domain: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def m1t1_sigmoid(x):\n",
    "    return (tf.nn.sigmoid(x) - 0.5)*2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Build model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To keep order in our model, we use name scopes which basically group the layers of our model in a simple to follow way. Eventually, when using [TensorBoard](https://www.tensorflow.org/guide/summaries_and_tensorboard) for visualizing the model, the graph is more readable and makes it easy to understand the model and find errors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Create placeholders for inputs'''\n",
    "# A place holder for input observations\n",
    "input_ = tf.placeholder(tf.float32, shape = (None, obs_dim[0]), name = \"input\")\n",
    "# A place holder for actions in a full episode\n",
    "actions = tf.placeholder(tf.float32, shape = (None, num_actions), name = \"actions\")\n",
    "# A place holder for discounted rewards in a full episode\n",
    "dis_rewards = tf.placeholder(tf.float32, shape = (None, ), name = \"dis_rewards\")\n",
    "\n",
    "'''Fully connected layers'''\n",
    "# A network channel for predicting the mean values of the action PDF\n",
    "with tf.name_scope(\"Mean_channel\"):\n",
    "    Mfc1 = tf.layers.dense(inputs = input_, units = num_Hidden ,activation = tf.nn.relu, name = \"Mfc1\" )\n",
    "    Mfc2 = tf.layers.dense(inputs = Mfc1, units = num_actions ,activation = m1t1_sigmoid, name = \"Mfc2\")\n",
    "\n",
    "    # A network channel for predicting the std values of the action PDF\n",
    "with tf.name_scope(\"Std_channel\"):\n",
    "    Sfc1 = tf.layers.dense(inputs = input_, units = num_Hidden ,activation = tf.nn.relu, name = \"Sfc1\" )\n",
    "    # We use a softplus activation for the last layer to ensure positive std value\n",
    "    Sfc2 = tf.layers.dense(inputs = input_, units = num_actions ,activation = tf.nn.sigmoid, \n",
    "                           name = \"Sfc2\" )\n",
    "    \n",
    "    \n",
    "with tf.name_scope(\"Action_PDF\"):\n",
    "    # Use the predicted mean and std for creating a probabilty distribution for actions\n",
    "    action_prob_dist = tf.contrib.distributions.Normal(loc = Mfc2, scale = Sfc2 + 1e-5, \n",
    "                                                       name = \"Normal_PDF\")\n",
    "    sample_action = action_prob_dist.sample([1], name = \"sample_action\")\n",
    "\n",
    "with tf.name_scope(\"Loss\"):\n",
    "    '''Define loss'''\n",
    "    # compute negative log probability using our PDF and sum over the actions in every time step\n",
    "    negative_logprob = - tf.reduce_sum(action_prob_dist.log_prob(actions), axis = 1)\n",
    "    # Modulate the loss based on our discounted reward - this is where reinforcment learning happens,\n",
    "    # we favor actions that produced high reward\n",
    "    loss = tf.reduce_mean(negative_logprob * dis_rewards, name = \"loss\")\n",
    "\n",
    "with tf.name_scope(\"Training\"):\n",
    "    '''Define optimizer'''\n",
    "    training_opt = tf.train.AdamOptimizer(learning_rate).minimize(loss)\n",
    "\n",
    "'''Define saver for saving and restoring model'''\n",
    "saver = tf.train.Saver()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define a writer for saving summaries to TensorBoard:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# writer = tf.summary.FileWriter(\"./tensorboard/1/\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have a model ready, we can start training it.\n",
    "\n",
    "Our goal is to achieve a model that uses the current observation to create the best probability distribution function (PDF) for the next action to be taken. This is **not a deterministic model** - the agent has a fine probability to take any action after receiving an observation, with larger probability to take favorable actions.\n",
    "\n",
    "To achieve that, in each step, we use our model to generate a PDF of actions, draw an action out of it and take the next step.\n",
    "When we reach the end of the episode, we compute the episode loss by feeding the observation, action and discounted rewards vectors (which we kept track of). Our optimizer minimizes the loss which makes favorable (unfavorable) actions more probable (less probable) due to the multiplicative factor of the discounted reward. By repeating these steps to each episode, our agent gradually improves its PDF and becomes a better player!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 0\n",
      "Episode ended after 104 steps\n",
      "Accumulated reward in this episode -103.63\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 1\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -185.01\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 2\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -226.28\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 3\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -228.93\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 4\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -220.76\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 5\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -247.88\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 6\n",
      "Episode ended after 64 steps\n",
      "Accumulated reward in this episode -119.74\n",
      "Mean of x velocity in this episode -0.02\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 7\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -281.90\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 8\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -297.82\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 9\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -319.02\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 10\n",
      "Episode ended after 77 steps\n",
      "Accumulated reward in this episode -105.96\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -103.63\n",
      "-------------------------------------------------\n",
      "Episode 11\n",
      "Episode ended after 50 steps\n",
      "Accumulated reward in this episode -102.82\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -102.82\n",
      "-------------------------------------------------\n",
      "Episode 12\n",
      "Episode ended after 57 steps\n",
      "Accumulated reward in this episode -105.07\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -102.82\n",
      "-------------------------------------------------\n",
      "Episode 13\n",
      "Episode ended after 71 steps\n",
      "Accumulated reward in this episode -100.42\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 14\n",
      "Episode ended after 76 steps\n",
      "Accumulated reward in this episode -102.11\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 15\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -106.23\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 16\n",
      "Episode ended after 65 steps\n",
      "Accumulated reward in this episode -102.29\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 17\n",
      "Episode ended after 82 steps\n",
      "Accumulated reward in this episode -103.30\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 18\n",
      "Episode ended after 50 steps\n",
      "Accumulated reward in this episode -104.41\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 19\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -104.07\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 20\n",
      "Episode ended after 69 steps\n",
      "Accumulated reward in this episode -102.96\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 21\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -103.69\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 22\n",
      "Episode ended after 93 steps\n",
      "Accumulated reward in this episode -106.16\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 23\n",
      "Episode ended after 70 steps\n",
      "Accumulated reward in this episode -102.32\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 24\n",
      "Episode ended after 95 steps\n",
      "Accumulated reward in this episode -104.36\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 25\n",
      "Episode ended after 77 steps\n",
      "Accumulated reward in this episode -104.31\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 26\n",
      "Episode ended after 114 steps\n",
      "Accumulated reward in this episode -106.08\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 27\n",
      "Episode ended after 66 steps\n",
      "Accumulated reward in this episode -101.63\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 28\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -101.27\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 29\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -238.17\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 30\n",
      "Episode ended after 116 steps\n",
      "Accumulated reward in this episode -129.48\n",
      "Mean of x velocity in this episode -0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 31\n",
      "Episode ended after 81 steps\n",
      "Accumulated reward in this episode -104.34\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 32\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -221.71\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 33\n",
      "Episode ended after 82 steps\n",
      "Accumulated reward in this episode -123.12\n",
      "Mean of x velocity in this episode -0.07\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 34\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -119.39\n",
      "Mean of x velocity in this episode -0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 35\n",
      "Episode ended after 91 steps\n",
      "Accumulated reward in this episode -121.18\n",
      "Mean of x velocity in this episode -0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 36\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -233.23\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 37\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -241.03\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 38\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -239.87\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 39\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -242.31\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 40\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -100.99\n",
      "Mean of x velocity in this episode 0.15\n",
      "Maximal reward so far -100.42\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 41\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -250.40\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 42\n",
      "Episode ended after 86 steps\n",
      "Accumulated reward in this episode -104.00\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 43\n",
      "Episode ended after 54 steps\n",
      "Accumulated reward in this episode -101.43\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 44\n",
      "Episode ended after 73 steps\n",
      "Accumulated reward in this episode -101.78\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 45\n",
      "Episode ended after 108 steps\n",
      "Accumulated reward in this episode -100.70\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 46\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -103.52\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 47\n",
      "Episode ended after 75 steps\n",
      "Accumulated reward in this episode -102.23\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 48\n",
      "Episode ended after 80 steps\n",
      "Accumulated reward in this episode -121.04\n",
      "Mean of x velocity in this episode -0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 49\n",
      "Episode ended after 112 steps\n",
      "Accumulated reward in this episode -129.43\n",
      "Mean of x velocity in this episode -0.13\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 50\n",
      "Episode ended after 51 steps\n",
      "Accumulated reward in this episode -105.59\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 51\n",
      "Episode ended after 103 steps\n",
      "Accumulated reward in this episode -123.69\n",
      "Mean of x velocity in this episode -0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 52\n",
      "Episode ended after 62 steps\n",
      "Accumulated reward in this episode -103.30\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 53\n",
      "Episode ended after 72 steps\n",
      "Accumulated reward in this episode -100.74\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 54\n",
      "Episode ended after 63 steps\n",
      "Accumulated reward in this episode -103.42\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 55\n",
      "Episode ended after 63 steps\n",
      "Accumulated reward in this episode -102.15\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 56\n",
      "Episode ended after 52 steps\n",
      "Accumulated reward in this episode -102.93\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 57\n",
      "Episode ended after 154 steps\n",
      "Accumulated reward in this episode -130.85\n",
      "Mean of x velocity in this episode -0.10\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 58\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -236.04\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 59\n",
      "Episode ended after 53 steps\n",
      "Accumulated reward in this episode -102.31\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 60\n",
      "Episode ended after 150 steps\n",
      "Accumulated reward in this episode -132.08\n",
      "Mean of x velocity in this episode -0.13\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 61\n",
      "Episode ended after 62 steps\n",
      "Accumulated reward in this episode -101.78\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 62\n",
      "Episode ended after 145 steps\n",
      "Accumulated reward in this episode -106.60\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 63\n",
      "Episode ended after 51 steps\n",
      "Accumulated reward in this episode -103.34\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -100.42\n",
      "-------------------------------------------------\n",
      "Episode 64\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -99.44\n",
      "Mean of x velocity in this episode 0.17\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 65\n",
      "Episode ended after 48 steps\n",
      "Accumulated reward in this episode -102.59\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 66\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -102.81\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 67\n",
      "Episode ended after 67 steps\n",
      "Accumulated reward in this episode -99.78\n",
      "Mean of x velocity in this episode 0.15\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 68\n",
      "Episode ended after 54 steps\n",
      "Accumulated reward in this episode -102.73\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 69\n",
      "Episode ended after 64 steps\n",
      "Accumulated reward in this episode -104.04\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 70\n",
      "Episode ended after 96 steps\n",
      "Accumulated reward in this episode -102.72\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 71\n",
      "Episode ended after 56 steps\n",
      "Accumulated reward in this episode -101.10\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 72\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -244.41\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 73\n",
      "Episode ended after 78 steps\n",
      "Accumulated reward in this episode -102.44\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 74\n",
      "Episode ended after 64 steps\n",
      "Accumulated reward in this episode -101.08\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 75\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -347.13\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 76\n",
      "Episode ended after 52 steps\n",
      "Accumulated reward in this episode -101.82\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 77\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -105.09\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 78\n",
      "Episode ended after 85 steps\n",
      "Accumulated reward in this episode -103.17\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 79\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -117.16\n",
      "Mean of x velocity in this episode -0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 80\n",
      "Episode ended after 79 steps\n",
      "Accumulated reward in this episode -108.51\n",
      "Mean of x velocity in this episode -0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 81\n",
      "Episode ended after 74 steps\n",
      "Accumulated reward in this episode -100.83\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 82\n",
      "Episode ended after 113 steps\n",
      "Accumulated reward in this episode -106.04\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 83\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -105.62\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 84\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -295.79\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 85\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -107.05\n",
      "Mean of x velocity in this episode -0.04\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 86\n",
      "Episode ended after 149 steps\n",
      "Accumulated reward in this episode -120.53\n",
