q_learning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The Agent-Environment Interaction\n",
    "\n",
    "In this exercise, you will implement the interaction of a reinforecment learning agent with its environment. We will use the gridworld environment from the second lecture. You will find a description of the environment below, along with two pieces of relevant material from the lectures: the agent-environment interface and the Q-learning algorithm.\n",
    "\n",
    "1. Create an agent that chooses actions randomly with this environment. \n",
    "\n",
    "2. Create an agent that uses Q-learning. You can use initial Q values of 0, a stochasticity parameter for the $\\epsilon$-greedy policy function $\\epsilon=0.05$, and a learning rate $\\alpha = 0.1$. But feel free to experiment with other settings of these three parameters.\n",
    "\n",
    "3. Plot the mean total reward (i.e. the undiscounted return) obtained by the two agents for each episode. This kind of graph is called a **learning curve**, and it gives us an idea of how our agent's performance changes during training.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## The agent-environment interface\n",
    "\n",
    "<img src=\"img/agent-environment.png\" style=\"width: 500px;\" align=\"left\"/> \n",
    "\n",
    "<br><br><br>\n",
    "\n",
    "The interaction of the agent with its environments starts at decision stage $t=0$ with the observation of the current state $s_0$. (Notice that there is no reward at this initial stage.) The agent then chooses an action to execute at decision stage $t=1$. The environment responds by changing its state to $s_1$ and returning the numerical reward signal $r_1$. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The environment: Navigation in a gridworld\n",
    "\n",
    "<img src=\"img/gold.png\" style=\"width: 250px;\" align=\"left\"/>\n",
    "\n",
    "The agent has four possible actions in each state (grid square): west, north, south, and east. The actions are unreliable. They move the agent in the intended direction with probability 0.8, and with probability 0.2, they move the agent in a random other direction. If the direction of movement is blocked, the agent remains in the same grid square. The initial state of the agent is one of the five grid squares at the bottom, selected randomly. The grid squares with the gold and the bomb are **terminal states**. If the agent finds itself in one of these squares, the episode ends. Then a new episode begins with the agent at a randomly selected initial state.\n",
    "\n",
    "You will use a reinforcement learning algorithm to compute the best policy for finding the gold with as few steps as possible while avoiding the bomb. For this, we will use the following reward function: $-1$ for each navigation action, an additional $+10$ for finding the gold, and an additional $-10$ for hitting the bomb. For example, the immediate reward for transitioning into the square with the gold is $-1 + 10 = +9$. Do not use discounting (that is, set $\\gamma=1$)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Q-Learning\n",
    "For your reference, the pseudocode for the Q-Learning algorithm is reproduced below (Reinforcement Learning, Sutton & Barto, 2018, Section 6.5 p.131).\n",
    "<img src=\"img/q.png\" style=\"width: 720px;\"/>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Example of a learning curve\n",
    "\n",
    "<img src=\"img/lc_example.png\" style=\"width: 550px;\" align=\"left\"/>\n",
    "\n",
    "<br><br><br><br>\n",
    "\n",
    "This is a sample learning curve and shows the reward obtained by a Q-learning agent across 500 episodes. Do not try to replicate this exact curve! It was computed using a different environment than the one described here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# environment\n",
    "import numpy as np\n",
    "\n",
    "class Gridworld:\n",
    "    def __init__(self):\n",
    "        self.num_rows = 5\n",
    "        self.num_cols = 5\n",
    "        self.num_cells = self.num_cols * self.num_rows\n",
    "        self.random_move_probability = 0.2\n",
    "        \n",
    "        # Choose starting position of the agent randomly among the first 5 cells\n",
    "        self.agent_position = np.random.randint(0, 5)\n",
    "        \n",
    "        # Choose position of the gold and bomb\n",
    "        self.bomb_positions = np.array([18])\n",
    "        self.gold_positions = np.array([23])\n",
    "        self.terminal_states = np.array([self.bomb_positions, self.gold_positions])\n",
    "       \n",
    "        # Specify rewards\n",
    "        self.rewards = np.zeros(self.num_cells)\n",
    "        self.rewards[self.bomb_positions] = -10\n",
    "        self.rewards[self.gold_positions] = 10\n",
    "        \n",
    "        # Specify available actions\n",
    "        self.actions = [\"UP\", \"RIGHT\", \"DOWN\", \"LEFT\"]\n",
    "        self.num_actions = len(self.actions)\n",
    "        \n",
    "    def get_available_actions(self):\n",
    "        return self.actions\n",
    "    \n",
    "    def get_start_location(self):\n",
    "        return self.agent_position\n",
    "    \n",
    "    def make_step(self, action_index): \n",
    "        # Randomly sample action_index if world is stochastic\n",
    "        if np.random.uniform(0, 1) < self.random_move_probability:\n",
    "            action_indices = np.arange(self.num_actions, dtype=int)\n",
    "            action_indices = np.delete(action_indices, action_index)\n",
    "            action_index = np.random.choice(action_indices, 1)[0]\n",
    "\n",
    "        action = self.actions[action_index]\n",
    "\n",
    "        # Determine new position and check whether the agent hits a wall.\n",
    "        old_position = self.agent_position\n",
    "        new_position = self.agent_position\n",
    "        if action == \"DOWN\":\n",
    "            candidate_position = old_position + self.num_cols\n",
    "            if candidate_position < self.num_cells:\n",
    "                new_position = candidate_position\n",
    "        elif action == \"RIGHT\":\n",
    "            candidate_position = old_position + 1\n",
    "            if candidate_position < self.num_cells:\n",
    "                new_position = candidate_position\n",
    "        elif action == \"UP\":\n",
    "            candidate_position = old_position - self.num_cols\n",
    "            if candidate_position >= 0:\n",
    "                new_position = candidate_position\n",
    "        elif action == \"LEFT\":  # \"LEFT\"\n",
    "            candidate_position = old_position - 1\n",
    "            if candidate_position >= 0:\n",
    "                new_position = candidate_position\n",
    "        else:\n",
    "            raise ValueError('Action was mis-specified!')\n",
    "\n",
    "        # Update the environment state\n",
    "        self.agent_position = new_position\n",
    "        \n",
    "        # Calculate reward\n",
    "        reward = self.rewards[self.agent_position]\n",
    "        reward -= 1\n",
    "        return reward, new_position\n",
    "    \n",
    "    def reset(self):\n",
    "        self.agent_position = np.random.randint(0, 5)\n",
    "        return self.agent_position\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class RandomAgent():