      "Mean of x velocity in this episode -0.08\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 87\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -210.65\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 88\n",
      "Episode ended after 98 steps\n",
      "Accumulated reward in this episode -119.80\n",
      "Mean of x velocity in this episode -0.15\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 89\n",
      "Episode ended after 96 steps\n",
      "Accumulated reward in this episode -107.98\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 90\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -172.50\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 91\n",
      "Episode ended after 120 steps\n",
      "Accumulated reward in this episode -117.21\n",
      "Mean of x velocity in this episode -0.10\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 92\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -227.12\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 93\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -214.38\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 94\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -221.50\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 95\n",
      "Episode ended after 86 steps\n",
      "Accumulated reward in this episode -119.99\n",
      "Mean of x velocity in this episode -0.17\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 96\n",
      "Episode ended after 77 steps\n",
      "Accumulated reward in this episode -111.56\n",
      "Mean of x velocity in this episode -0.08\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 97\n",
      "Episode ended after 53 steps\n",
      "Accumulated reward in this episode -110.32\n",
      "Mean of x velocity in this episode -0.07\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 98\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -192.00\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 99\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -190.83\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 100\n",
      "Episode ended after 126 steps\n",
      "Accumulated reward in this episode -102.38\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 101\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -191.95\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 102\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -199.91\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 103\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -193.94\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 104\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -206.17\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 105\n",
      "Episode ended after 68 steps\n",
      "Accumulated reward in this episode -106.33\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 106\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -223.22\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 107\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -241.93\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 108\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -246.12\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 109\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -250.70\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 110\n",
      "Episode ended after 64 steps\n",
      "Accumulated reward in this episode -104.43\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 111\n",
      "Episode ended after 81 steps\n",
      "Accumulated reward in this episode -101.75\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 112\n",
      "Episode ended after 63 steps\n",
      "Accumulated reward in this episode -104.57\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 113\n",
      "Episode ended after 81 steps\n",
      "Accumulated reward in this episode -101.25\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 114\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -213.57\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 115\n",
      "Episode ended after 107 steps\n",
      "Accumulated reward in this episode -122.25\n",
      "Mean of x velocity in this episode -0.07\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 116\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -215.17\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 117\n",
      "Episode ended after 101 steps\n",
      "Accumulated reward in this episode -103.92\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 118\n",
      "Episode ended after 176 steps\n",
      "Accumulated reward in this episode -107.04\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 119\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -196.47\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 120\n",
      "Episode ended after 56 steps\n",
      "Accumulated reward in this episode -103.14\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 121\n",
      "Episode ended after 220 steps\n",
      "Accumulated reward in this episode -130.92\n",
      "Mean of x velocity in this episode -0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 122\n",
      "Episode ended after 82 steps\n",
      "Accumulated reward in this episode -101.03\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 123\n",
      "Episode ended after 105 steps\n",
      "Accumulated reward in this episode -124.12\n",
      "Mean of x velocity in this episode -0.11\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 124\n",
      "Episode ended after 67 steps\n",
      "Accumulated reward in this episode -101.81\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 125\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -225.05\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 126\n",
      "Episode ended after 86 steps\n",
      "Accumulated reward in this episode -118.06\n",
      "Mean of x velocity in this episode -0.04\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 127\n",
      "Episode ended after 71 steps\n",
      "Accumulated reward in this episode -101.24\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 128\n",
      "Episode ended after 81 steps\n",
      "Accumulated reward in this episode -123.86\n",
      "Mean of x velocity in this episode -0.14\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 129\n",
      "Episode ended after 66 steps\n",
      "Accumulated reward in this episode -104.60\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 130\n",
      "Episode ended after 75 steps\n",
      "Accumulated reward in this episode -103.48\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 131\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -209.97\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 132\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -217.04\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 133\n",
      "Episode ended after 68 steps\n",
      "Accumulated reward in this episode -104.13\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 134\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -229.71\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 135\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -202.32\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 136\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -196.35\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 137\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -204.27\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 138\n",
      "Episode ended after 109 steps\n",
      "Accumulated reward in this episode -106.61\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 139\n",
      "Episode ended after 91 steps\n",
      "Accumulated reward in this episode -121.70\n",
      "Mean of x velocity in this episode -0.11\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 140\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -143.29\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 141\n",
      "Episode ended after 89 steps\n",
      "Accumulated reward in this episode -121.75\n",
      "Mean of x velocity in this episode -0.16\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 142\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -115.96\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 143\n",
      "Episode ended after 85 steps\n",
      "Accumulated reward in this episode -117.92\n",
      "Mean of x velocity in this episode -0.06\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 144\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -132.13\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 145\n",
      "Episode ended after 1151 steps\n",
      "Accumulated reward in this episode -123.29\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 146\n",
      "Episode ended after 4618 steps\n",
      "Accumulated reward in this episode -221.61\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 147\n",
      "Episode ended after 134 steps\n",
      "Accumulated reward in this episode -125.13\n",
      "Mean of x velocity in this episode -0.07\n",
      "Maximal reward so far -99.44\n",
      "-------------------------------------------------\n",
      "Episode 148\n",
      "Episode ended after 66 steps\n",
      "Accumulated reward in this episode -98.87\n",
      "Mean of x velocity in this episode 0.17\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 149\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -101.68\n",
      "Mean of x velocity in this episode 0.16\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 150\n",
      "Episode ended after 110 steps\n",
      "Accumulated reward in this episode -99.29\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 151\n",
      "Episode ended after 1299 steps\n",
      "Accumulated reward in this episode -125.40\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 152\n",
      "Episode ended after 57 steps\n",
      "Accumulated reward in this episode -101.33\n",
      "Mean of x velocity in this episode 0.18\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 153\n",
      "Episode ended after 100 steps\n",
      "Accumulated reward in this episode -102.16\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 154\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -100.97\n",
      "Mean of x velocity in this episode 0.16\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 155\n",
      "Episode ended after 65 steps\n",
      "Accumulated reward in this episode -103.40\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 156\n",
      "Episode ended after 57 steps\n",
      "Accumulated reward in this episode -100.29\n",
      "Mean of x velocity in this episode 0.21\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 157\n",
      "Episode ended after 64 steps\n",
      "Accumulated reward in this episode -105.73\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 158\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -101.48\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -98.87\n",
      "-------------------------------------------------\n",
      "Episode 159\n",
      "Episode ended after 73 steps\n",
      "Accumulated reward in this episode -98.40\n",
      "Mean of x velocity in this episode 0.22\n",
      "Maximal reward so far -98.40\n",
      "-------------------------------------------------\n",
      "Episode 160\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -105.36\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -98.40\n",
      "-------------------------------------------------\n",
      "Episode 161\n",
      "Episode ended after 57 steps\n",
      "Accumulated reward in this episode -101.48\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -98.40\n",
      "-------------------------------------------------\n",
      "Episode 162\n",
      "Episode ended after 61 steps\n",
      "Accumulated reward in this episode -97.54\n",
      "Mean of x velocity in this episode 0.25\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 163\n",
      "Episode ended after 62 steps\n",
      "Accumulated reward in this episode -101.15\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 164\n",
      "Episode ended after 71 steps\n",
      "Accumulated reward in this episode -101.72\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 165\n",
      "Episode ended after 78 steps\n",
      "Accumulated reward in this episode -102.63\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 166\n",
      "Episode ended after 98 steps\n",
      "Accumulated reward in this episode -103.75\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 167\n",
      "Episode ended after 47 steps\n",
      "Accumulated reward in this episode -105.20\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 168\n",
      "Episode ended after 113 steps\n",
      "Accumulated reward in this episode -126.62\n",
      "Mean of x velocity in this episode -0.11\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 169\n",
      "Episode ended after 70 steps\n",
      "Accumulated reward in this episode -119.75\n",
      "Mean of x velocity in this episode -0.11\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 170\n",
      "Episode ended after 79 steps\n",
      "Accumulated reward in this episode -101.78\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 171\n",
      "Episode ended after 84 steps\n",
      "Accumulated reward in this episode -104.01\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 172\n",
      "Episode ended after 83 steps\n",
      "Accumulated reward in this episode -98.42\n",
      "Mean of x velocity in this episode 0.17\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 173\n",
      "Episode ended after 83 steps\n",
      "Accumulated reward in this episode -123.27\n",
      "Mean of x velocity in this episode -0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 174\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -141.33\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 175\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -159.61\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 176\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -140.30\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 177\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -163.68\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 178\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -186.69\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 179\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -207.96\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 180\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -199.95\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 181\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -198.78\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 182\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -205.02\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 183\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -194.52\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 184\n",
      "Episode ended after 3000 steps\n",
      "Accumulated reward in this episode -207.00\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 185\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -196.80\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 186\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -202.05\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 187\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -201.71\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 188\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -193.12\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 189\n",
      "Episode ended after 113 steps\n",
      "Accumulated reward in this episode -105.43\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 190\n",
      "Episode ended after 2300 steps\n",
      "Accumulated reward in this episode -184.59\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 191\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -114.61\n",
      "Mean of x velocity in this episode -0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 192\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -197.57\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 193\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -211.84\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 194\n",
      "Episode ended after 91 steps\n",
      "Accumulated reward in this episode -103.31\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 195\n",
      "Episode ended after 3553 steps\n",
      "Accumulated reward in this episode -243.76\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 196\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -204.70\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 197\n",
      "Episode ended after 227 steps\n",
      "Accumulated reward in this episode -108.33\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 198\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -213.04\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 199\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -104.34\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 200\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -216.21\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 201\n",
      "Episode ended after 51 steps\n",
      "Accumulated reward in this episode -102.06\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 202\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -219.00\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 203\n",
      "Episode ended after 53 steps\n",
      "Accumulated reward in this episode -103.20\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 204\n",
      "Episode ended after 55 steps\n",
      "Accumulated reward in this episode -103.76\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 205\n",
      "Episode ended after 49 steps\n",
      "Accumulated reward in this episode -102.14\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 206\n",
      "Episode ended after 65 steps\n",
      "Accumulated reward in this episode -103.50\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 207\n",
      "Episode ended after 67 steps\n",
      "Accumulated reward in this episode -101.25\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -97.54\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 208\n",
      "Episode ended after 53 steps\n",
      "Accumulated reward in this episode -105.93\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 209\n",
      "Episode ended after 91 steps\n",
      "Accumulated reward in this episode -104.00\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 210\n",
      "Episode ended after 116 steps\n",
      "Accumulated reward in this episode -106.64\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 211\n",
      "Episode ended after 86 steps\n",
      "Accumulated reward in this episode -103.09\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 212\n",
      "Episode ended after 72 steps\n",
      "Accumulated reward in this episode -101.76\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 213\n",
      "Episode ended after 130 steps\n",
      "Accumulated reward in this episode -119.64\n",
      "Mean of x velocity in this episode -0.03\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 214\n",
      "Episode ended after 115 steps\n",
      "Accumulated reward in this episode -99.87\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 215\n",
      "Episode ended after 118 steps\n",
      "Accumulated reward in this episode -101.16\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 216\n",
      "Episode ended after 59 steps\n",
      "Accumulated reward in this episode -104.70\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 217\n",
      "Episode ended after 65 steps\n",
      "Accumulated reward in this episode -101.18\n",
      "Mean of x velocity in this episode 0.17\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 218\n",
      "Episode ended after 110 steps\n",
      "Accumulated reward in this episode -101.25\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 219\n",
      "Episode ended after 54 steps\n",
      "Accumulated reward in this episode -102.94\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 220\n",
      "Episode ended after 68 steps\n",
      "Accumulated reward in this episode -104.28\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 221\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -122.44\n",
      "Mean of x velocity in this episode -0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 222\n",
      "Episode ended after 84 steps\n",
      "Accumulated reward in this episode -104.36\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 223\n",
      "Episode ended after 71 steps\n",
      "Accumulated reward in this episode -99.90\n",
      "Mean of x velocity in this episode 0.15\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 224\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -102.17\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 225\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -270.83\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 226\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -273.56\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 227\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -270.12\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 228\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -267.87\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 229\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -264.10\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 230\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -258.44\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 231\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -254.04\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 232\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -245.24\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 233\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -243.21\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 234\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -239.64\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 235\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -221.87\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 236\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -104.27\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 237\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -212.97\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 238\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -213.34\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 239\n",
      "Episode ended after 201 steps\n",
      "Accumulated reward in this episode -128.19\n",
      "Mean of x velocity in this episode -0.05\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 240\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -189.19\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 241\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -175.12\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 242\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -171.23\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 243\n",
      "Episode ended after 117 steps\n",
      "Accumulated reward in this episode -118.34\n",
      "Mean of x velocity in this episode -0.04\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 244\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -166.72\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 245\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -165.84\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 246\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -158.37\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 247\n",
      "Episode ended after 84 steps\n",
      "Accumulated reward in this episode -121.63\n",
      "Mean of x velocity in this episode -0.13\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 248\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -154.00\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 249\n",
      "Episode ended after 58 steps\n",
      "Accumulated reward in this episode -104.33\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 250\n",
      "Episode ended after 102 steps\n",
      "Accumulated reward in this episode -102.18\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -97.54\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 251\n",
      "Episode ended after 63 steps\n",
      "Accumulated reward in this episode -103.46\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 252\n",
      "Episode ended after 218 steps\n",
      "Accumulated reward in this episode -139.03\n",
      "Mean of x velocity in this episode -0.10\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 253\n",
      "Episode ended after 120 steps\n",
      "Accumulated reward in this episode -99.87\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 254\n",
      "Episode ended after 48 steps\n",
      "Accumulated reward in this episode -104.06\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 255\n",
      "Episode ended after 2470 steps\n",
      "Accumulated reward in this episode -255.61\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -97.54\n",
      "-------------------------------------------------\n",
      "Episode 256\n",
      "Episode ended after 92 steps\n",
      "Accumulated reward in this episode -94.08\n",
      "Mean of x velocity in this episode 0.24\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 257\n",
      "Episode ended after 2065 steps\n",
      "Accumulated reward in this episode -244.64\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 258\n",
      "Episode ended after 1895 steps\n",
      "Accumulated reward in this episode -245.52\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 259\n",
      "Episode ended after 2563 steps\n",
      "Accumulated reward in this episode -296.88\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 260\n",
      "Episode ended after 1818 steps\n",
      "Accumulated reward in this episode -253.41\n",
      "Mean of x velocity in this episode -0.02\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 261\n",
      "Episode ended after 2321 steps\n",
      "Accumulated reward in this episode -295.56\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 262\n",
      "Episode ended after 3303 steps\n",
      "Accumulated reward in this episode -374.64\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 263\n",
      "Episode ended after 2840 steps\n",
      "Accumulated reward in this episode -345.10\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 264\n",
      "Episode ended after 2609 steps\n",
      "Accumulated reward in this episode -329.53\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 265\n",
      "Episode ended after 3717 steps\n",
      "Accumulated reward in this episode -423.50\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 266\n",
      "Episode ended after 4894 steps\n",
      "Accumulated reward in this episode -522.10\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 267\n",
      "Episode ended after 3396 steps\n",
      "Accumulated reward in this episode -399.77\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 268\n",
      "Episode ended after 4062 steps\n",
      "Accumulated reward in this episode -453.26\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 269\n",
      "Episode ended after 3852 steps\n",
      "Accumulated reward in this episode -437.05\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 270\n",
      "Episode ended after 109 steps\n",
      "Accumulated reward in this episode -95.51\n",
      "Mean of x velocity in this episode 0.20\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 271\n",
      "Episode ended after 3030 steps\n",
      "Accumulated reward in this episode -355.99\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 272\n",
      "Episode ended after 1955 steps\n",
      "Accumulated reward in this episode -270.61\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 273\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -103.00\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 274\n",
      "Episode ended after 86 steps\n",
      "Accumulated reward in this episode -102.77\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 275\n",
      "Episode ended after 53 steps\n",
      "Accumulated reward in this episode -104.11\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 276\n",
      "Episode ended after 75 steps\n",
      "Accumulated reward in this episode -101.90\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 277\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -106.38\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 278\n",
      "Episode ended after 138 steps\n",
      "Accumulated reward in this episode -106.43\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 279\n",
      "Episode ended after 60 steps\n",
      "Accumulated reward in this episode -104.68\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 280\n",
      "Episode ended after 48 steps\n",
      "Accumulated reward in this episode -102.67\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 281\n",
      "Episode ended after 52 steps\n",
      "Accumulated reward in this episode -102.36\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 282\n",
      "Episode ended after 73 steps\n",
      "Accumulated reward in this episode -98.04\n",
      "Mean of x velocity in this episode 0.22\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 283\n",
      "Episode ended after 2823 steps\n",
      "Accumulated reward in this episode -315.87\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 284\n",
      "Episode ended after 115 steps\n",
      "Accumulated reward in this episode -101.26\n",
      "Mean of x velocity in this episode 0.13\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 285\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -374.88\n",
      "Mean of x velocity in this episode -0.00\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 286\n",
      "Episode ended after 3696 steps\n",
      "Accumulated reward in this episode -393.59\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 287\n",
      "Episode ended after 4882 steps\n",
      "Accumulated reward in this episode -478.70\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 288\n",
      "Episode ended after 3452 steps\n",
      "Accumulated reward in this episode -371.76\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 289\n",
      "Episode ended after 3319 steps\n",
      "Accumulated reward in this episode -358.44\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 290\n",
      "Episode ended after 2484 steps\n",
      "Accumulated reward in this episode -297.92\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 291\n",
      "Episode ended after 3997 steps\n",
      "Accumulated reward in this episode -398.38\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 292\n",
      "Episode ended after 3388 steps\n",
      "Accumulated reward in this episode -352.70\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 293\n",
      "Episode ended after 3199 steps\n",
      "Accumulated reward in this episode -335.73\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 294\n",
      "Episode ended after 109 steps\n",
      "Accumulated reward in this episode -101.43\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 295\n",
      "Episode ended after 114 steps\n",
      "Accumulated reward in this episode -98.42\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 296\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -304.79\n",
      "Mean of x velocity in this episode -0.00\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 297\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -293.91\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 298\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -289.85\n",
      "Mean of x velocity in this episode -0.00\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 299\n",
      "Episode ended after 171 steps\n",
      "Accumulated reward in this episode -96.71\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -94.08\n",
      "-------------------------------------------------\n",
      "Episode 300\n",
      "Episode ended after 144 steps\n",
      "Accumulated reward in this episode -88.76\n",
      "Mean of x velocity in this episode 0.23\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 301\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -248.41\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 302\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -224.79\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 303\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -207.90\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 304\n",
      "Episode ended after 164 steps\n",
      "Accumulated reward in this episode -90.75\n",
      "Mean of x velocity in this episode 0.19\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 305\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -180.13\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 306\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -190.42\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 307\n",
      "Episode ended after 83 steps\n",
      "Accumulated reward in this episode -95.88\n",
      "Mean of x velocity in this episode 0.19\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 308\n",
      "Episode ended after 79 steps\n",
      "Accumulated reward in this episode -96.77\n",
      "Mean of x velocity in this episode 0.18\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 309\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -220.96\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 310\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -246.12\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 311\n",
      "Episode ended after 1734 steps\n",
      "Accumulated reward in this episode -207.98\n",
      "Mean of x velocity in this episode -0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 312\n",
      "Episode ended after 3206 steps\n",
      "Accumulated reward in this episode -287.82\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 313\n",
      "Episode ended after 3323 steps\n",
      "Accumulated reward in this episode -299.28\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 314\n",
      "Episode ended after 4607 steps\n",
      "Accumulated reward in this episode -378.47\n",
      "Mean of x velocity in this episode -0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 315\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -261.27\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 316\n",
      "Episode ended after 1598 steps\n",
      "Accumulated reward in this episode -167.00\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 317\n",
      "Episode ended after 1750 steps\n",
      "Accumulated reward in this episode -188.74\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 318\n",
      "Episode ended after 3047 steps\n",
      "Accumulated reward in this episode -265.03\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 319\n",
      "Episode ended after 122 steps\n",
      "Accumulated reward in this episode -91.97\n",
      "Mean of x velocity in this episode 0.22\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 320\n",
      "Episode ended after 109 steps\n",
      "Accumulated reward in this episode -93.55\n",
      "Mean of x velocity in this episode 0.21\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 321\n",
      "Episode ended after 159 steps\n",
      "Accumulated reward in this episode -94.84\n",
      "Mean of x velocity in this episode 0.16\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 322\n",
      "Episode ended after 249 steps\n",
      "Accumulated reward in this episode -97.58\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 323\n",
      "Episode ended after 135 steps\n",
      "Accumulated reward in this episode -109.46\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 324\n",
      "Episode ended after 115 steps\n",
      "Accumulated reward in this episode -103.52\n",
      "Mean of x velocity in this episode 0.11\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 325\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -310.64\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 326\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -336.94\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 327\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -345.94\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 328\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -348.16\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 329\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -346.42\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 330\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -347.15\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 331\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -352.97\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 332\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -341.56\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 333\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -347.06\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 334\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -346.35\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 335\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -341.78\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 336\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -337.15\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 337\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -330.45\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 338\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -330.51\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 339\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -321.34\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 340\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -315.79\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 341\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -311.93\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 342\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -294.68\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 343\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -280.41\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 344\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -253.41\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 345\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -236.57\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 346\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -232.18\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 347\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -206.45\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 348\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -206.33\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 349\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -197.10\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 350\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -190.25\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 351\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -180.35\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 352\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -196.07\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 353\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -191.18\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 354\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -199.89\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 355\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -168.92\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 356\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -181.49\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 357\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -189.25\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 358\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -171.49\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 359\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -175.84\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 360\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -174.77\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 361\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -171.45\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 362\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -163.55\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 363\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -167.50\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 364\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -154.95\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 365\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -180.34\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 366\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -190.64\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 367\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -181.62\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 368\n",
      "Episode ended after 171 steps\n",
      "Accumulated reward in this episode -96.59\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 369\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -197.15\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 370\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -166.72\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 371\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -169.81\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 372\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -169.38\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 373\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -230.46\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 374\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -217.97\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 375\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -162.75\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 376\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -197.63\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 377\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -210.48\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 378\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -182.94\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 379\n",
      "Episode ended after 3813 steps\n",
      "Accumulated reward in this episode -243.63\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 380\n",
      "Episode ended after 113 steps\n",
      "Accumulated reward in this episode -101.79\n",
      "Mean of x velocity in this episode 0.10\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 381\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -250.66\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 382\n",
      "Episode ended after 1557 steps\n",
      "Accumulated reward in this episode -176.26\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 383\n",
      "Episode ended after 94 steps\n",
      "Accumulated reward in this episode -103.92\n",
      "Mean of x velocity in this episode 0.07\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 384\n",
      "Episode ended after 67 steps\n",
      "Accumulated reward in this episode -108.70\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 385\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -103.33\n",
      "Mean of x velocity in this episode 0.09\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 386\n",
      "Episode ended after 70 steps\n",
      "Accumulated reward in this episode -106.81\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 387\n",
      "Episode ended after 133 steps\n",
      "Accumulated reward in this episode -104.79\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 388\n",
      "Episode ended after 98 steps\n",
      "Accumulated reward in this episode -107.02\n",
      "Mean of x velocity in this episode 0.08\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 389\n",
      "Episode ended after 117 steps\n",
      "Accumulated reward in this episode -108.98\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 390\n",
      "Episode ended after 118 steps\n",
      "Accumulated reward in this episode -108.96\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 391\n",
      "Episode ended after 102 steps\n",
      "Accumulated reward in this episode -119.84\n",
      "Mean of x velocity in this episode -0.13\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 392\n",
      "Episode ended after 101 steps\n",
      "Accumulated reward in this episode -120.24\n",
      "Mean of x velocity in this episode -0.14\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 393\n",
      "Episode ended after 127 steps\n",
      "Accumulated reward in this episode -121.62\n",
      "Mean of x velocity in this episode -0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 394\n",
      "Episode ended after 93 steps\n",
      "Accumulated reward in this episode -119.26\n",
      "Mean of x velocity in this episode -0.15\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 395\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -120.98\n",
      "Mean of x velocity in this episode -0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 396\n",
      "Episode ended after 2411 steps\n",
      "Accumulated reward in this episode -225.11\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 397\n",
      "Episode ended after 148 steps\n",
      "Accumulated reward in this episode -129.31\n",
      "Mean of x velocity in this episode -0.08\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 398\n",
      "Episode ended after 619 steps\n",
      "Accumulated reward in this episode -126.98\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 399\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -303.16\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 400\n",
      "Episode ended after 4085 steps\n",
      "Accumulated reward in this episode -358.88\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 401\n",
      "Episode ended after 1991 steps\n",
      "Accumulated reward in this episode -220.71\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 402\n",
      "Episode ended after 106 steps\n",
      "Accumulated reward in this episode -100.65\n",
      "Mean of x velocity in this episode 0.14\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 403\n",
      "Episode ended after 114 steps\n",
      "Accumulated reward in this episode -95.95\n",
      "Mean of x velocity in this episode 0.20\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 404\n",
      "Episode ended after 112 steps\n",
      "Accumulated reward in this episode -93.39\n",
      "Mean of x velocity in this episode 0.22\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 405\n",
      "Episode ended after 154 steps\n",
      "Accumulated reward in this episode -93.70\n",
      "Mean of x velocity in this episode 0.19\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 406\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -366.49\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 407\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -357.46\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 408\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -359.89\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 409\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -357.51\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 410\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -355.68\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 411\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -351.51\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 412\n",
      "Episode ended after 89 steps\n",
      "Accumulated reward in this episode -102.05\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 413\n",
      "Episode ended after 139 steps\n",
      "Accumulated reward in this episode -95.78\n",
      "Mean of x velocity in this episode 0.18\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 414\n",
      "Episode ended after 167 steps\n",
      "Accumulated reward in this episode -93.67\n",
      "Mean of x velocity in this episode 0.19\n",
      "Maximal reward so far -88.76\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 415\n",
      "Episode ended after 180 steps\n",
      "Accumulated reward in this episode -101.70\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 416\n",
      "Episode ended after 1938 steps\n",
      "Accumulated reward in this episode -176.60\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 417\n",
      "Episode ended after 743 steps\n",
      "Accumulated reward in this episode -134.67\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 418\n",
      "Episode ended after 905 steps\n",
      "Accumulated reward in this episode -127.46\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 419\n",
      "Episode ended after 2200 steps\n",
      "Accumulated reward in this episode -132.59\n",
      "Mean of x velocity in this episode 0.06\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 420\n",
      "Episode ended after 2596 steps\n",
      "Accumulated reward in this episode -184.37\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 421\n",
      "Episode ended after 3827 steps\n",
      "Accumulated reward in this episode -254.29\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 422\n",
      "Episode ended after 3311 steps\n",
      "Accumulated reward in this episode -242.09\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 423\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -234.47\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 424\n",
      "Episode ended after 749 steps\n",
      "Accumulated reward in this episode -130.53\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 425\n",
      "Episode ended after 123 steps\n",
      "Accumulated reward in this episode -98.55\n",
      "Mean of x velocity in this episode 0.12\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 426\n",
      "Episode ended after 4956 steps\n",
      "Accumulated reward in this episode -326.28\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 427\n",
      "Episode ended after 2146 steps\n",
      "Accumulated reward in this episode -176.52\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 428\n",
      "Episode ended after 88 steps\n",
      "Accumulated reward in this episode -96.06\n",
      "Mean of x velocity in this episode 0.19\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 429\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -215.96\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 430\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -184.00\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 431\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -138.42\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 432\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -172.09\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 433\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -203.32\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 434\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -236.15\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 435\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -250.44\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 436\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -272.71\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 437\n",
      "Episode ended after 204 steps\n",
      "Accumulated reward in this episode -93.16\n",
      "Mean of x velocity in this episode 0.16\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 438\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -262.61\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 439\n",
      "Episode ended after 2049 steps\n",
      "Accumulated reward in this episode -184.40\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 440\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -250.18\n",
      "Mean of x velocity in this episode 0.00\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 441\n",
      "Episode ended after 3664 steps\n",
      "Accumulated reward in this episode -268.55\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 442\n",
      "Episode ended after 129 steps\n",
      "Accumulated reward in this episode -94.21\n",
      "Mean of x velocity in this episode 0.20\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 443\n",
      "Episode ended after 2216 steps\n",
      "Accumulated reward in this episode -221.28\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 444\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -284.00\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 445\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -295.90\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 446\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -314.91\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 447\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -322.85\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 448\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -323.29\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 449\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -336.55\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 450\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -336.64\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 451\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -339.39\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 452\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -336.11\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 453\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -343.57\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 454\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -344.60\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 455\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -339.72\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 456\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -334.19\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 457\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -334.54\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 458\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -331.28\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 459\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -332.96\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 460\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -324.50\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 461\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -330.97\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 462\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -319.34\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 463\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -312.72\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 464\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -304.95\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 465\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -301.72\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 466\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -295.73\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 467\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -290.58\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 468\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -288.77\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 469\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -279.38\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 470\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -268.74\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 471\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -265.86\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 472\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -251.02\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 473\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -245.42\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 474\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -233.45\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 475\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -218.98\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 476\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -212.36\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 477\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -225.99\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 478\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -207.57\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 479\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -214.72\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 480\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -190.98\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 481\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -174.65\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 482\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -179.34\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 483\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -167.14\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 484\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -167.01\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 485\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -154.94\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 486\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -162.92\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 487\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -164.03\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 488\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -160.06\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 489\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -169.73\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 490\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -160.03\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 491\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -161.21\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 492\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -157.44\n",
      "Mean of x velocity in this episode 0.04\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 493\n",
      "Episode ended after 5000 steps\n",
      "Accumulated reward in this episode -195.68\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 494\n",
      "Episode ended after 2197 steps\n",
      "Accumulated reward in this episode -185.61\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 495\n",
      "Episode ended after 4170 steps\n",
      "Accumulated reward in this episode -303.51\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------------\n",
      "Episode 496\n",
      "Episode ended after 793 steps\n",
      "Accumulated reward in this episode -123.87\n",
      "Mean of x velocity in this episode 0.05\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 497\n",
      "Episode ended after 1608 steps\n",
      "Accumulated reward in this episode -168.80\n",
      "Mean of x velocity in this episode 0.03\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 498\n",
      "Episode ended after 2290 steps\n",
      "Accumulated reward in this episode -213.50\n",
      "Mean of x velocity in this episode 0.01\n",
      "Maximal reward so far -88.76\n",
      "-------------------------------------------------\n",
      "Episode 499\n",
      "Episode ended after 1901 steps\n",
      "Accumulated reward in this episode -189.00\n",
      "Mean of x velocity in this episode 0.02\n",
      "Maximal reward so far -88.76\n"
     ]
    }
   ],
   "source": [
    "ep_action, ep_obs, ep_reward = [], [], []  # Allocate space for episode actions, observations and rewards\n",
    "tot_ep_reward = [] # Total episode reward\n",
    "mean_reward = []\n",
    "mean_vel = [] # mean of velocity in each episode\n",
    "\n",
    "''' Run TF session '''\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "\n",
    "    ''' Run episodes '''\n",
    "    for ep in range(num_episodes): \n",
    "        obs = env.reset()  # Reset and save first observation\n",
    "        ep_obs.append(obs) # append observation\n",
    "\n",
    "        ''' Run steps '''\n",
    "        while True:\n",
    "            # Propagate forward to compute action probability distribution and sample an action out of it\n",
    "            action = np.squeeze(sess.run(sample_action, \n",
    "                                         feed_dict = {input_ : obs.reshape((1,obs_dim[0]))}))\n",
    "            m = sess.run(Mfc2, feed_dict = {input_ : obs.reshape((1,obs_dim[0]))})\n",
    "            # clip action values to fit into the allowed box\n",
    "            action = np.clip(action, env.action_space.low[0], env.action_space.high[0])\n",
    "            \n",
    "            obs, reward, done, info = env.step(action)  # Take action and save observation, reward and done boolean\n",
    "            \n",
    "            ep_action.append(action)     # append action\n",
    "            ep_obs.append(obs)           # append observation\n",
    "            ep_reward.append(reward)     # append reward\n",
    "\n",
    "            if done: \n",
    "                # Stack vertically episode parameters to one np.array\n",
    "                ep_action = np.vstack(ep_action)\n",
    "                ep_obs = np.vstack(ep_obs)\n",
    "                ep_reward = np.hstack(ep_reward)\n",
    "                mean_vel.append(np.mean(ep_obs[:,2]))\n",
    "                \n",
    "                # Discount rewards\n",
    "                dis_rewards_arr = discount_rewards(ep_reward)\n",
    "                # Compute loss and optimize\n",
    "                sess.run([loss, training_opt],\n",
    "                         feed_dict = {input_ : ep_obs[:-1], actions : ep_action,\n",
    "                                      dis_rewards : dis_rewards_arr})\n",
    "                \n",
    "                tot_ep_reward.append(np.sum(ep_reward))  # Compute total reward for episode\n",
    "                mean_reward.append(np.mean(tot_ep_reward))\n",
    "                \n",
    "                 # print info\n",
    "                print(\"-------------------------------------------------\")\n",
    "                print(\"Episode {}\".format(ep))\n",
    "                print(\"Episode ended after {} steps\".format(ep_action.shape[0]))\n",
    "                print(\"Accumulated reward in this episode {:.2f}\".format(tot_ep_reward[ep]))\n",
    "                print(\"Mean of x velocity in this episode {:.2f}\".format(np.mean(ep_obs[:,2])))\n",
    "#                 print(\"Mean reward so far {:0.2f}\".format(np.mean(tot_ep_reward)))\n",
    "                print(\"Maximal reward so far {:.2f}\".format(np.max(tot_ep_reward)))\n",
    "                \n",
    "                ep_action, ep_obs, ep_reward = [], [], []  # Clear episode values for next episode\n",
    "                      \n",
    "                break\n",
    "                \n",
    "    saver.save(sess, \"models/BipedalWalker/model.ckpt\") # save model for later\n",
    "#     writer.add_graph(sess.graph) # Save graph for displaying with TensorBoard\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Assess model performance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to get a sense of how good our model is, and to see if our agent improves in the training process, we can plot the mean rewards gained in one episode over the episode number. This value would grow as our agent becomes more skilled and gains more rewards in each episode."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(mean_vel)\n",
    "plt.xlabel(\"episode\", fontsize=14)\n",
    "plt.ylabel(\"velocity\", fontsize = 14)\n",
    "plt.title(\"velocity over episode number\", fontsize = 16)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgQAAAEcCAYAAAC4b6z9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzsfXeYHEeZ/lvdkzYnaZVzlizLQZZly2EdANmYnEww2PiwCSbdDx+YcMCBDzgwNkc48HFgbIINGINxTlrJtqxk5RxXcRV2tXl3Unf9/qiu7urqMD0zm2T1+zx6VtNTXV1d01311fu931eEUooQIUKECBEixNkNZagbECJEiBAhQoQYeoQGQYgQIUKECBEiNAhChAgRIkSIEKFBECJEiBAhQoRAaBCECBEiRIgQIRAaBCFChAgRIkQIhAZBiDcoCCHvJIT86xCefzMhhBJCJhdax5kIQsgDhJCmoW7HQIMQ0kQIeWAQr3dW9GuIoUVoEIR4o+KdAAqe0Pvh/BBvbLwLwHeGuhEhQvQnIkPdgBAhQthBCIlTSlND3Q43EEKiALL0LM9oRindMNRtCBGivxEyBCHecDCo3I8BGGfQ9lSkWwkhswghjxFC2gkhfYSQVYSQpUHOJ4QkCCH3EkK2EkK6CSHHCSH/JITMLrCtjYSQVwghbyOEbCCEpAB82vguQgi5ixCykxCSIoQcI4TcQwhJCOdvJYT8WvhcRQjRCCFHpOu8Sgj5s/D5DkLIa4SQ00Y/rCKEvFU6Z7Jx758mhPwXIeQYgBSAauP7awgh6wkhSULIPkLI7QHveRsh5FGX4xcb13un8Xmm8TudNK5xiBDyF0JI3guZgH0p3u+Pjev2EkKekF0/ssuAEDKaEPI7o94UIaTZOK9eKDOGEPIgIaTFKLOZEPIRl7YG6ldCSCkh5AeEkAOEkLTx92uEEEUoU04I+anRdylCyAlCyAuFPq8h3tgIGYIQb0R8B8BIABcBeLtxLAUAhJCxAF4B0AXgDgAdAD4D4ElCyA2U0qf9zgcQB1AB4LsAmgHUgk3gqwghsymlxwto70wA/21cdz+A08bx3wN4G4AfAFgJYI5RZjKA9xhlXgJwg1BXg9HWcYSQmZTS3YSQMuNefi+Umwzg1wCawMaBtwF4ghByvdEHIr4GYC2A2wCoAJKEkDkAngKwDsCNYP3yLQDlALQc9/sQgG8TQmoopW3C8Y8Y9/6U8fkJAO0APgWgBcA4ANejsIVMkL7kuAvARgC3AKgH8J8AniOEzKOUZnzuaRKAOwEcBjAKwDUASgHA+A2WA6gB8FWjzEcAPEQIKaWU3m+UC9SvhlH0LIC5xn1sAbAYwDfAnsn/ZxS9F+wZ/iqAPQDqACyBYdSFCGEDpTT8F/57w/0D8ACAIy7HfwQgC2C6cEwFsAvA+lznu9Sngg36XQC+KBy/GQAFMDnH+Y0AdADnSccvN87/qHT8w8bx84zP7zI+TzI+3wfgcbDB/3bj2FKjzGyPNihgRsFzAP4hHJ9snLceAJHO+QPYJF0mHJsAIA2gKcc9TwCb3G4XjkUBnALwC+PzCOPab++HZyFoX/L73Q5AEcotMY7fKhxrAvCA8LkbwOd82nCHUUeDdPwFACcBqPn0K4CbjPqukOr7mlG23vi8FcCPB/PdC/+duf9Cl0GIsw1XAFhFKd3LD1BKNQB/AnAeIaQyVwWEkPcTQlYTQtrBjIsesBXcrALb1EQp3SgdWwo2sD9q0N0RY1X4nHAfAFt16gCuNj5fDcYavCQda6aU7hTu4UKD0j5h3EMGwJs87uHvlFJZM3AJgKcopT38AKX0MIBXc92sUW452KQm3u8IAA8an1vB2JLvE0I+QQiZkateHwTtS46/Ukp1ob2vAjgCds9eWAvgTkLI5wkh8wkhRPr+CgBHKaWN0vHfg7FRc43PQft1KYCDAFa63FMUjC3g7bqZEPJVQshCQojqcw8hznKEBkGIsw21YFS/jOMACBil6wlCyNsAPAJgB4APAbgYjI4/BSDhc6of3NpTDyAGtvLMCP9OGt/XAQCl9DSATQCuIoSMAHAOgGXGvwaj7FXGZ34PEwC8CNYXnwVwqXEPz3jcg1v7xgA44XLc7ZgbHgSwhBAyxfh8E4C9lNJVxn1RMANlHYDvAdhNCNlPCPlUwPpFBOrLHPdwAsxl4YUPgDEz/wZgM4CjhJB/F/z5fs8d/x4I3q/1YC6KjPRvjfE9v6fPAvgVgI+DGQcnCdPAlPrcS4izFKGGIMTZhtMARrscHw1GwZ52+U7EjWAT1838AGHK+1rPM3LDTbHfCiAJRne74Zjw/2VgE9JVxnmbwSafekLIEgDng00KHEsBVAF4P6XUFB/6TBJu7WsG85PLcDvmhkcB/BzARwghPwHz73/PdlFK9wP4qLHaXgBGu/+CENJEnToHP+TTl4D3fcksjtjWk2BalM8QQmaBiVK/DWYo/g/Yc+XGvvBnsdX4G7RfWwEcAPB+jyY1Ge3qBtNE3EUImQTgvQC+D8aYfNnrfkKcnQgZghBvVKQAlLgcXw5gsagaN2jUDwDYQCntynF+KRjFLuImMC1Bf4Kv1qsopetc/skGwTgAtwNopAwnAWwDm5RUMBeCeA8AW1ECYIp+MF95ULwG4HpDLMfrmBC0DqOf/wHWd+8z7vUhj7LUcKnwvBDn5NFOIL++BID3Skr9JQDGg91zTlBKd1FKvwqgTWjrcgDjjbpEfAiMqdhhfA7ar8+AaQu6Pe6pxaVdByml94AJEPPtwxBnAUKGIMQbFdsB1BoU8zoASUrpFjDV9c0AnieEfBNAJ1iUwEwAbw1w/jMA3kkIuRdMBX8hgM+BqeH7DZTSRkLInwD8lRDyYzAqWAcTvl0P4MuU0t1G8RVgIr1rwFapHMvAVtWHjNU2xwtgRs2DhJB7wGjqbwM4hOCLhO+CTeTPEUJ+CEbJfxvBXQYAcxt80DjvFUrpAf4FIeRcAD8Bc8/sBTNqbjba/ZJQLgvgd5TSW70ukmdfAiyK5O+EkF+B+fe/BybSfBAuIIRUgfXpHwDsBDO03gHmfuI6hQcAfB7A3wghXwPTJHwYzC1yu6FjAYL36x/AoiBeNH7DTUbZaWBRBe+klPYSQl4Dc2VsAXOZXAnGtvzOq79CnMUYalVj+C/8NxD/AJSBCQXbwCjvJuG7WQD+DhZymASwCsDSIOeDTZjfBaOZe8FWfufDqTq/GcGjDF7x+E4Bm0Q2Ge3sMP7/X2CrXbHsakiRBLAiEB5wqfv9YJNXEoxJuBFs0hL7abJx/r94tO9aABvA2JT9YAyFrY4c966CUeQUwG3Sd/Vgk9Zuo59PG339Fqmc6/0V0pfC/X4awI/B6P5eAE8CmCLVZ/7eYKGBvzL6sRvMyFwL4EPSOWPAWJAWo882A/hIof0Kxnp8y/gdU0YfrTWORYwyPzDq6gATv26BTzRE+O/s/kcoPasTjoUIESIEAJaYCMwv/wlK6a/9S4cI8cZDqCEIESJEiBAhQoQGQYgQIUKECBECocsgRIgQIUKECBEyBCFChAgRIkQInGVhhyNGjKCTJ0/u1zp7enpQVlaWu2AIT4R9WDzCPuwfhP1YPMI+LB793Yevv/56C6V0ZK5yZ5VBMHnyZKxbt65f62xsbERDQ0O/1nm2IezD4hH2Yf8g7MfiEfZh8ejvPiSEHAxSLnQZhAgRIkSIECFCgyBEiBAhQoQIERoEIUKECBEiRAiEBkGIECFChAgRAqFBECJEiBAhQoRAaBCECBEiRIgQIRAaBCFChAgRIkQInGV5CEKECBHCC09taUY6pQ91M0KEGDKEDEGIECHOehw+3YtP/2E9frs1DQBYsfsUVu1vzasOXQ/3hQlxZiM0CEKECHHWY/2hNgCAQtjnj/5mDW68f1Xg8/93xX5c/l/LsMGoJwTQ0ZvBPzYexdH2PtfvdZ1i94mufrnWgZYerNzX0i91nc0YFgYBIeR9hJBthBCdELJQ+u4uQsheQsguQshbhONLjWN7CSFfGfxWhwgRohjc/tA6fOeJ7YHKdqeyeHXvwA34Gw61AwDGV9iHREop9p3qRq5dYZtae3C0vQ8f+NUqPLft+IC180wBpRSffXgDPv/wRtz6wFrXMl94ZCPefO8KbD/WGajO7lQWGc3dpXP/in248y+bzc/ff3onJn/lyfwbfpZjWBgEALYCeDeAFeJBQshcADcCmAdgKYBfEEJUQogK4OcArgMwF8AHjbIhQgwqspqOrKYjmdHw+1UHkcpqnmW9BrOzFU0tvWhq6QlU9gsPb8SHf70aJzqTvuU6+jI4maMMRzpr/R6bjjCDQB4QV+xpwTX3LMdb//sV37p0SlGZiGDW6Ap8+dHNaOlOBWrDYOG/X9yDHzyzc9Cu98KOk1ix+xQAoCeddXy/63gXHt90DACw5kAw18w533wWtz3ovhdNKqvb3q9fLt+Xb5NDYJgYBJTSHZTSXS5fvQPAw5TSFKX0AIC9ABYZ//ZSSvdTStMAHjbKDktQStHemx6Ua3n5MXOtcADgV8v3YePhdpzoTOKlnSf6u2lnHB58rQmffOh1/Mc/t2Pr0Q7bdwdaeqDrFO/+n5X44P+uwoOvNeHrf9+KB1e67yGy/1Q3ZnztaTyztXngG54Daw6cxuHTva7fHW3vw5YjHa7f9Qc0neKpLc3Iajo0SqEFeC4BmNRyX9rb4AKABd9+Dov+88Wc9W092oGZX38az29nz3l7bwYAIL8+p3vYxL692X8Vq+tASUzF998zH229GTw7zFiClftasHJffpqIYvDI2sMYWRHHO88bC93FDn5p50kAACHAeoOdccPjm45h8leeNA2sZbtOuZbTdQo94LMUwhvDPcpgHADRkXfEOAYAh6XjF7tVQAi5DcBtADBq1Cg0Njb2awO7u7t968zqFC8eyuJPO9O4/dw4xpUT1CQUVMSIWYZSikNdOiZWKKAA/rYng6YOHbedG0dlnHjWvflUFs82ZfD5CxJ4aHsaLX06DnXp+MT8OCZWKqhNMHuvL0tx96o+TK9WcV69ihk1KsqiBEe6dMRVYONJDVOqFHxvdRJ1CYKIApzopfjKogSSWYrH92XwsXkxPLo7gzl1KhQCXD4ugtIoQVea4q6Xe3HbuXGcOzL/x6k7TdHR3ePbh1tbNHRnKNYez+Kd02OYUOG0Y/9vSwplUWBEiYLLxkWw8lgWaQ1YOiWad5s47n6+Bylj/nll+0FUxgj6shTHuinaUhTn1KnY2soK7Glmg9q9z+3Arr178aZJUcRVYrQ/i7/sZhPOX1dsRqLFzfYtDsdOd+P+x17EzBoVaY1iS4uGU70UfVmKd82IAQC2t2royVD8fCMbXK+fEsXmU1nMrlVx4+wYIgrBvzb24nSS4huLE5hWrfZb+0706OjLUrSnKO5bn8J1U6Lo7s6iResL9E729TE/9OrVq7G/lKAvC3SlKZp7dJxfz567DSetlej77nsGb54UxYwa93v45kpW31MrNyF6MobuHmYgJdNpW3u2btth/n/ZsmUgxP19PNqcQiat4cSu9YgpwIr1OzGu70DO+xostJ7uQ1pHzr5e3ZzFlCoF9aWFrxVbO7qxbGcP3jQpgpZTJ9GX1BzX/fvqPkyoUDCqlGDl7mY0Nrobofe+xn6nf7xgMTRu93D8RBLJlPM6/T3eDxZyzSsDhUEzCAghLwAY7fLV1yil//A6zeUYhTuz4WoeUkrvB3A/ACxcuJD215aSWU3HLQ+sxZySLL56g3udh0/34m0/e8VcffxuRwYZjaIspuJjl07GiztOoiuVwfSR5Vi26xSunz8a88ZW4Yn9bMKomXoOxlQlEI+omFBb6qh/0wt7sO313VjTNwovHz0AVSHQdOC+9SlcOq0Of/zEYgDAtx7fhiPdTTjSnUXjkSzqymJ4+ctX4eZ/f9ZWX0xV0Jpk5nwiquD+bTpO9zBm4w/7ItjdksTONoq0puNPO9O4enY93nPBeHRn1uORfQo+9z73fvDDvz6yEVubknjuy+7n7j/Vjdt/8jJSBr17+fzxuKlhJp7YfAyjKhPo6M3gcw9vQK+wcozVjsOD29lgfNWi+Xh84zH84L3nYkdzJ2pKo5heXxGobfN3rMS6g0wktqddh7wA4cYAALSnKL7+1jl4fvsJ/HX3abx6QsX33j0fDbPqcbPgy5w7fTIaGmYFur4bjnckGSPx8UXmM/HKnha8vGs9XmnOomFWLZ7cfMy20h07YSKqS6L4r7V2yvipA+y5PNKdRV+sCp+/ZiZOJ18DAGxJ1uLWhvN92/LKnhacP7EaZXH/YeRoex8+d98KdCazuH7+aADH8fSBDCbVlaKqpgQNDYtz3nfp2mVAXy8WXXwxbv3dWuw/ZbkaXv63S/HX14/gJ+v3mMfWHtdwoFvFuq9f41rfwWfYb3L+vJlouGQy4qtfAvr6EInG2LazxvfTZswCtm4BAFxxZQNUxd0gePzkRjT1nsbVV12FaZtfxsvNPRgzbiS+9fZ5Oe9tMHDvtlehZjQ0NFzhW+72rz+Nj182BRfMHY+1TaexrqkNF06qwZvnjcKI8niga/392Zeg0T4sOW82thxpx77uFsdWvp9d9izedcE4VJdEse6lvbjiiiuhuPTtd9cvB9CNyxYvAlYxj7LbGP6XY+uhtgvXMX4/r3qHO4ZqC+lBcxlQSq+llJ7j8s/LGADYyn+C8Hk8gGM+xwcNqkLw8p4WnOjx9gv/6LldpjFw86WTkczoIADmjq3ET1/ai65UBomIatJgT205jh8+a60eNZ3i2h+vwOX/tcy1fk63PrG5GQ2zRmLff16P+2+6EABwsLUXB1p6cPNv1+C1fa1YOKkGt142BW9fMBatPWm0djtdGJ++ahoe/dQl+MWHL8DfP7MEY6oSqK9gg8DpHnYfNWXWivulnSex7uBpAGzQ//zDG9CX1my+2Vxo602jO23NXgdaemzhXo+sO2yjAuvK2Gr3jj9uwPt++Rq+8+R2mzEAAGnN+nz7Q6/jyS3N2HW8C3f9bQt++tJe87tDrb2eCmiA+aOXzhuNSXWlpjGw9+7r8P/eNBPT68sBAF+4dgbiEQUzR5Xj1sum4JHbL8Ejty1GZSKKjz+wFluPdqA0Zq1S2/sygfsGYEblu37xqulyenT9ERxo6cEf1xwCADS19OAj/7ca645nkcro+OcmZgx85qppuGrWSADA/zTuw/eethsDF0ysBgD86ROL8f13z8ere1vxuT9tAADUV8RNA8yzb3ozuOk3q/GPjblfuz+vPYzOJFu9P73VotI7+jLQAobq8SGdUmozBgDg2h8vx09e3IN3nz8Oi6bUoszs79wTAXex8d9Xpp01ge/OunHfQj2KwR5MHVGG3rSGB1Y25bz+YCEj+dg9yxmamC8/uhl3/W0LHl1/BF99bAsWfvcFPLTK3R0mg3ehSggUQlyp/KxOEY8oiEfZb5Xx6NvelFN/4H5N6uouDeqSCsEwLDQEPngcwI2EkDghZAqAGQDWAFgLYAYhZAohJAYmPHx8MBtGCEFUJdB8nrd1TW24fv5o/ObmhfjGDXNxw7ljcMuSyfjTJxZjwzfehBV3XoXf3HwRFkyoxsO3LcbyOxvwlnmj8O7zmVckl0+MD1aprIbaUjZRvnneaHz44onoy2j4z6d2oHHXKew60YXxNSX4xg1zcdVsNknwlT9HSVTFJy6figsn1eL6+WMwe3Qlnvzc5fjZhy4AYAni+ACeiLJH56ktzSgxXup/bDyGOf/+DN73K7bKfHHHCfzviv2+95DRqK0Pf9m4D1/6yybzcyqjoySqYvmdDQDYQCLiYKvTF+62YtV0ilTWbqxc8cNlWPL9lzzb1tGXQXVpFNUlzAiqKY0ioir47DUz8MAtF+HdF4zDrZdNwY/ffx5+9L4FJp188dQ6/PKmC6FTYPuxTqSyOj5z1TRMris1DcSg2HasExsOtaPJuE+uBSFgsfIr9jBjsjdLkdV1lMcjuPWyKbjzLbPx21sW4fYrp+LCSTUgBHjr/DFmvXe+ZTae/+IVuGRaHd6/cAKiKsHxziSqS6OoLYvlnKhTWQ2UAr0ugjEZ3aksSmMq5o2ttLEs6azu6l/2A2/WF6+diabvvxU3nDsGqayOd543Fj983wKMqy5x6ACC1MfvVz43Izycfn2iUZjsweiqhHk8iHZnMJDV9Zy/KaUUOgU03X4PABsfvvX4Nvzviv2OsUMG77KIQkCIs08BNrYpxhgK2PtZRI9h7PuJdQH227hN/kENzhAMw0JDQAh5F4CfAhgJ4ElCyEZK6VsopdsIIX8GsB1AFsBnKKWacc4dAJ4FoAL4DaV022C3O6Io8FtIaTpFRTyKq2ePAgBzcgWAGmOlO6G2FP/4zBLz+K9uWoitRzvwtw1HkfWzNsBeXIC9TCItNrIijtM9adv6qM6g+6qMya1NEDnWlsXw6pevRknM6W/lgxyfSPmE/C+XTcXPlu3Fic4UblkyGVUlUdz3AqNsNx1m/vRbf8cUwZ+4YqqtzqPtfXjg1QP4xBVT2YpE6MOkNGlrOoWqENQa/cVf8JrSKNo8Jlc3hiKr6dA06jAo/NDRl0FVSRRVhrFVY/wFgPE1pfjx+88DALz13DGOc0eUs7J7TnZB0ylGVSZQVRqz9XsQcEOM/+VjHiHAD5/dheYOxnBkdDbwpjUdEeFZuOu6OQDYxL2uqQ1PbmGixvrKOKaNZCyHohCMLI/jWEcS9RVxY1Xn3y7ej0EG3GRGQyKqOgy1jCEszAfcSFaNpcxPP3g+fvjeBeazS0h+q0JeH/8rnyqyAn73yiY49v+RFRa1Tilr01Ajq1HPSZfDMoooakrZOHH5jBG45/0LUBJV8f5frcLdT+3ApiPttrFMBu8mRSEghDj6FGD+XbaoYj9kJqsDLh4JbnDmYh116r6ACg2C/DAsDAJK6WMAHvP47m4Ad7scfwrAUwPcNF8whiDHIFEAB8Opx76Mv1XMXwB5EuADUmfSmjC5/68y4TQIqkuirsYAIBgEEkMwUdA0nD+xJnD4GKUUV/+oEamsjmkjy5HV7QxBVqe2l1ijzCCIGB3JJyIvcRcAdCWdq9asTh11c+g6dfgZkxkNqayOqtKoaURVlwYXKJbHI4hFFOw8ztTx9RVx1JRGc66unO02DAJjQOStJyBIZ3VTdc8flXRWd/Vzl8YiGFVprfpkf/DIygSOdSQxqjKB9t5MAHaKGu0LYhAwlqdMesYymvvv4Qb+e2vS708IsT27CiEWixJgIrYMAuOz9H1QhkB0GdyyZDKe2HwMW492sjEggOtioJHWcjMEfCzTdIp0Vkd9RRwP3LLIfJ4e+/SluP4nL2NHrogLgSFQiDtLQg0DyjQIPNwZvP9zubB0nbqyTfksAEIMf5fBsEZUVeDnltOp/8TlhYhBo+Wk5oyHPZ3VbRNafUXCcX6dsWLlkxvXBABARcLbLlSlgZj76UZXJUy67/wJ1Y5JqNvD93f4dJ/5cmd0ylbuwkub1XTbS8wHWl6/uIrxAl81i9AMY8BtUGzpSYFS1haODsPXX1UiugxijnO9QAhBXVkMO5qZQTCyIoHqkmgBDAFr79Nbj+PhNYdsDEFG182+FO8q4iGi4nqQmKqgUvrN+XcjK+JQSO6VVV4MQVZDPKq4unLyDRXj1/MS9wVpO2AZC6YhIBkGHCJL5ze5cCYLAOIRFdedw1ij4eLDzmrUVwMBwJxQNcrYhERUtfVzIqri+vlj0NTa60vh2xgCuGsIdMqMt1iETUHpHPqGXC4D3SOENUwnnR9Cg6AIRFUFWZ/njQo0Yj7gKw034Z8IceBTiZMh2Huy2zzGKexK7jIQjIWKhPfKV2Y4+KAYUQnGVpdgRHkc42tKHKuxlUJWuae2NJuT9IbDVmrXrKYjo1FbH8qTNh9o+STHB2g/d8rRNqdB4McQHGtP4v9eOYDz/uN57DJW9KJBYDEEwQ0CgLliePz0qMo4qktjeWsI+H0+tOogvvK3LaCwNARZDxeI6kFLVZdGEVMVjCiPOQxVbhDUVySgKO6DuAiuXwmyAktlNCQiKso9tB1BwFtrGgQehnYQd4cI2WXgMAiESdRvcuETnNgOwOmCGCpkJEPbDXxC1XUWScQNfhEzRpVD0ymaWtzzWLB62F+TIXApw10sMZMh8G9bLpeB5uEyCBmC/BAaBEUgGiEOhmBd02nz4dUp9Ry4/MCt8lZh0u7ozeChVQeRFNwINoPAxWUgvgt+LgNfhkCyaPg1FULw5rmj8J4Lx4EQ4rjP5butBCKf/sN6fNDIC7/psBVvzAYp3eYyyEgrGc0QHymGQMmaiLwHiGMdzkx1nIlwNwj6cP+K/ehOZfGNf2wFYCWqqS6Jma6CfFwGAEzdA8B+k+rSKLqSWRsTkQvyfZpjHiGek2nEZSBnpxCMrIhjRIXTWctZpfqKOFQPZbi9Xdwws7dvzYHTuPKHy9AjMER9GQ0JD4YgXx8vn7S8QslEQyfIm2dGF0jRBhyZgAyB7B5UiHV8OIBFDwRzA2k6RSarIxZxuhFnGCG7z207jg2H2vC7lU0OlwC/Z4UwDYFsSFFKWT+LGoKcDIH/97xOr7aECIZhoSE4UxFVFNsgsaO5E+/95Wv4+JIp+Pe3zS3cZaBwl4GV/vQHz+7EH1cfwmv7WvCLD7PQQpEiE1cnnA0QwUWFiaiCmKrYDIJKH4ZApp+zgkHwtbda2aIV6T7lneL4JL3zeCfmjKnEjuZONvlr1GZUyZO2LlCxEYUgmdVxqivlO5GIq4m6shhae9IGQ2B3T5REVfRlNLy2rxUnu1hf7z/FWBWRIagUogzyATfCRpTHEY+opuuhoy9j/h65IK+MLA2B9yDqRacDwIIJVagqcT4f9ZVx86+ieBsbHOLkIWLn8U4cbO3Fqa4USmMqHl1/FK3dadSWxfrFZUDNycb9e/F4IA2BFF0g37YWUFSo6Xbjn78Pw2WBmtVzZ/Izf1PKGIKYi2E5dWQZSmMq7nl+N+55fjcAtrj4wrUzzTKihoAQp5HFPzMNgV207IVUJgdDIDyPokEcMgT5IWQIioDsMuAU/64TTHQjCo3yAV/9iC6DR18/AsBK+cnr51CFXzIeUTGuusRWJ4/fJ4RTy+LKAAAgAElEQVSgsiRi0xf4MQRe7ZcHZLlYhxRvzw2LVFY3J8aMpjOxk2DZy9SmRq1zVYXg96sO4pp7GnNSjFUlUURVYoZPcUNDXHHz/ABPGcr7d543Fi3daXzxkY143UhIJGoICnEZAMCiKTW2872iI9zgGNAEwZzXYOelIQCAX3z4Qnzv3fMdxyfUlJp/FYKc4YBeokI+cPdlNBw63Ysv/WUTdh7vYlEGLsLVwAyBcUt8letl9PgZQ27QHC4DaltlyhEvXtAptbMTZzBDoOsUGU03V+8iElEVL/6/K/HJK6dhYm0pKuIRrDlw2laGd5OqsjwE8lX5Z4UQRH00BOLvkMrBIPB+lnUEWo57Hi7Yf6obq/PcbnsgEBoERSCi2l0GVHr0xVCkfMBXGuKkncrqmDqyzOFf55Ap1PnjqgAwi/7aOfVIRK3BuLIkijabqNB75es1wMrMh1xOzjnPv89qOmIRBapCbIOUaOFTKqzchAiAiKKgN62ZSW5kcIESwFYel0wbgcVT6wBYhoYc0QBYrpkLJ9cCAB7bcNTcHEXUEOQjKgSATsMoumAiMwi4gdCax8Y3MiVvMQTEkyEoxAhdMr0Of779Epw7vgpqAA2Bl6iQu7SSGQ2dfdbvVOISdggEX0HLGgJvQ1V0GQRITETt9VKpTWKkTy6XgfgOmAzBMFihUkMkmNWpq+Kfw5pUgUyWuhoEADCmqgRfuW42VvzbVZg3rlLQ9XBXKSvHEhM5jSJdYHliYtihBNFNkMoVcaXze7UfHy6izlz4n8Z9uPOvm3MXHGCEBkER8BIV8oFIp96+Tj+4aQgANqhmPQwC2Yc/rb4MALB03mj8+mMX2b6rTERxOqCGIChDIJeTQyZNUaBOEVFYQhJR+czvKyOtPEUqNtfqb9O/v9n8v6oQPPjxRbj1simsXo0ZGrbMc8KEGlUJ5o5xpjROxBQsmFCNmxZPwpLpdb7XlzFrNKtvyfQRAGCG/XH3RBB4MSHER03vpSHwAyEEi6bUghjZ5XINpBZDoGPPiS5TRJo01OB9Gc22y108qhQlKjTLU3+DIIgtZPqwYa1CrUyFdt1GXyYgQ6Db30H+rA4De8BzEeFVTtMZcyca2F5guVh0/E/jPkz/2tPMnWdzGTiNS91kuSwNQU86i1t+u8a2ShYjlXJpCMSQSfs9nRk7jKY13TMyazARGgRFIKYqtgdQHkMZjZh/vXwwkWn3eESxrZ7FQVueLOeMqWRtcqm/siRq230xH1Ehh8wQyIaPPO5w5XtWYz6+qKIgbUQZAJY/nE/SIiVtMQTenVkaU1ESU01Dhect4JMjD1vyCiOrKY2ZtDkAk96OKgoSURXfeec5ebsMblkyBcvvbDB/C67kz8cg8BIV8igDN+RLm8tQXIRgjnYJ0R5vuncFPvTr1QAsl0Eqo9uEhWJiohKBrcrXILBcBt5tzwXxPXWLMhDbJDJd/pkK7e/6cBIVBhVGiixdOuvuMpARUQlae9Lm1sqtPSkpMZGfhsDKVLjmQBuW7TqFj/12jVlOzCeSOzGRh8vgzLAHoOk0506eg4FQVFgEIlLqYsvyZZ8pLYy+9YpM4LR/VqeIScIveRK4/pwx+M470njXBeMd9VQkIrZBohCXgZMh8KwCgMgQ6IgoCqIRhTEEZiY+yXUgDNR8XPKb6LhmImIYGqrgZgBYchxeH4c4ONaWxWzJevoyGhRSGMPDoSoEk+rKzM887O9klzMKwgsyQyA+Y1753/0MpyBQlXwYAsllkLVcBhnhu0RERVmcPb/liYjJIAWdMLkByg3HIMyV16snXtN0GQgGgXhPXlE9jjp1altR8/YOC4PAth9DEIOA9XMskvs5iigEzUJUjyaIF1nYoTNTofgMc6Pjlb0sKimZ0dHRm0FVadRmUAZJTAQAVCqWK/fCcAGlbMyhkhZlsBEyBEUgqvqnLi447NCD8uUGgVtyHvk6ikJw0yWTXWnamGD515XFcNHkGu+2eBoExPezVz1ZQwUcMTQEpotAStGrCQYCvzevFcvkulL8+fZLWDuMIpwZ4Nc1GQIzvMwezTCinKnr77hqOgA2MUQKSTPpAx72d6ozH5eB/QHjTWYhVu7neOUhCArGEPiX4QOtPEmKosJeG0OgoCzGnkWRkcrbZaBzhiDYc+kG8ZK64VcXXQaiEM2uIfDZ3MhDQzAM7AGbf95PZCca4V6iQhmqQmyrd0rlxERuGgLje2LpfrYetbIfbmtmocnis5E7MZH9Hsx7Gg4+mwDg7cxl+Aw0QoOgCMibG/H/iquDYkSFMviGQqbfXbh4PitZceC678bzfEPgvNoiH85l+PAJOquxgTOqcpeBlbUQsPumAR7f7a8hKE9EzL0hTFeBwo0IK7oBsFYS3DDgkxMX/H3pLbMst0MBvvhcGFkRz89lIA3gpmvFZ6ArniEIsrGWF0PA2pfM2H2ioqhQZKQCMwTGXzHs1bVcgDwEdoaA2l0I8GYI/NqqUfu1+U8wHCYk8X68WCVAcEXqTIQYzGVgL6MJwl2TIZDOsUJHiW1xMnMU21uDG5Viu3MmJvIQuQ6H/g8C/mwNtdsgNAiKAGMInA8cgbVzWCH0j9cCLxHxYQjymATECSPXRO5laMgDcq7bFF0GUUVBLKIgldXNwdjTdSAwBF4TnbiSV03jQbF9lgcZPtFyoZ+YREg2KvoT9RXxvFwG8qpU3nXSDf2hIQieutjePj6B9mU027bUiaiVqbAiXjhDwPvD+7m0/u/17tk1BPZVpUNDIDIEPqtrXacQ7UfevmHhMhBYJr97EBmCVDaoqFDWDlGBAfCKMmB/iRB2CFh5O/gzJIp+uUFPCMuC+rk/bbBFTFibU53ZBkFvjmiKgUZoEBQBh8tAZAuElyJfeFHVcZMhcFrD+bgmxJVvLmYhKDWbaxLi96TpFKrhMkgKE4ZsCIj3mIshEFOsikmMxOty3zavn6+URhkJeepEg8CoT1799AfqKxlDcNuD63JuDQ0AaSmMha+U/DK7FWvIBEldzFeTNsGaZu2tkMxokqhQQV15DIum1GKh4KIqWFQYIOzQs+3SRCJ+ptSeqyIZUFQ4rF0GNlGh93MjhgBnNN22eveC/E6KBkFEJYCLhkBMLiW+uzysV35XAcsgUAjBq/ta8PimYzZjzVtUOAx+gADgzQwZgjMYsqjQzDMvWMWF7WXgfjwuMQTitfNxGYgGR67Jw3vglT8HYxoyGkXUcBmIK0gr/NC+AtYptSUmcm2j4mIQGAMN/4qvOjSZITBS9opuk4hkVPQnRlWw3QSX7z6FtU2nc5aXB3BuEPj5GotlCNQA+wG4GaVpTTf7OeUIO1QRVRX8+fZLcMlUK3wz6HjNHy/LZeBeLsit2w0AF8bAgyHwE1pqOnV1GQwHhiAbkCEwJ1Wdawhyd2ZUkV0GzjwEAKTVPPsruwxqypgriQuARVFq2jASFAL0pdn33UIUgqkhOENdBrydyZAhOHMRkxgCMSRMFNbkCyK8SCLEKANAylSYx2VsK5kCGQJ5/g/qMmCbFSmIqsSVjuWrGVseAmmSlyH6OuXJnBihTXwClTUKs8dUYMn0OiyaUmvVYdQXxIeaL3iK4FRWD7TzoTyA84xufiKrYrUPCsk9kLppCNJZ3Uwgw0SFdpcBh/hM5e8yMAyCgOGwbhCv6EiVLRkEvWlN0O7kYAiGaerifMMO8xIVSs8a23XQ+M7QELDj9jKAPcoAAGo5Q2Aa724uA2vc6BIYKK+9KM6UxESmhmCIDYIw7LAIsEyF9tUGAFsyjkIjSHj4nIi44W/TNPukBuSpIRBe4pwMQYGZCgFmMPF74NfkK4+oqtjijGXfOB8MNGoNrl7qefEerLJ21oBPVLKGoDIRxR/+ZbFrfQMhKuSbCAG5t7cGnK6BpBDn74WiowwCuAzcNASprOgysIsKE4KvWJzM8x2w+WrXi7kK8h6IoWkOlwGcVHVVSRTJjO6r0NepXfvDmzccVqg2DUGOSAlWJrioMCq7DHQrG6KqWLkidUqhmgnb+NhIbNfgwmD+jIvGMH/eCSyDwc4QOBkrfi9nAkJR4RsAcqZCjTqNg0LCDgF3YaHFEHDFvFi+MFFhTqrfk5olvp8BmHHn4jXN7YxVYqOUzUyFkope03WHLoDDYg4EhoD7/21uEcUZZWAMNm46gVwuimIwUthpMIhB4FDxc0rex2VQdJRBgMREmkukSzqre4oKxdrEdyJoal8+teQOO8xdl5yHQLcZCM4+Nw1xH+NF3reEt88vVfBgQTQCfEWFRjE+IQcRFcrGJxNp8u+IOS7ZusF0Gdg1BHw3UTlEGLCYMYUQc9IUDU4zU6GsIThD9jLgP1FvaBCcuWCZCq3PfLBiLgPu6yxscHYzJDh1KSfvAfKbBFQXVb4XvNwXQTQEYu56lqCEGnkIFERVxeYvkxkCS+DkvuoHrIx3rqJCkQVRiTPKwHgD3frNdBn0cx4CwHIZAEB7XybnClJmCPpMg8B74OiXKIOADEHWtprWzMkkmdFsA7Y4EakFMAT88eL0t9dr5bbBkAxZRCgbCPJvIuf/8KrTdS+DYTAfBXUZ8HeCs2lBRIUyi6YJokJVIa6bPIkaAvH3Ko1FEIsoFkMgPVu8TtNlIDIEOv8rGQTDwCALAt7OUENwBkMWFfIHmBDrBy7UZeA2qHNRoazIB/IzPESaL4gh4dYWJ0PgPI8nogFg7CNAzWvGJFEh22uAOsRqbPtj97ZyA8lNJBmR7jElKZf5ddzcAlaUQf8zBHVlcUFo5UxPLcNhEKQHniFg2x/7l3HTEDCXgZWpsDftXMEB9meHifqCD9qcmeivxERiZj3+nTxpBtEQaNTOEAwHUeG+U93oSWVtz5Bfbn/dZAjYbxhEVOgIO9QlgwBOhsBLcF0aUxGPKO5hh4aRQCk1DQLR4BRTT4sYDi6bjr4MDrT0+JbhhsxQawhCg6AIRFUFGoVtZz4GYvopC2YIXAY8B0MgigrzYQjyCDsE3O9BPuRWplRwGWR03TYJRxyiQt2+CYtAAVq5BezX4AaSLYzSQ0PAVx28Xj5IuoV4WhqC/n89VIXYUiTnchvIFC8PyfLXEBTpMlByT9IWg2PXEFgMgY6elIbLZ4zAey8cjzfPHeXZvnwGbW4MB4l+8Xr1ZMW7zdUH56TJnzM/94au29+BoU5dnNF0XHPPctzxx/W2Z8hv23BzlWoYm9Ei8xBEFMUyfkFtZQDnbpSlMRWJqOrrMsjqVNAQWMa0PCYOJw3HR3+zBlf9qNHRlq5kxhyHQg3BGwBcEJPR7RRXsWGHgD9D4JY2ttDEREFWk25lguQhKI3ZN7ExDQIj7FAcK/n2rGJ5wO6b9WIIRGrfW0NgMQRi2mK3VdBAJiYCrGRIQG6DICMNIsl0bpdBsSmX1QAuA1OAJgm/RIagJ53FlBFl+NH7FkhRBva68qF1rQE/N0Pgtf2x2KVi2mL+nWyEBY4yEO7L2v7Y85QBxdG2PgDAq/tabeLkXPsxANakFMxl4J2pkIUdOl0nlvjaXldJNIJE1HIZ2MSdwjE3DQEvahkbcNQxVNh0uB0AsPtEl3ksmdEw/1vP4cuPsi2PeZ8NNUMQRhkUAT6ZZDWKeMS+gjANggInlSAMgbj6yIeJECeMIOe53UOQTIXxiMo2ytGpbSOjiKI4FMxZXXfNqGZnCOzn8EkmYtMQKI5jESHsEGCDh7nSdLm3qOpugPQXJtaV4kBLD7pT2QAMgZeGYOAYApJXpkKrXG86aw7Mh9t60d6bwfiaEse58rOTz6SZyeEyCPIa+CUm0il13LuV/8O7oZpDVOi81mDiQCujqKtLohJD4JOYSOcMgWEQBBIVygyBkIdA9dIQ8MWSC0MQUT0yFVqsgWvYIXWOieI9DQc8sfkYbvntWlwzpx5/WH0IANC4i23qREOG4MyHyRBIyngmKmRlCt25yl1UaM9DUDBD4CLC822LSxnHXgZu4jzFSjxiYwhU4liZZzX7QGxLXewRZWAaBKJ7gDjbw1wG9iRIfLBxC63i5w5EHgIA+I+3z8ODty4CgJy5CLxFhQPpMnBml5Ph5rYS9RAnjA2cLppcCxkOl0GASZO/R5ppyLmXyz9Tof0efKMMfAwXcc8Nsb1DYRCc6Eziyc3NAFj2PzHKwB4mTfHtf27D1qP2zYR4kwPtZeDqMrDcOrwfxFBPM4urVD13GSQzLi4D4Xnn2iO/sEOLoRlag0DM1vnzZftwvDNpGgMAcO74KgBWu4eaIQgNgiLA6TJOydnDDv19nbngtiqPqnaGwEuolQvigBxIVFighiCqKuYqg03C3GXgZAgymu6aYpXliDcYAsmIKDEZArt7QL4vJiq0D4qiwFEGr28gRIUAy4o4d0wlgAI0BGYegoGLMuCsTpB2iZvldEoCyZKoinPGVTnOlZ+VfGjdXJsbBdMQCNd2cRnIBooVZeAXw49hk5joY79Zg7++fgSAczdC8R3r6Mvgt6824cO/Xg3AabwUu7kRS0zE/u+mIZB/w5KYikTUChG2uaOEe+CGgC3sUGJN5cyWQ4Umg6mZXl9uO/6WeaOwaEqtadzwZoYGwRmMmOAyAABNszbg0MyHvrC63WLhzR0Di2QIRJ97IFFhgS6DiEoEg8DKEc81BCKyugdDQP0YAufE7ZqbQMhDwOvmvnm3SX+gXQYAm2RKY2oADYH7JDSQUQbi8+sFPjlmhEQcncZAzVmhy2aM8GVgOPJZxZmbG3kZBAHuXQ47dGQqlIwweQ8R8dwvPrIRr+1rNVwGQjuIVWawsfO45avu6MvY2v3J37+OJzYfA2DdT1rK4slRUJSBYGDZExOJZdhfmT0tiaqIR9wZAjeBocgQ8Gs6ogyGMMqjoy+Dj/1mDQDgpx88H0989jJcPbseADBnTCXK4xEzEoe/A8nQZXDmgq9GzRh6/qDDygVfaJQBH9jiEXFiM2hTV1Fh8LrzZQgCiQpd4r8jimJODkxD4OcykDQEgtHjtblRnOchcMmrEJEMKdnYyPpEGZg6hAHIQyCitiyGtjwZAo4B38sgoIYgKYgbOUPwqYZp+M475uHH718QqH2BXAb8uj7aDyCYi842OemWgaAqxMNl4J6HIJXV8diGo/jg/65i4leXPAQDKWrbebwT87/5LI6295nHjgn/BwyDQPJ13PHHDfjW49uEideudOcoTENgMQQKcd/1kRtJ8i8VURVJVOgvLulyYQisXDDcxTREqk4A6w+2oaU7jQsn1WBGfTnOGVdlutDmjKlEaUw103uHqYvfAOBhORkpBIsQy+IrOA+BcWIiqpq0En/5+KBo28sgj8krn90OAXejxpGYSDgQN5KLRFUi+F8tDQHLVCi7DHw0BMb15URBCZewQ7cQRXnQykp6Bhk8T8NAuQw4astiOJ1TQ+BlEAzcXgZsYszFEBgGgTCAcbZj7thKvGXeaM9znaLC4JMm7w8vQ9tmmHrU4RQVsv9HFAIKp6hQFvNabRGFqnIegoF3Gfx57RF0pbJ4fOMxZDUdk0aUmZPtP++4DC/tPIl7X9jtmv3ugZVNeO+F4437YOfIxmeQKAPZsNeNzY1YUiJBQ+DCELj9hvGoahqZfiGSgLuGwHyWhoHLgOce+NVNF5rj3dWz6/HE5mO4aHItXtpx0szWyo3ioc5UGBoERYC7DByiQlLc9seANYnFXPzjbhqCvLY/zmO3Q7EtIuSVmHifMZUZBBGVYEx1Avtbemwug6jqHmVgS7HqIip0aAhizk2IvBITidB1atMzeN3vQLoMACb4ysUQeKnC/QbL/shUqFO2kvNacZsMgZAP4W8bjiKqEpw3odq3/kIYArNszsRE1v+92i7nIeCTSVRVoFPNke9fFvNabREMWEemQue1+ht15Sz3f2t3Cr9+5QAAYNaoCkyoLcGcMRVYf6iNfS89Y0vnjcYz247bMv0Bzt8h0OZG8m6HhoHF+8KtH/xCshMR1RZi6Ae/xES87qGK8jjU2ot1B0+jIhGxba0+a3QFnvzc5QBYnhaTIQhFhWc+ZJeBe9hhYXXzF4r7L8VjbhqCfK7jlmI1aHnrPO/PMb5yVxTcf9NCfGDhBGiCqFBViGNTlIxGbZOcZtKYVhsdGoKIS5SBh4ZAhE3P4Koh4NqEgXcZyIO1DJnuDYJiXR1BVrdeg/XbFoy15Vlwg2y8BqHVZZGYl82Tb6ZCURUfVYlr6mLOcslMhvi8is8pIFLlOZtTMPizfrwzaR7bdaILX3rzLERUxdwb4FRXynYeNyS6knYRqHx/gTY3cjAE1LbzI4GzH6w8BM7firkMDIbAx2VQEY+45iEwM8TCPlYONq744TI8teU4powo8zRMy2IR9GY0I302OxamLj6DIbsMrLBDAi8lbVCYBoGrhsBpEOTDELjl/veDWxE/hoC3OaoSlMUjKI2rNpo+qhJHFrSMnKlQ53+thC9yWxMuUQZeGgIRbL93oy1umQqN8kFEVcUgEENQwIDWH5kKAf/VldckvnhKXc765S4PkofANAi4yyBAHoLALgPj+hFVsaXP5rASgtmPi0yC7iEqHMgVKp/Qtx3rBACUxVTc94Hz8PYFYwEAVSXMIGjpthsEPFNmZw6GIBbJf7HAojSs988/yoB9Xjy1FvXGpl/xiGptVe7Dgo2siNtcBmIiM1t7hsAgENkQv/G/JKZC0ylSWWvsG+o8BKHLoAjwVa7MEFBQXz9ZEFgGgXMf+WKjDGz+9TwYArsrxLtObhCY4XsKMfIMcLrXxWWgUdec64yKVVzbamYqzFNDwMIOjba4TPqWUTGw9nJdeQw9aQ3JjGbL5CeiMIagSJeBYHh6NMtz5eWWiEhGMS4DU4cSIFOhF+TdDU2GQCHohbvaXiHO46LPXXYZmHkT+tkgWL2/Fa/tb8UXrp1pUv7cV/2zD12AqwwVOwBUlzImQGYIRhgMgRwmKk+eMdXjxxcgP2vcZaCYBoE3Q8C/e/i2S8zvRIbAb3U/oiKO/S09SGU1m1t1OIQdiobWHCO82A1lRibX3rRmJSYKGYIzFxZDIG/ZS4tmCPh5CcFlYIYdahadbpbPYxKI5Bl2yCdkW7iiT9ghVydHzclZkVblziiDjG5nCLKCxc/fd7kr+STqFjXhpyHISm2Rwe9zILY/FlFjDNjtve4bHLHnKP96+0NDAORiCNwNlQm1pYHrt+oK4DIwtz/OFXaYsyrP3Q6jEcU1yiCiEEQUxTG5iwYslVwG3GDpbw3BB+5fhfte2IOspttWyIB9J00AGFvFXDdNrT22d6e2jDME1nO392QXntt+wnZ+NABDIBvNnP423znOlLi4U91+wkTUYBM1pinyGj75NuLdyayrGNkMQxwCg+B4B3PhfP6aGfjGDXM8y5Uau8H2pLLmszXUDEFoEBSBqBBSBwgb8uhOWixf8IEkMEOQj6gwTyrc3G1QOM8vU6HMEERVgqww4atueQgcGgKrT63B1X5Nt7BDbuD4aQh0KoQduvhJB8tlUFvGKN3WnpTr93zCCaL2FlEsQ8D722+i9lp5ja7y1w8A7qFqQWFqCIJkKgzQDXKUgU6tULVx1YztiKiKa7ImuQ9cowwGKOrtRFfKQfnL2o2RFXFUl0aRzOiYWV9hHud7jIiiwmt/vAIv72mxnR9IVKjaXZvciFUlhkCE32KJL4CSBo3u1YaRhtujW5hMAeC2h1ieBXOvjSEwCJo7WOjnFTNHoDTmTcLz3WBf299qukhDhuAMBh+ozY1zeDig4JcsNHUxh11DIEUZFOgyyHfC4BOEeJ5z+2PRILCHA/L47rQwCcvphuXdDrMaha6zJCd8kpdf7RKXvQzcGALZLZDV/MMOB3K3QxGcIWjrcWcIeBtvWDAGX7h2RuDfrWiGIIAgzstYCDSJFJGpkBtJxeUhsBueVtZKtuEWt0tnjLKyy6mG20uE/Fm8dbcc/sVCFJw1t/ehK5nBrFHWRF9bGrOVJ4SY34uRH5zBk10GMgKJCo0xKeZpELBycrpo1j5nfXxMTWdZ5tK4Rxu44dmVzDoWCnf8cYMjlfFggjMEucS1fDfYf/vrZlPn0ZfRhiSZFUdoEBSBUsEHBAg0N6XFMwTGX98og35IXRwEfIIQBwinQWD9Py7tQsgnMh5OFFGILelJTGXiOVFVzFKg2v3F8ntiZSoU+8hJ9zv8nLp950UZJrMxwC4Drvb2ykXAWYz546rwhWtnBmJ2ePx3MeCX8aNbi1l5yW6qfKIMTJYpSOpij7rsUQYWGxeN8CgD1u98Mj3c1mswBPblvhye6JaYqD/no70nu83/H+tIojuVtWk23Nx/s0azezh3gpVCmi8yOpP+BkE8j8REvCylsBkElmEknuXNEPAIpTRnCDzaMMYwCLpTWdfnx23RNFho7kiCEKC+wt8gKHNhDygFlt73Mna0Dg1TEIoKi4BsEIhWabF5CPj5osvAL1NhPm6AfDftMRMD2QwCexk5D4HYJj5JcyaFZSq06oqrhPkMbXsZWCs3iyGwv9w87FCcuE33hs0gcMZK+2UqlNs9ULAYAneDgLtQLNeLYov7d0N/6B5MUaGfhkBaHS+ZXod3nDcu8DVECr4Ql0Gg7Y89ynilLo4oCnThGoun1eFXK/ajuiSGiOLcElrOBWEPO3Req1jwTYgA4HN/2gAAmFFfjrvfdY65mZSM2aOZqO3ccW4MQdb1HI5gexlwg8DI5kipTWCpmMa8kyFw3/+EHctoOrI6tbnLoiox+3xMFTOEupNZ1z7m1xgKl8HxjiTqyuI5Mz2K28OL2HWiC9qEuOt3A43QICgCJcYP2ufDEBQ6QJsMQcSdIZBXbwPJEFix/d6Drc1lIK3c+eTMKc+IomBSHROfffLKaXj89f043pnCb189YNah6bqjD+X3PhZxMgTm5kYuyYrEun0ZApf7HQhUlURBiD1xzL5T3ZhqxC5zFws3eArZfa4QBBIVSt999uoZWDw1d8ghh5nyvSQAACAASURBVEoINOS/isuVujjI7TsTE7H/x1TFtpfB5dNH4K+fvAQLJlTjN68ecIkysBtnImuhBujDfLFyXytGlMdtYYQViSg+fPEkz3Pedf44VJZEcM44S+3OJ28/hkAhwcYJ/rzxd5FrMiyGwMmU8LHLrXpeTyrLdiQVhY0xVUFGY2MIZ9e6U1lfncZQJCZq7UmZokc/lMXt029ZTEWPMZeURwd27PHCsHAZEEJ+SAjZSQjZTAh5jBBSLXx3FyFkLyFkFyHkLcLxpcaxvYSQrwxFu7lgxGQIXKIMCmZvTVGhaCEbfjrNGSs9oBoCaUJyuyfx+jGJcldNg8BKBjRvbBWavv9WfOW62ZhSqWLF7lNYua/VrENkCLzo4ajERIjXsoUdShN7ZzKL57efsOVaF8GNioEWFUZUBVUlUZMh2H2iC9fcsxzPbjuBU10p7D/FKGKu3A/Snv5gCHgdfgNtIRvhiBDJlyCDNq+9X8IOzRUqZ/PsehJRp7Bwci2ihubFoSGQ+sCmZ+xnlwGlFCv3teLSaXajqzzhv6Yriam44dyxNgM+FsBlEJRF5O8Kf+d1riEwExNZ7ecw+8RHQ8AZArEdovugwrjvrpQ7Q8AxFAxBW28GNUZSKD+IEWSAFXUAAGVns0EA4HkA51BKzwWwG8BdAEAImQvgRgDzACwF8AtCiEoIUQH8HMB1AOYC+KBRdlChKgQRBejNMOrNCpXzV9IGgaUhsGglPtZnBYPDbEteUQZ5ugykSdbtnsRDlqjQPrGaLgNp0ppe7WyPplvCTNNlIN3z2OoEIgox1eBi3X5hh1/68yZsO9bpOVhHB8llAFj7GRxs7cGynScBsB3pLrr7Baw5cBoAzC2Eg+RF6BeXAffX5zHQFuqGAqwkVEHAV+UkQJSBp4bAJiK0tCr8eU1pukOLobq6DCSGwKYhYH/7SyC271Q3WrpTWDK9Dp9qmGYe70350/4y4hHFXGTIqYtFBI1skTOqajoTZTpcBsI51Gds5MZK2mAIxHaI/6+Iswm3O5nNy7U1GGjrSZvuQD+MqkhgwXhL21EhGATlsaExCIaFy4BS+pzwcRWA9xr/fweAhymlKQAHCCF7ASwyvttLKd0PAISQh42y2wepySbiKhz5qHlyDqBwg8CsX7CKCSHMl6k7N2DJZ+7KlyGQUwe7na64rEDkidViCOyNnV7j9KWJokJ+XfnVnlRXhu3/sdTmqwsSdpgrXbA6SAwBwJThT25uxpObmx3f/fSlvZg6oszMOFdXHrPtbOeGfnUZ+KyuZIFdvkmcRGYmkMtACoX0ZAgC5SFgfyNGqmL+2Uw0lqWO+t3CDuXPbinB+4uy3mLoB86bUIMPXDQRCyfV4NbfrUN7jkgBEX/79KUYW1ViMiF+UQZBdjoErHfEZAg89jKw6TaMv+4aAjtDYBMfC/9PRFkoaHcqk5drazDQ1ptGTVluhkBRCH74vgV4870rANhdCIncOaEGBMPCIJDwcQCPGP8fB2YgcBwxjgHAYen4xW6VEUJuA3AbAIwaNQqNjY392VbEFIr9h46isbEFx46zcJP29g6sX89EP5s3b0L2aP6/bmcnG/iPHj5oHmtsbAQBxf6mg1i+4pit/JrVq7A3Eewlbu2zBvMg/dHexu4r1csyolFKHee1Ja06TzYfBQDs2rkDje17sPcIG3h272MagTWrXkON0NYq9GLp5BieabIGqAMHD+FlwibJfXv3oDHdhCNH7cIpt7YfOcwm+107tqOxbTcA4NhRbwPArY6DB1g79u/dg8ZUk+e5/QGtL+n7/aho0mxjLONfFgCymUzRz/juo+z+X1u1CvtK3Z+plla7YbJh/Tqc3B3cKNA1a3W6YeNGpA77vyNdxvvQ1dMLAHj1lZfNzcVEbD1l1dvT2+PaF9sNBTfVNXR2dWHTpk0AgPY25rJqOnQYgG47N5XsQ/PxlO3YxhP2FfbePdbzcrKXvQ/btu9ATcde33sLgmd2pBBVgCPb16F5JwF0iuumRHFFVVtev3cngL4s1zoB5VGg28Uu0LPBnqPjPew+u7s6QADsb2pCJptFX7objY2N2Hac9dGaNWtxvJL9xltbWP9v2rgBvQftv/vO0+y7Na9vQEtrGuIu39m09fwvX74cCZVi576DeJXax0IRR5ub0djYlvM++gs6pWjvzaDjVDMaG1tzlm/utm4w3dtl/r+nx/3ZHWgMmkFACHkBgNueqF+jlP7DKPM1AFkAf+CnuZSncHd1uJqClNL7AdwPAAsXLqQNDQ35NTwHEi8/haq6EWhouBB/ObYeaG5GeUUFzl0wB1izCuefdx4umRZcbMVRvuVloLMTc2dOx9/27AAANDQ0IPbSMxg3fjwuuXQ68OLzZvnLLr0U9TniXjlOdiWB5S+adebC7w+uA06dQE11JQ50tiOiKo7zTnYlgUZW57Qpk4CD+7Bg/jlomDcara8fAbZuwqix44H9B3DFZUtQV26JbhobG/HLTzZg8leeNI+NHTceiy+ZCix7EbNnzULDxRPxQvsW4NAhs4xb2zdmdwP79uBc49oAsDq5E2ja53pvbnUciB4Adm3H3Dmz0bBwQs7+KQaPn9iIDSeP4pxxldh6tNM8vmBCNeaPq8THLpmMGUb428reHVh3Yr9ZJh5RzLzvHGUliUC/qR86Nh4FtmzEwosWYerIctcyP9uxEmizBtoliy/G5BFlga+RePl59GSYoXbO/HNxxcyRvuXv3fYq0NGOaCwO9CXRcOWVrqtYsvsU8PoaAEB5WRkaGq50lInsaQHWrkYiFkVpWQnmnTMTeH0dRo+qB040Y+SoMYifbLb1Y+WG5agbUY6GhgvNY92bjwEbNpif+XMKAEfaeoEVyzBzVv88Q7/c/RrmjtNxzdVLzGPXXF1YXamsBrzwDACgpqIE3aedrFN5WbDn6PDpXuDlZRhZV4u97a2YOHEi9rcfQHVFJRoaliC17Tiw8XVcuHAh5o1l9DjZfQpYtwYXXHABLpxUY6uv4mAbsGYl5p4zH42n9kFRgD3tzHVWXVGO4z1dIIS9t9WrXkLViFpcvHgW0PiSa/tGjhyFhobz8+meotDemwZ99nmcN2cGGi6bkrP8wdYe4JVGAMD4USOwq425DcvLy4t+jwvBoBkElNJr/b4nhHwMwA0ArqGW4+0IAPFtGg+Am4NexwcVcZWYokJNyFho+ckKq9cKO7QPeqpCmOCOyi6DfESF+dG7cqIeN6pP9XEZcIpSjDJww9wxldjezCZF0S2Sj3vaLYmSWz6Bb75trqePWY6OGEjwneo+cflU3PW3LeazdMdV0/GmuaNsZcdKWQBLY6rDIOgPDQEJQHc70vvmLSrMz2XAS/PNnrxuM8jtW9EriimCA4SkOJruEKIqhL136ayOqMr0BX6JidzC7QrFPc/twqr9p/Fhw9goFqIvXgxrFhFUEyLqixRCjCytonuR94N1jl+OlrioIdB1lEUj5h4qfFzhCckqEhFH6mIZg5mHgFJqZnsMIioE7GNpaXzoCfthISokhCwF8GUAb6eU9gpfPQ7gRkJInBAyBcAMAGsArAUwgxAyhRASAxMePj7Y7QYMDYEUdshfCqCIsEPTILC/sBFVMQR3hYsKCw075BO8m0HgmqlQShIkRhm44cFbF+EH75mPurKYLdWx26Di2VbTCHEmKxLxprmjcPMSdws+6qJDGChMqmOr6gsm1mD7fyzFo5+6FJPqSrFocq2j7Ogq+8ZBXFsQs+kl+iHKIIBC3hllUIyoMPigLaa/doNdVOhexr7dsSXQFcNjZaM1ohK0dqcw8+tP4/9eYa4vWVQoihD7MzHRE4a+5COLvcML8wEhVmKwRNTaJllEUFGhJeJVoCiWhoDfP+8ROfcDb4fjusIOsppOETEMDbFNPH6/3NgC2W9ckJNHDSSe2XocnzXyQwQRFQL2sbA8PkTCAQHDwiAA8DMAFQCeJ4RsJIT8EgAopdsA/BlMLPgMgM9QSjVKaRbAHQCeBbADwJ+NsoOOuErMPAT8odeFFXyhWeP4My7ToqrCknPIDIHbrn1eyD9EzHrpAXc/jjgQW7sQ8pW2PTGR12A+ojyOD1w00RRwyYN/kLHVLcogHnU+5nwydYNpAA0CQ/CNG+bgsU9faoYWXjipBsvvvApVLoP0GIkhqDTuwdw7QiH9FHbI/ubay8DGwhQYuQLkJ/zKaGzDG6/3yh76514HFYx1lmbcHmVwvDNp7ghotVfBodNsrfLYBqaRcYT+2gwC9rc/RIUdfRl8ZPFE353z8gV/ZuIRFfPHVTm+DyoqFCOJVEKMsc+a6LhdZWMIePSQy+8TNVkaDRmN7XTKy1lGjGEQJCKOvQxkFLBZaEFIZTU0d1gah5qyYAaB+B64ZS4cbAx9CwBQSqf7fHc3gLtdjj8F4KmBbFcQxFSgLS2FHfZH6mKXPAQAjCgD3alwHkiGwCjODQm3S4kLqosm1+Lfls4y/YN84jjS1mfU4z/YmNsl50hM5AZumIj3eO5454Dn9/JZ+Q0G3l4ujUVw/sSa3AUBTBlp99GXGxRjPKqgK8UGzP50GfgZBLpOkYiq6DbC3vJ3Gdjryt0mmG3yi9wJlofAeq54ql0ApkjxWHufY/JViZWAzNzUzCfKwHS7FEkR6DpFe28a1SXBJpigiEcUdBl/f/bBC3DvC7vxwMom8/t8XQYRVWEuA5khcHE/+UUZmAxB1mIIWB3UfAZEhuBQa28Ol8HAWwSUUsz6+jO2Y0FdBiITJScqGgoMF4bgjIWNIRDCDv1ibfNBQtqQ3tQQFJWYqLAQMX6em17BzhCo+HTDdPPl5uzFxsPtRln/66kqsblFeN0fvWRSzvzqbgzBBS4Trp/mgk9ug6EhyAeViSh2fmcpPriIyWdKYxF85brZZsrgeETpl+yKQbLsZXUd50+sxtsWjMX180ejPM/Vjc1lkE9iIs0ZEigin8REnCGQ8xCc6ExhtCTQjSiKuRMdp67lTIXipc3kTkUSBN3pLHQKV1q/GHC3XjyioKo0ig9cZBc+BmURebmoQqAIBhZ/dyyXgXWOX9I2M2eJpiOj64ioxBwv0oZeRtQQdKWyvjqNwUhDIO4xwVEIQ1ASU3HehGr8/EMX9Fvb8sXQmyRnOOIRoDfDNQTsgdXFpDpF72XgxhC45CEYQIZATuXrKioUMwNK9csTay43SkRRbMJJXt+cMZXY9d3rsHJvC1o8cgmoLqt72ajKhcFKXVwIElHVNMzK4io+eeU0vLavFQ+tPICRFfF+SaYUZDLTdIq6shjuu7EwBXfeeQgMZHXdl84OsrAVNQN9mu7IVAg4d6rjO3YCMNPpOkWF/e8y6OhlMYGVPi6uQiDT7/I4U5ioUNztUBYgOzUEbuNIXGXtyRibG4kagikjyrC2qQ23X8kSM5XHI7kTEw0CQ/DKXmvb6JEVcfzgPfNRmQjKEAjjJiH4+2dYFElj467+bWRAhAZBkeBRBie7kjhpbDCi0eJTF/ONfGT/N2cIHJkKB3A1K6cudruUeJ/y94WIGN00BByXTh/hea4bQwAAz3zhcuxs7sIXHtmY8/rmfgiDkKmwEPD+4NTpJdPq8PNrS/GnQ6UmW1UMzM2NcmgIijE+xFV+PpOmmBbXDUE0O6YBoCjQdc0SGQr3M7rKaRBwmMlz5N0OXTZWKpYh6DCSB1X3s0HADQCeAjguGc1BdjoE7HuH8GyOOqWmYeYmrvRL2saNrbSx2Rl3RQBMY9T0/beaZcvjUfRlNGSyQxtlsGq/lW9gyogyXD17lE9pO2xG5DBgJEODoEjEVUZlLbr7RfOYrvs/9PmADz6WcEyBplGHWGYgnyV+D5aGwN9lIN9zIWGOYurivPQRLhoCgO36VhuQxotKYZPDDbx/SwWaPqoQfPNtc/tlAAyyutV0Gmg17gWbqDDAIs5Nwe+GfFwGEZUYux1anzkcLgPhO8tl4M3S8dsrNuyw3WAIqgOq1oOCb5zFo1zkqIJ8GAJC2LuiENafzGVgCJD5syQ8l36LJXMvAyPsMGLUz68lgu/j4Lcng/wbDQSaWqzAuHx/b3HhMgzsgdAgKBZxl0lD3DCl2LBDVSH4zjvmmcmNOEMghtMopPBohiAQhUP8eo4y4mAoFRD74Jcfye0fi6hspz/ZZZBXW13OCWqYXDyFiSIXTKjOXXgIwG+tTNo+lQ/uxcLUEPgYFx19GZTHC1+1ipNnvsI7v5VUfnkIeOpiHoZoPR+yy0CkgLnB6NjtUHEaLcUaaO19zDXW3xqC44YifoqRTEpmIvOJGokohIUdEoPZo9ZvxH9msRfEzaVkcAMjremG0WkFj8rGHs/975eCWQ4N7W9QSnG0vQ/T68ux92R3zu3JZfi5WocCoUFQJNxCR0WhUqG/sS7QmjddMtk8HlVZlIHIVg70g2QyBIq3hsAmqJK+7zIs+CXT67D0nDE5r1dbFkdrT9rKQ5DH/Y2pSiAWUVxFPUH7iYsihyu4ur1kgMKUTJeBx2qnO5VFb1rD6KrC92zPN+xQ/OX8fkfx2fSqlh+PGKJCahoE1rn1lfZ7EyfkmBBlEFVZGDC7trONxRI2nCHwC5MtBDx3Ct+GvFCGAGDG06jKuGBgORMTBc1DQAhBTFWQ1nRkNLbbIT/TiyHo8DEI5MRd/Y2Ovgy6U1lcMLEae0925224iX0wkIu6oAgNgiIxw2VjHp6cAyj+R3bzx8uZCot1S+SCJRzyzkNACBMVsZAj+3cTjRj7j/js2y5iZHkce0905dzIxg2XTKvD+m+8yQzJEyFvuHKmgmd8LBugRCbW5kbu358wsivWVwRLle16jQJFhYD/8y5+5eXysDMEgstAYJBkn734mU92WWMFaxoEtrBD/zYEBZ/s+tsg4JhssEqyZiAWCf6SPP35y5GIqnho1UEz86OcmEjsBmoyBO7XiKkK0lmLIeAWgcMgiPu7DBRiuUYGCjyU+qpZ9Zg9uhI3nJt7weOFfMa5gcLwVE2dQZhQoeCu62bbjokug4LzEBh/5WckoiiOKIOBDo/j9ftpCABn7DHH1JHl2Hv3dbhufrCXZWRFHKe6U6ZbJB8JAiHE1RgArPu4LgBLMZzBVz2lA8QQmImJPCYz0yCoLIIhCDBxe8HvcbcxBB5l+OWiqiK5DKxz5RwUog+fv3sZTbcJEe2JiYhxreINgkRUyTtSJih47Lv8zgbNVAgAFYkoooKoUBMZAoX3g1U+V46WWEQxUxdHVOKZtyAXQ1ASVZHKk8LPF9wgGF9Tio9fNiXwfjJuGA4LlZAh6AfcfuU0xCIKvv1PtvuypouJiYq0CKT1uFuUwUCrUxXTIOB5CDzKGQlE3CjdfJL81FfEkdEo2nrYi95flnMiqqLxSw0YU134SzscwBmCkgGaJHJt3Xuqi0XTyH72fKDmyRC4xfi7waZNCMgQWGGH3s+oSAXzVSdTwRMzKsZNWFusy+BYe1/gNLj54JHbFuNkV8rz+3wzTwLsPeUuA0Vi48TfIpfgOqoqbPtjjYUdWnosezlLQ5CVqwDA4voHmiHg25GPrynJUTI3wiiDNxBESk+nKD4PgfFXPr22LIbNR9tt6tmB1hCYGwb55CEADENBK97SHVnBVp5845/+vL98duQbruAMQdDQsHxh+r89ZjPLZVA4QyA+Q/3pMhAfFa/FuRlloBipdnnqYiGmXoZoEHChGlvBKlAJgQZqM5R5FcWICjv6Mnhhxwm86/zxBdfhhYun+u/AGi3g2SLEyMFCxbTWTuMyFyMUiyjoTWtGemylYIYgEVXRlXQ3FvoLh0/3oiym9ovoc6Bdv4HaMNQNeKPAbhD0Qx4Cfr50/PyJ1Th8ug/Hjf3hVYUUvIKeNjLY5OhgCHK4DIq1dE2DoKOvX+p7o4HvCeG2R0N/IJdC/kRnCqUx1dM1EwTipBvEZSA2xc+FJFLfXnS9Z+piYxKsTDjvq0pIHcyN8YxGEVWs/SPE95AQYuzSV7hB8MzWZiQzOj60qH92OcwHBTEECt8syqnXEXvBdKd6vNdRlZhZTWeMKrdFXIngkR9tve5JyhJR1XxXBgr7TnVj6v9v78zD5ajKhP97q7vvkn3fQxYIZDEkQAjRkOGCBCJ8CqJ+4sq4MSoqM+qIDOPogKjj6KjMODqMOuIz+rngAiqKIF4UZI8hLAkkQEL2Pfcmufvt8/1RVd11+1YvVV3dXbfv+3uefrr71Hb6dFWdt9518qhIHAKrkCm9KDHoQn3gzSRm+xDYn8M+3V5xpv1UMHHkwKcwtz7A3/3oScCJ/w1xjE03ruOua9eUtG5uSeF85767XrmSblZD0D3guIqNG9pUKbtyMXX33vYupo5pKusmGDQPgXdiLZy6OPs5X/+9iYm8ScRcp0K/rID+JoN0JiEPDJ7gLJGyTAYvHewglRCWzIiuqFGphNE+uWGHAwWCwb4UWadC//00JBMZ2/yKORMySdpy76UjG5OMaUpm1s2lOZXIpDuuFC8dPJEJ3SwX1RDUEbkagv4yfQg+fMEpbLpx3aCqd0tmDCzUk7KsUBqC5oZE3lroubimgtyLPBe3udz5e0quhiAGF0qccH0IKm4y8Hm6PdrRw/3PHfCtkBcEywkvy3ecXLxrFBKAS9E8eBMTebV5rlOhn0e/146fNRmYARUmc89TO+omvESwr72LKaObaqIhC5OUyxaABgoE2cRE2fWK+RC4RaZmT2hm2timglEJs8aPYPuhE777aU4lSJvB+SKioqu3n11HO1UgUAYzNkdDUG61QxGhuWHwhN2QtPjx37ySq/9qPqmEMHN8c9XyELgq1XyHs4oIDKUyqjHJzHHNPLmzDYhHwo444Q5v5TQE9nuuycAYw+fu2sTx7j4+eP7JZR0jYUm2kE0JT3HerpRa3Ci/D0FuYqKBx/DLQ+81I7hhhl6nQvvYA7eRMjUEe9u6BqVQrhbhTQZ2QaFBGgLPesXMqe59ZtmscQO29bsPzBrfnHeMm5z7Z6UcC18+3IExML9E02sx4nCfU4EgIgY7FeZPvlEuK+dN4B8uWcTzn30NM8c1BwrLC8PCaaNZMmNMJsxNfDMRRGcyEBHefPbsjLovDhdKnPj6W8/kgy0ns2DKqIrs38qjIWh97gA/fnwnH2w5mYXTylNjWyKkkhYTRjbwn3/YypZ9xwpvUGLeDe+i/D4E9rvrwe4KPm4CLT8NgTcCoc/rVGhlc+3nnqdRaAhyUyhXklvfcRafdEKowwgElmSLG+WaF30TE+W5j7irzp/snN+u+TWPhiAfzY6PTaVCD593ztnoNASR7Ka8PtS6A/VC7tNaJsteBf9kEdt/oNIJLc5fOIVff2SNJw9B/v5ANBP4m8+eXdREMVyZM3Ekn1i3sGKZzfKVP35022FSCeHaCxeUfwzLPs63r1pBX9qweW9hgaBUk8HAsMM8+8poCKwBOUOOd9se6fkqC/7thQtIWkKP16kwIYOy8rkkRAKnZfb20fXVqBYXLZnGeadOBoLlIXCx3DBOvImJBvsQFEpdDLDHSas835loXR8Cv/+9ULifG5ZbKQ3BfZv3M7Y5xaLp0fh4xOE+pwJBhNzyljO4fPkMIJtettJ/ctIK51QYBquIBsDtRhTdmTqmibWL7KphqiGoLu54595Hn9rZxmnTRpfse1LsGJYlGadZ1y8iH17hpNBcNTAxUR4NQTrrM+A1GVy+fCZLJlp8+AL/tNV/e+GpXLxk2uCww7w+BOFNBsciSA8dBve3FCoxnQ9LJKPVyyYmspf5Jybyv653O75Dbohw1kF78LoFBQLXZFABx8Le/jT3PruPVy+aEkqb4kccBALNQxAhr1s2gx2H7cpX2fzmlf2T33DmrEy8fqXJhhX6L3cLk0T15HrN+afQ3tVbVbWp4l+hzhjDxp1HufT0GZEcwxJbs9XkqHV3H+3iqZ1tLJ3l76zonVBKzUOQbzJOZyYYe2VXeB83IsXfn93MjHH5J5lkQjImg14ncY57PeQKrlKGyWCf85RcTQ0BwOzxI7hg4RSWhyjslRDJCEu5/kTe/6JY6mJ3+byJroYg//qFfCxcwbUSoYdb9h2nvauPltOmRLbPODz4qEAQMZmbTH/wtLthuHBx6bW3yyUbRZBPQyCRCkBLZ43lB+9bFdn+lNLwizJ43rkBnp5nwg5zjIQlNDpq3f9s3cptD21j/afW+q5fqkAQJA9BMqMJSQ/aNh92Fj3jbGdoSlmZcMXcB0VXhR4GNxtkOfUiwtDckOA7f312qG0tC3p7B46tO6J+iYmkyL3RjbAqVDm2UBZHV0NQiQJHO4/YD35zJuT3YQhKDBQEKhBEjWt/zUjKcfiXI8K9HvPWMrDi4RijlId7DntrGfzosR2kEsKFi6IRQE+aMII9bV0ZDUF3X5re/h6MMb7n10CTQWkagmLVDi2PhqDUp7OUU4kPnDwEjcls8p2c47lheGE46mTfi7rscSWxPBqCbNjh4CiDYhqC33/svIyGxLutn6/UxFH5BYIRqcqZDHY7KYtnRpCy2EU1BHWIe5PprZIPQTVxL+68YYcRawiU2uD+z67JoLc/zc/+spOLlkzLJI0ql4+uPdVzPDIZA7v70kXDKUsNOyxWyyBTtbDflCzINiSyk57rVJjxucg5Xjk+BJWuclgJEpZkhKVBmQr9NAR59nPy5FGcPDkbQZMrwHkpVM+jqYICwa6jnTSlLCb6lFkPSxzunepUGDG5Mdwx+I8jI7ekqd/yOJzUSnlkTQb294dfPMTRjl4uWxaN/wC4qX3tV5PHSbGzx9/e651rC51ipUQZZMMO7dtfX3+6ZL+XZMLKpC52ww4Tmf3kCgThUxcPRYFAPE6FuRqCIMWN8uEnCBb635oqaDLYdbSTGeOaI430icO9s6CGQEReIn8V0QEYY+ZH0qMhjnsh1LPJoFCUQRzUXkp5ZEwGzp377mf20pxK8FdOSFrUNKUsOp0og87efsb768prmQAAIABJREFUrFOqycBrly7Vh6A3bUoO3R1gMki7iYkG9xEcDUHIuaits5ekJYzwSU4WVxLCIJOBnzklbJ2XUu8tjUmL7r50NuywEgLBkU5mFnA+DYKrIYvDvbOYyeA/PJ9HAR8FHgUectpeCawEvhx914Ym7mTZl4kyqGVvoqV42KHUlUZkuGJ5JrhjXb3cuWE3FyyaUrHMiPZ+7SfijnwaAs/nQjdO78Se34fAcVJz8mr0hzQZ9Du+B4unj+HpXe2Dij1ZMtiMUCptnb2MbU5VLNdEJUhYknG4zE1S5qetCfqwVCy8+o5rVjN7wghe++8PZFT6EH0egnTasPNIJ2sjyj8g4ORuiGR3ZVFQIDDGZCZ6Efku8C/GmM951xGR64ElFendECQbylR/GgL3p+T7SQlLTQb1gLfa4f88uI32rj6uXlM5BaC3JkN+k0F2Rik0SZbiQ2AyJoOsU2GpuTySCQtj7LHp67cFghsvewVXnDkrm1nP08+wToXtjkAwlBARevNkFx1gMgiZtK1YkbNpY5uYMLKBiaMa2HW0M6MZiFpD8JMndnDoRA+ripSQLhXX+TQOVV2D+BBcAfzYp/0nwOui6c7QJxtlkD+71lBFimgIRCQWai+lPNz/8PHtR/ja77dw6dLpLAsRl14qXs1DR49//XrvvFqo7s6A1MV51snmIXBs/+l0yYKsm4Smtz9N2th5CJpSCd/JwS2vHIa2zt68GRPjSkJ8nAozNoPsemF9CPKt72ZQdZd/5AI7k+aCKaOB6PMQ/M+D21g2ayyXLY/Gp6ZYOHc1CSIQnABafNpbgI4oOlMPuP9pX386FiqgKMn4EOQ5ayyJh9pLKY9UwmJsc4p7nt3HxJENfOENSyt6vEavQJAnY2GpJoNKRxm4k09Pf9oJV8x/Cy2nlsFQ1BBYVn4fAt88BBH5EIxttj39XZ+XCxdPZdsXLs0kLYpSQ9DTl+aFA8d51SmTIjPnZFK+DzGB4CvA10XkmyLy187rm8C/O8sUPE6FaRMLiS9KivkQqMmgPkhYwnXr7CI3H77gFEb7VP+LkqaAJoNCqWJLyVRo5zrw5iFIl6zZclP69val6XfKH+fvS3lhh0NNQ+D9vZkoAwb7EGSKGwWNMsjzt//rG0/n5MkjmZATAuj+V1FGGbx48Di9/YaF00ZHtk93FOJw6yw5D4Ex5osisg24Fvi/TvMm4CpjjJ8pYVjizVRYb5Nj1ocgv8mg3n7zcOWt55zEqvkTIqvkVoiBJgN/gcA7oSQLCgTFbQZp44bI2t/7+v2TIfmRCVVMG/r6CwsSIoQubmQ7FQ6tNDHesUjmaAi8dSXs8Q++/3z3lvMXTuH8hYNTCDdWQCDYvMcuwlVutU8v7s+Kg7m1pDNORJLARcDvdfIvjDfKoN7mxmJ5CBIadlhX5DrJVQrXGxzIhB/m4p1QUgWcCLzXXCGTgW3e8jgVBjUZlKwhCC4QGGNo7+obeiYDz+C7ppRsHoLseoZw2tOg9xa3YmOUJoPNe4+RSgjzJ0cnKLtalDg8TJVkMjDG9AE/A6LTk9Qpw8NkkH95nf1kpQp4NQSdJTgVpgrY7cVzDuYXCFxtlv29L50u2X6bMRm4PgQFhJNEyFoGx7r76E+bIS4Q2O9uU2754zD3xjBhiqmERBp2uHlvOydPHhVZhUPI3k8j3GVognThScC/LqiSwVVTdXT31Z2DXfHERGoyUILjzVSYNw/BAJNB4XPMPQfzTcUmR0PQG8Bk4E4EfWlTVEMgIX0I9hy18/hPHxtdnvxq4B0KV0OQ+S9yEhOFuU2E0T42JKxINQTP7T3GoojyD7i4514cck4EEQg+A3xZRC4XkdkiMsH7qlD/hhxuudI9bV11FXIIxX0ILEtNBkpwGlPRORXCwOx4ftkKbZNBVnjtT5uSq5K6AkBPn6MhKJgTIVzqYreS3qwIC+dUA++1n8jRJno1JcaEc6ALc28Z1ZTMpIEul6MdPexp6+K0CB0KIWuCjUOUQRCvlV877z9joPAtzvehk2Ozgkx3Ql12He0cciq/YpRS3CgG57QyxCjFqdB7wynkQwDuU6m9hd/kk3EqdISAvv7STQapjKOa3c/CYYeSCYULws4jdiW9WeOjK61bDbwPQIWiDNIhzalhtpk7cSTbDp4IvJ0fm/e6DoURCwRDzanQ4fyK9aKOmDCygYakraaKwf8bKaWEHcZBylWGFt6ww/xRBtkZpVCUAQw8P/2mY1dl7a7X3Vd6RND4EXZo2772bqcvhTUEYUwGO4900Ji0mFSgtG8cGWgycAQC56+KwocgzIQ5f/JIfvfMvsDb5dLXn+ab97+AJbB4RqVMBpHuNhRBwg7vr2RH6gURYfrYJrYf6qg7e3qxxEQadqiEwZuYqLO3BKfCEk0GYE/+iZy4GGNcbZbdvnnvMdYtmVZSX13v8uecp8WCSZJCOhXuPNLJrPHRVtKrBt6HgWxiosE+BIaQPgQhNpo3aSSHTvTQ1tHL2BHhNbb3btpP63MH+MxrFzNldFPo/fgRJw1BYL9GEZkhIqtE5K+8r3I6ISI3ichGEdkgIr8TkRlOu4jILSKy1Vl+pmebq0Rki/O6qpzjR41rNhhqF3QxsmGH+ZwK4yHlKkML12SQsCS/D4Hnc6rIjbNYtkI37NA7wbxr9dyS+jqmKcXk0Y08v88WCIqFHYZJVGgLBEPLXAAD73dZk4FNrg9BKJNBCC/8+ZPs0NkXDx4PvrGHu5/Zy7gRKd6+ak5Z+/HDHYk4PEyVPMSOINAK7AQeBFqBP3he5fCvxpjTjTHLgV8B/+S0vwZY4LyuBr7h9GUC8GngHOxqi58WEb+qqTXB9Q6OgcAXKVmnQv/lzakEzUOoXKsSD9w8BONHNBSIMijdZDAgN5HPhJx1Ksy2rZxXul/0yZNH8ty+EjQEEi518Z62TmZEVFq3mvgnJhoc8eEKZOXsv1TmORqdl8rwI+jpS3Pvpn2sXTS16LkXBitGJoMgv+6rQD+wGLt2wRrgTdjZCteV0wljTLvn60iy589lwPeMzcPAOBGZDlwM3GOMOWyMOQLcU24foiSrIahxRyKmmA/BdesW8vkrKpv3Xqk/ls0ax7LZ45gzcUQmF34uA00GRTQEnonD3W7bwRPM/eSvueupPZk8BO4T7ajGZCBt3smTR/HiAXuCKRZ2GNSpsKcvzcHjPUwbE61auhp4J+zcSS63lkEY7WkYk4F7L3Z9PsLw0IuHONbVx7pXlGZWCkqc5okgAsF5wHXGmM3YE/YBY8zPgOuAm8rtiIjcLCI7gLeR1RDMBHZ4VtvptOVrjwWTRjUCcKzL3x46VHFP3Hz3wLmTRkaa0lMZHrxi5ljuuGY1Y5qSmSqhuQyMMijdqdCdiDbuagPg10/tydYycFYLmnXOq84vVtwoqILgwHF74po6pjHYhjHAO7G5zpbZxETZZaFTF4fYaERDklGNSQ4cCy8Q/PbpvYxsSLD6lEmh91GIjHAUsu5FlASJMmgGDjqfDwNTgOeBZ4HTi20sIvcCfiLWDcaYO4wxNwA3iMj1wIewTQJ+Z4Ap0O533KuxzQ1MnTqV1tbWYl0NxPHjxwftc99uWxDo6OmP/Hi1ZNcx++lt//79kf4uvzFUglEPY9h2pIsjHWnf39Hd05P5/OLWLbT2bMu7n77e7Lp//NOfaE4KTzvX5KED+zmWEHp7+nn4L08BMLI/O3aljOOObdm49q1bnqO180Xf9drbOuntJ9D/svWIbTLZv/15Wjv89xtXdu3Ijvvjjz3G7lEWfY6G5IUXX6RVdtrr7eqmtzf4vfHhP/+ZUQ3BhYKRiX6eeeFlWlv3B972WI/hF+s7WDY5wcMP/inw9qXQ65zbD/75z4xvsgXMWl3PQQSCzcBCYBuwAXi/80R/DbCr2MbGmAtLPM4PsHMefBr7yX+2Z9ksYLfT3pLT3prnuLcCtwKsWLHCtLS0+K0WmtbWVnL3mdxykP/a+AjAoGVDma37j8GDf2T6tGm0tCyLbL9+Y6gEox7G8Ce719O2p933dyTv/x302hPxksULaVkxe9A6Lo0P3gs99hPh6nPPZUxTikNP7ISNTzJ92jSSltB87CAnzZsPTz/L2rNOo+XceUBp47jj4e2w+Wm7L4sW0XLWLN/1vv3CI5zo7qOlZXWxn56h86k98Mh6Lly9MvLwtkqzvvd5eHELAK9adQ5zJ420TSa/u4u5c+fR0rIAgN8e2sjm9v2ln6+/tVPgrFlzbqjcLnM2P4RY0NLyysDbfv43m+juf5GbrlzNgqmVydz/z+N28YnbN3LJhefR6GTtrNX1HEQg+BrZJ/wbgd8CbwG6gbK8/EVkgTFmi/P1ddjCB8CdwIdE5IfYDoRtxpg9InI38DmPI+FFwPXl9CFKcstw1gvFEhMpSjk0JKz8JoMBmQqLRRl4tnNcEvqd7d0SvZYIbznnJBKW8NZzgnmON3iOX7jaYfDUxXvb7bTF08YOPR+CRk8+iUJRBq5TZ1DChuVNHt3I5r3txVf0oXXzAVafMqliwgDAZctnctnyeFi8g+Qh+L7n83oRmYutMXjZGHMw33Yl8gUROQ1IA9uB9zvtdwGXAFuxHRnf5Rz/sIjcBDzmrHejMeZwmX2IjKGWUKRUijkVKko5JC0p0akwSGIie0O3DHHCsusQiEBjMsE7Xjk3RD8HT3z+/QgeZbCvvZuGhMX4MmLma4WvQJDXh6A6ToVgCwR/2hLch6Cts5fn9x/j0tOnhzruUKRkgUBEphtj9rjfjTEdwPooOmGMeUOedoNtkvBb9h3gO1EcP2rG16mGwL33qTygVIJU0sovEHg+J4sEpA90KrTfXQ1BwhJ6+8urRJryTHyFogySzrGCsLetkyljGodkDhNvCupkRiCw05mbHA1BGMLkIQBbIGjv6qOrt39AH4vxl5ePYAysmBObiPaKE2SId4nIcyLyXyLyFif8T/EhytKYcSIbSjT0blZK/InMZOC5/NzJx9UQ2CaDcHHw2X6WZjIY05TiWFfphXVOdPfxh+cOcPqsseE7V0O8GgJvRIAwMIWzMeEm99AaAifqK2ikwaMvHSZpCctmjwt13KFIkL/lVOBfsfMEfJGBAsKVFemdEkvUh0CpBKlEfpNBOoDJwDtxuHJEfzqrIQirsvY7fqFaBmOaU7R1lC4Q3P7ETto6e3nfmvmh+1ZLGn00BJAVwlxMlX0Ipjr+GDucKpKl8sDWg5xx0jhGNgZxtRvalCwQGGO2GmO+ZYx5uzFmNrAE+DPwbuD7hbdW6gFX6lcfAqUSJBOFTAbeTIWFz7/rL1nEm50oBFez4Coe3MmpnFPYm62uUB6CcSNSHOvuKzk50S+f3M3CaaM546ShqaLOpyGwRHIyFYa7h4TVTC6daWtcNuw4WvI2R0708NSuNs49ZXKoYw5VgqQutkRkpYhcJyK/AR7BDv37PrZQoHg479TJnFZBz9RakClupPKAUgFSjsnA+NiYgzgVXrxkWkbN687FfY6gkbAk9BNq9viDU/T64YbItXcW1xLsb+/iiZeP8JpXDF1LbGMe3woRv0yF1evXhJENzJ88kvXbj5S8zRPbbf+BV548sYI9ix9BdCFHgS7sHAE/BN5vjNlekV7VAbe9e2WtuxA56kOgVBLXNt+XNoP8BIIIBJAVWl3NQndfViBIp8vTcjUkSosycAWCo529RR2NH9h6EGPgwsVTQver1rgx9DBwfCUnY2PY4kblcOZJ47n9iZ3c8+w+1i6eWnT9Fw7YxZBOm1ZfD3XFCOJD8BQwBruY0NnAChGpTC5HJZZkUxerQKBEjzvR+5kNBpgMSlBRueeoqyHocQQCY0zZT6gDfAgK9GWcEzrYVoKG4Old7TSlrCGtVXSLVMFgH4LcKINq30He6CSP+sydzxRd9y23Psznf7OZSaMaQyVCGsoE8SFYDYwHrsXWFvwdsMMpS/y1CvVPiRFxqsql1B8ZgaCvPJMBkMmI40YXdPfZKYH70qZsp0KvD0Oh/PoZDUFHT951XJ7Z3cai6WMqUk2vWng1BLmFjrxuFEETEwWtNeHHqvkT+cirF7CnrTMjHPpx6Hg3D714KLLjDjUCnX3GmE5jzL3AfwBfB34KLMKuPaDUOe4lrD4ESiVwzQQHjnfx2V89O+DGnQ4QdgiDJ3x3X32Oj0LYmHYYaDIo7ENgmwmKaQi2HzrBIy8dZskQS1WcS6NHQ+A1K9phh94og2APFT//4Gru/eh5ZffvpAkjSBvYdbQz7zobd7ZlPo8eRtEFLkGcCt8kIv8pIpuwaxd8GdsH4cPYJZGVOkczFSqVxH3yv+HnT/OtB17ivs3ZYjRBqh1CVmh1JyLXh8DWEJTrVBjMh6CQU2Fvf5p1X7WL5iyfPTSjC1y8ToVecn0IgmpoxjanOGXKqHK7x0kT7CqVLx/2Dz+8Y8Mu3vXdxzLf3aq1w4kgItAtwP3YNQ1anTLIyjBCnQqVSuJOtNsOnQBgTHP29uSdUIqFHUL2CTTXh6CvP03alHcOD8xUmF84yZoM8gsEf3n5KJ29/bz33HlcvnxG6D7FAa/JwItlDfQhKFdDE5Y5Ex2B4NAJIBtO2Nef5pbfb+GW+7YCdmbDT792MWsWDK+QQwhWy2DoxsMokSDORawmA6USuBPtvnY7o5wrIOSGIZamIZAB27oagn5HQ1DOKZwqMVNhQ9JiREOioMngga0HsQQ+fMGCIe0/AAOdCr3kZiosV0MTlsmjGmlMWnyj9QUWzxjL3c/s5bp1C7lv8/6MMPDNt5/JwmljmDtp+PkPQDANASIyFXgHcDLwKWPMQRFZDew2xrxUiQ4q8UFNBkolSeVMrm60QW5aglQJj5eSE2XgOhX2pg1tnb2MGxG+3oj3+MUiHiaMbGB/gZS5f9i8n6WzxjF2CBYzyiWvhkBkQJRI2lD1KAOwNRVnz53AA1sP8oZv/BmAS5dO55cb9zBuRIrWj7eUdV7UA0F8CM4CngPeBrwHOwQRYC1wc/RdU+KGJiZSKknuk3+fk14wN+agJJOB8z5YQ5Bmb1sX08aEtw+nfKr65WPhtNE8u2dg6d3DJ3roTxs27WnnqV1tXLZsaJsKXPI5e+aWgbbDPmtzE/neu1dywyWLMt/vf/4A9z67j0uWTh/2wgAEizL4EvA1Y8wZgFfkvRtYHWmvlFiiPgRKJUnlOKX1pe1JPLc6XiCTgfPdFQi6etMcPN7NtDFN4fvpzVRYRDhZMmMsLxw4zonuPsD2ZTjzpnu47qcb+dFjO2hIWLz+jJmh+xIn8t0XcqsdQu0eKixLeO+aefzm2jWcNGEE/3bP83T29vPa0+tDKCuXIALBWcBtPu17gOKpn5S6QeUBpRLkPmFmNAS5JoOSwg7td1eYcJ0K97R1kTbZgjeh+mmVriFYOnMsxpDREmx3HCZvf2Inv9iwi7VLptZtuXQXa1CUQW18CFxEhEXTx7BmgZ1Xb0RDgpXzJtSsP3EiiA9BJ3ZiolwWAvt92pU6Q30IlEoyyGSQdk0GAyWCUjRUmSgDJ5WBqyHY5VS8K0dD4E1GVKwkr1vK+FO/eJqjHb3M8zirHe3o5cqzZ4fux1Aht9phuamjo+L6SxYxcWQDC6aODl1Jsd4IIhDcAXxaRN7kfDciMhf4F+wERUqdk0oIly2fodK0UhFyBYJ8ToWlIBmTgashsJ0K27ts1f3UMgQCL4XCDgGmjGni/eedzDfvfwGAve1dAKyYM55XL5rKuafUf/Z3vyiDmngV5jCqMclHLzqt1t2IFUEEgo9jFzY6AIwAHsA2FTwI/GP0XVPihojwtSvPqHU3lDqlVJNBKWTDDu3v3TnpaqeVYTLwkijBfPGJi0/j4iVT6ezt563//QgAt3/gVZEcfyggIoOLGw3tCMu6pSSBQERSwD3AO4GZwJnY/gfrnVTGiqIoZdGQoyHoz2MyKAV3ms71IQA7VHBCRB7lJRVasoQzThqfKcE83LCsrFPhsa5entndxiVLNa1NHClJIDDG9IrIPPujuQ+4r7LdUhRluDHIZJCJMgi+L/cJ1E9DMKopWbAoURCC2J6TCYsfvPccxgyzCnpC1ofgew9t50RPP29fNafGvVL8CGIyuA14H/D3FeqLoijDmNwQvqzJwH6fOqaRJTPGlrSvbGKiwRqC5pR/Ap0wFHMqzOVVw8BnIBeD4RcbdmOJ8PMNu1i7eCrLZo+rdbcUH4IIBCOBt4nIWuAJ4IR3oTHmI1F2TFGU4UWuySDjVOh8f9+a+bx3zfyS9pU1GdhZCjt7+zPLohQIotI01AujG5Mcc3IuuOw4bFcX/NlfdrFs9jj+/S3qhxRXgggEi4D1zufcqzKEUk9RFCVLrskg40PgPNwHSYiVDWsz7DwysNxtU4QCgTKQR254Nff/8U++y847dTKfu2Kpjn+MCVLc6PxKdkRRlOHN4EyFA50KgzyLW55aBm6525MmjODlwx00N+iEVClGNCRpTvr/U9++asWQL+BU7+i/oyhKLMj12M/NQxBEO59NTGTY4QgE8yfbSYGiNBkopaPCQPwJVO1QURSlUnhNBpZknQpdx8AgJgN3VQO8fKiDppSVyU6oKuvq8qU3LWNinadnrhdUIFAUJRZ4Q/iSCctjMrAJ4tBveaIMXj7cwUkTRmQEjqaUPqlWkzeeNavWXVBKRK8MRVFiR9KSTCIf12QQSEPgvBsD+9q7mDa2ORPWqCYDRfFHBQJFUWLFlNGNtkCQHpiHIJBToaNt6E8bevoNDQkr46OgToWK4o8KBIqixIZffuhc7rp2DamENSgPQRCTwTgnG+DRzl7602mSlmSc2lRDoCj+qECgKEpsWDprLJNGNZJMSDYPQSbKoHSJYPLoRgAOHOumL21IJiSTVTAKp8Jls0rLmKgoQwl1KlQUJXYkLYve3CiDANuPbU7RkLA4cKyb/rQZYIJojMCp8KcfeFWoGguKEmdUIFAUJXYkE0JfOrzJQESYPLqR/ce66Os3JKysCSI3RXK4/qlyVak/9KxWFCV22FEGOU6FAQsJTRrd6JgM0gOiFkopWawowxEVCBRFiR2phJXVELhhhwH3McURCPodH4IeR8DITZGsKIqNXhmKosSOxAANgd0WVEMwOaMhMAM0BClLb3uK4odeGYqixI5kwqI3p7hRUE3/lNGNHO7ooau3n4SVzXyYylN8R1GGO7ESCETk4yJiRGSS811E5BYR2SoiG0XkTM+6V4nIFud1Ve16rShK1KQ8T/TpjIYg2D4mjGzAGOjqTTsmA9eHIFa3PUWJDbGJMhCR2cBa4GVP82uABc7rHOAbwDkiMgH4NLAC2wn5CRG50xhzpLq9VhSlEthRBrmZCoNJBN5iSQNMBhohoCi+xOnK+ArwCbJRRgCXAd8zNg8D40RkOnAxcI8x5rAjBNwDrKt6jxVFqQhJy6KvP80dG3bxxHZbzg+qIcgVCM6ZNxHIlkFWFGUgsdAQiMjrgF3GmCdzHIdmAjs833c6bfna/fZ9NXA1wNSpU2ltbY2u48Dx48cj3+dwQ8ewfOptDNuPdnGsx3DtDzdk2jZt2sTYo1tK3seW3X2Zzzte3s4ZJ6f4t5Zmdm96gt2b/Lept3GsBTqG5VOrMayaQCAi9wLTfBbdAPwDcJHfZj5tpkD74EZjbgVuBVixYoVpaWkppbsl09raStT7HG7oGJZPvY3h/25/nN6jndDenmlbsngxLctmlLyPExv3wMb1AJxy8nzOP/+UotvU2zjWAh3D8qnVGFZNIDDGXOjXLiJLgXmAqx2YBawXkZXYT/6zPavPAnY77S057a2Rd1pRlJqQtIR+Jw+BS5BaBgCpRHZ9TUakKMWpuQ+BMeYpY8wUY8xcY8xc7Mn+TGPMXuBO4J1OtMEqoM0Yswe4G7hIRMaLyHhs7cLdtfoNiqJESzKRzUPgUo4PQUIFAkUpSix8CApwF3AJsBXoAN4FYIw5LCI3AY85691ojDlcmy4qihI1qYRFb46GIOiUnutUqChKYWInEDhaAvezAa7Js953gO9UqVuKolQRby0Dl6CZCpNek4GGGipKUfQqURQldnjzELgENxmoD4GiBEEFAkVRYoebh8BLOSYD9SFQlOKoQKAoSuzwcyoMGmXgTVHsNR8oiuKPCgSKosSOVMLK1B5wKc9koLc6RSmGXiWKosSOhCV095UrEGiUgaIEQQUCRVFiR8pnAi8nykB9CBSlOCoQKIoSO/zCBMtxKtQKh4pSHL1KFEWJHX5OgEE1BBploCjBUIFAUZTY4WfzDzqnJzUPgaIEQgUCRVFih19UgAQ0GjSohkBRAqECgaIosSPlazIItg+vVkBTFytKcfQqURQldvhN4M0NiUD78GoF1GSgKMVRgUBRlNjhp+KfPrYp0D68TohqMlCU4qhAoChK7PAzGUwe1Rh6f5q6WFGKowKBoiixw8+psBw/AE1drCjF0atEUZTY4achKAf1IVCU4qhAoChK7EhE/ESvPgSKUhwVCBRFiR25Nv+GZHm3KvUhUJTiqECgKErsSOVoCC5fPqOs/akPgaIUJ1nrDiiKouTifaL/whVLecNZs8rbn5oMFKUoKjYrihI7vBP43Ekjy65WmFCTgaIURQUCRVFiRzLiOgS5JghFUQajV4miKLHDqyGIQiDQKANFKY4KBIqixA6viSAK+7/6EChKcVQgUBQldiQi1hBYKhAoSlFUIFAUJXZ4MxWqul9RqoMKBIqixI5kxCYDRVGKowKBoiixIzXAZKC3KUWpBnqlKYoSOwb4EIhqCBSlGqhAoChK7BiQh0CTCilKVdDUxYqixA6vU2E5PgSfvfwVPPTCoSi6pCh1jwoEiqLEDm8xonKiDN6+ag5vXzUnii4pSt2jJgNFUWJHUn0IFKXqqECgKErssCzBlQnUh0BRqoMKBIqixBLXsVDzEChKdVCBQFGUWOLmItBMhYpSHVQgUBQllriCgPoQKEr5rUkzAAAKSklEQVR1iIVAICKfEZFdIrLBeV3iWXa9iGwVkedE5GJP+zqnbauIfLI2PVcUpVK4FQ9VQ6Ao1SFOYYdfMcZ8ydsgIouBK4ElwAzgXhE51Vn8dWAtsBN4TETuNMY8W80OK4pSOZIJIWEJohoCRakKcRII/LgM+KExpht4SUS2AiudZVuNMS8CiMgPnXVVIFCUOiFpWaodUJQqEieB4EMi8k7gceBjxpgjwEzgYc86O502gB057ef47VRErgauBpg6dSqtra2Rdvr48eOR73O4oWNYPvU4hr09XWBMVX9XPY5jtdExLJ9ajWHVBAIRuReY5rPoBuAbwE2Acd6/DLwb8Hs8MPj7Phi/4xpjbgVuBVixYoVpaWkJ2vWCtLa2EvU+hxs6huVTj2M4+olWOvu7q/q76nEcq42OYfnUagyrJhAYYy4sZT0R+W/gV87XncBsz+JZwG7nc752RVHqgFTC0qREilJF4hJlMN3z9fXA087nO4ErRaRRROYBC4BHgceABSIyT0QasB0P76xmnxVFqSzJhGjIoaJUkbj4EHxRRJZjq/23AX8DYIx5RkR+jO0s2AdcY4zpBxCRDwF3AwngO8aYZ2rRcUVRKkNCnQoVparEQiAwxryjwLKbgZt92u8C7qpkvxRFqR0pSzRtsaJUkViYDBRFUXJJJkR9CBSliqhAoChKLEklLPUhUJQqogKBoiixJGGJ+hAoShVRgUBRlFiSSlgkLb1FKUq1iIVToaIoSi7vfOUcDp/oqXU3FGXYoAKBoiixZM2CybXugqIMK1QfpyiKoiiKCgSKoiiKoqhAoCiKoigKKhAoiqIoioIKBIqiKIqioAKBoiiKoiioQKAoiqIoCioQKIqiKIoCiDGm1n2oGiJyANge8W4nAQcj3udwQ8ewfHQMo0HHsXx0DMsn6jGcY4wpmulrWAkElUBEHjfGrKh1P4YyOoblo2MYDTqO5aNjWD61GkM1GSiKoiiKogKBoiiKoigqEETBrbXuQB2gY1g+OobRoONYPjqG5VOTMVQfAkVRFEVRVEOgKIqiKIoKBIqiKIqioAJBaERknYg8JyJbReSTte5PnBGR74jIfhF52tM2QUTuEZEtzvt4p11E5BZnXDeKyJm163l8EJHZIvIHEdkkIs+IyLVOu45jiYhIk4g8KiJPOmP4z077PBF5xBnDH4lIg9Pe6Hzf6iyfW8v+xwkRSYjIX0TkV853HcOAiMg2EXlKRDaIyONOW02vZxUIQiAiCeDrwGuAxcBbRGRxbXsVa74LrMtp+yTwe2PMAuD3znewx3SB87oa+EaV+hh3+oCPGWMWAauAa5xzTsexdLqBC4wxy4DlwDoRWQX8C/AVZwyPAO9x1n8PcMQYcwrwFWc9xeZaYJPnu45hOM43xiz35Byo6fWsAkE4VgJbjTEvGmN6gB8Cl9W4T7HFGPNH4HBO82XAbc7n24DLPe3fMzYPA+NEZHp1ehpfjDF7jDHrnc/HsG/GM9FxLBlnLI47X1POywAXALc77blj6I7t7cCrRUSq1N3YIiKzgEuBbznfBR3DqKjp9awCQThmAjs833c6bUrpTDXG7AF7sgOmOO06tkVw1K5nAI+g4xgIR9W9AdgP3AO8ABw1xvQ5q3jHKTOGzvI2YGJ1exxLvgp8Akg73yeiYxgGA/xORJ4Qkaudtppez8modzhM8JNwNX4zGnRsCyAio4CfAn9rjGkv8LCl4+iDMaYfWC4i44CfA4v8VnPedQxzEJH/A+w3xjwhIi1us8+qOobFWW2M2S0iU4B7RGRzgXWrMo6qIQjHTmC25/ssYHeN+jJU2eeqvJz3/U67jm0eRCSFLQx83xjzM6dZxzEExpijQCu2P8Y4EXEfjrzjlBlDZ/lYBpu+hhurgdeJyDZsU+kF2BoDHcOAGGN2O+/7sYXTldT4elaBIByPAQscz9oG4Ergzhr3aahxJ3CV8/kq4A5P+zsdr9pVQJurQhvOOHbXbwObjDH/5lmk41giIjLZ0QwgIs3Ahdi+GH8A3uisljuG7ti+EbjPDPNMbsaY640xs4wxc7Hve/cZY96GjmEgRGSkiIx2PwMXAU9T6+vZGKOvEC/gEuB5bBvkDbXuT5xfwP8D9gC92JLue7DtiL8HtjjvE5x1BTuC4wXgKWBFrfsfhxdwLraKcCOwwXldouMYaAxPB/7ijOHTwD857fOBR4GtwE+ARqe9yfm+1Vk+v9a/IU4voAX4lY5hqLGbDzzpvJ5x55BaX8+aulhRFEVRFDUZKIqiKIqiAoGiKIqiKKhAoCiKoigKKhAoiqIoioIKBIqiKIqioAKBoihlIiJzRcSIyIria4c+xhtFREOiFKWCaOpiRVHKZQcwHThY644oihIeFQgURSkLY9cH2FvrfiiKUh5qMlCUYY6TDvUTIvKCiHSKyFMi8nZnmWsOeKuIPCAiXSKyWUQu8mw/wGQgIikRuUVEdotIt4jsEJEveNYfLyK3icgR53j3isiSnD69U0S2i0iHiPwKmOrT79c6leK6ROQlEbnZSSWuKEoIVCBQFOWz2OmkrwEWA58H/ktELvWs80XgFmA5dtngO0QkX/nVjwCvx851vwB4M/CcZ/l3gXOwa7yvBDqA3zr1BRCRc5x1bnWO90vgRu8BRORi4PvAfwBLgHdj58r/XMDfriiKg6YuVpRhjFNY5SBwkTHmT572rwKnAh8EXgL+0Rhzs7PMAjYDPzbG/KOIzHXWOdsY87iI3II9SV9ocm4wIrIAuwbIecaYPzptY4GXgY8ZY74lIj8AJhtj1nq2+xbwHmOMON//CNxjjLnJs87lwP8Co3OPqyhKcdSHQFGGN4uxC9D8NseLPwVs83x/yP1gjEmLyCPOtn58F1uL8LyI/A64C/iNMSYNLALSOftrE5GnPPtbhK0V8PIQthbD5SxgpYhc52mzgGZgGnYxLUVRAqACgaIMb1yz4Wuxn9K99GJXWQuEMWa9ozVYB1wA3AY8KSJri+zPFUhKOaYF/DN2Jb1cDpTcWUVRMqhAoCjDm2eBbmCOMea+3IXOxA6wCrjPaRNs2//t+XZqjDmGPVn/RES+CzwMnOIczwJeCbgmgzHAUuB/PH1albPL3O/rgYXGmK3Ff6KiKKWgAoGiDGOMMcdE5EvAl5yJ/o/AKOwJOA38zln1AyLyPHYt9g8Cc4Bv+O1TRD6KrbLfgK1leCvQDuw0xnSIyB3YTotXA0eBm53lP3B2cQvwZxG5HlvoaMF2UvRyI/ArEdkO/BjoA14BrDTGfCL8iCjK8EWjDBRF+RTwGeDjwDPY9v83YDsKunwS+CjwJLYp4PXGmJ159ncM+HvgUewn+eXAa4wxHc7ydznL7nTeRwDrjDGdAMaYh7H9BT4AbASucPqXwRhzN3ApcL6zj0edPuaaPRRFKRGNMlAUJS+5EQS17Y2iKJVENQSKoiiKoqhAoCiKoiiKmgwURVEURUE1BIqiKIqioAKBoiiKoiioQKAoiqIoCioQKIqiKIqCCgSKoiiKogD/HzoWQcJ7bl8LAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(tot_ep_reward)\n",
    "plt.xlabel(\"episode\", fontsize=14)\n",
    "plt.ylabel(\"reward\", fontsize = 14)\n",
    "plt.title(\"total reward vs. episodes\", fontsize = 16)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Watch the agent play an episode"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is the fun part!\n",
    "Now we get to see how good our agent really is by watching it play an episode."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n",
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n",
      "INFO:tensorflow:Restoring parameters from models/BipedalWalker/model.ckpt\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -73.31\n",
      "Sum of x velocity in this episode 0.018663935171400952\n"
     ]
    }
   ],
   "source": [
    "ep_reward = []\n",
    "ep_obs = []\n",
    "env = gym.make('BipedalWalker-v2')\n",
    "# env._max_episode_steps = 3000\n",
    "# env = gym.wrappers.Monitor(env, \"recording/LunarLander\")\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, \"models/BipedalWalker/model.ckpt\") # load model\n",
    "    obs = env.reset() # Reset env and save observation\n",
    "    t = 0\n",
    "    while True:\n",
    "        env.render() # Render game\n",
    "        # Use our model to create a probability distribution of actions based on observation\n",
    "        action = np.squeeze(sess.run(sample_action, feed_dict = {input_ : obs.reshape((1,obs_dim[0]))}))\n",
    "        # clip action values to fit into the allowed box\n",
    "        action = np.clip(action, env.action_space.low[0], env.action_space.high[0])\n",
    "        obs, reward, done, info = env.step(action)\n",
    "        ep_reward.append(reward)\n",
    "        ep_obs.append(obs)\n",
    "        t = t+1\n",
    "        if done:\n",
    "            ep_obs = np.vstack(ep_obs)\n",
    "            print(\"Game ended after {} steps\".format(t+1))\n",
    "            print(\"Total reward earned in this episode: {:0.2f}\".format(sum(ep_reward)))\n",
    "            print(\"Sum of x velocity in this episode {}\".format(np.mean(ep_obs[:,2])))\n",
    "            break\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n",
      "\u001b[33mWARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.\u001b[0m\n",
      "INFO:tensorflow:Restoring parameters from models/BipedalWalker/model.ckpt\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -73.07\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -68.69\n",
      "-------------------------------------------\n",
      "Game ended after 1475 steps\n",
      "Total reward earned in this episode: -156.08\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -71.52\n",
      "-------------------------------------------\n",
      "Game ended after 1531 steps\n",
      "Total reward earned in this episode: -169.94\n",
      "-------------------------------------------\n",
      "Game ended after 1162 steps\n",
      "Total reward earned in this episode: -149.89\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -59.02\n",
      "-------------------------------------------\n",
      "Game ended after 1582 steps\n",
      "Total reward earned in this episode: -170.76\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -74.33\n",
      "-------------------------------------------\n",
      "Game ended after 1048 steps\n",
      "Total reward earned in this episode: -139.96\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -82.42\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -78.26\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -78.92\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -53.24\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -67.17\n",
      "-------------------------------------------\n",
      "Game ended after 274 steps\n",
      "Total reward earned in this episode: -100.87\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -64.98\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -80.73\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -57.52\n",
      "-------------------------------------------\n",
      "Game ended after 887 steps\n",
      "Total reward earned in this episode: -137.63\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -61.19\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -55.40\n",
      "-------------------------------------------\n",
      "Game ended after 1112 steps\n",
      "Total reward earned in this episode: -136.36\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -69.85\n",
      "-------------------------------------------\n",
      "Game ended after 521 steps\n",
      "Total reward earned in this episode: -114.90\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -62.63\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -68.90\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -60.57\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -76.81\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -72.66\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -61.80\n",
      "-------------------------------------------\n",
      "Game ended after 622 steps\n",
      "Total reward earned in this episode: -124.98\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -52.54\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -60.86\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -55.05\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -63.35\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -76.27\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -81.04\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -64.67\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -66.42\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -87.52\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -67.77\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -87.39\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -73.52\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -68.62\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -72.65\n",
      "-------------------------------------------\n",
      "Game ended after 1080 steps\n",
      "Total reward earned in this episode: -132.13\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -66.51\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -49.02\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -65.76\n",
      "-------------------------------------------\n",
      "Game ended after 1312 steps\n",
      "Total reward earned in this episode: -149.12\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -60.69\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -74.59\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -65.59\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -88.17\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -60.73\n",
      "-------------------------------------------\n",
      "Game ended after 126 steps\n",
      "Total reward earned in this episode: -126.46\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -69.40\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -85.90\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -68.27\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -74.43\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -72.44\n",
      "-------------------------------------------\n",
      "Game ended after 1182 steps\n",
      "Total reward earned in this episode: -139.09\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -57.37\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -63.38\n",
      "-------------------------------------------\n",
      "Game ended after 576 steps\n",
      "Total reward earned in this episode: -115.71\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -79.73\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -70.88\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -88.92\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -72.19\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -74.11\n",
      "-------------------------------------------\n",
      "Game ended after 1076 steps\n",
      "Total reward earned in this episode: -150.87\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -65.44\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -78.80\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -78.90\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -76.33\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -77.58\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -86.15\n",
      "-------------------------------------------\n",
      "Game ended after 811 steps\n",
      "Total reward earned in this episode: -137.42\n",
      "-------------------------------------------\n",
      "Game ended after 1138 steps\n",
      "Total reward earned in this episode: -140.94\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -76.11\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -88.33\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -35.84\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -57.59\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -76.08\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -82.46\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -77.38\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -82.64\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -67.14\n",
      "-------------------------------------------\n",
      "Game ended after 721 steps\n",
      "Total reward earned in this episode: -116.37\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -78.98\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -86.21\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -72.17\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -73.94\n",
      "-------------------------------------------\n",
      "Game ended after 1544 steps\n",
      "Total reward earned in this episode: -177.04\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -73.06\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -75.46\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -68.45\n",
      "-------------------------------------------\n",
      "Game ended after 1601 steps\n",
      "Total reward earned in this episode: -42.87\n",
      "-------------------------------------------\n",
      "Game ended after 1177 steps\n",
      "Total reward earned in this episode: -126.75\n",
      "===================================================\n",
      "Average reward in batch is -84.77\n"
     ]
    }
   ],
   "source": [
    "num_ep = 100\n",
    "ep_reward = []\n",
    "batch_reward = []\n",
    "env = gym.make('BipedalWalker-v2')\n",
    "# env._max_episode_steps = 1000\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, \"models/BipedalWalker/model.ckpt\") # load model\n",
    "    for ep in range(num_ep):\n",
    "        obs = env.reset() # Reset env and save observation\n",
    "        t = 0\n",
    "        while True:\n",
    "            # Use our model to create a probability distribution of actions based on observation\n",
    "            action = np.squeeze(sess.run(sample_action, feed_dict = {input_ : obs.reshape((1,obs_dim[0]))}))\n",
    "            # clip action values to fit into the allowed box\n",
    "            action = np.clip(action, env.action_space.low[0], env.action_space.high[0])\n",
    "            obs, reward, done, info = env.step(action)\n",
    "            ep_reward.append(reward)\n",
    "            t = t+1\n",
    "            if done:\n",
    "                print(\"-------------------------------------------\")\n",
    "                print(\"Game ended after {} steps\".format(t+1))\n",
    "                print(\"Total reward earned in this episode: {:0.2f}\".format(sum(ep_reward)))\n",
    "                batch_reward.append(sum(ep_reward))\n",
    "                ep_reward = []\n",
    "                break\n",
    "print(\"===================================================\")\n",
    "print(\"Average reward in batch is {:0.2f}\".format(np.mean(batch_reward)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}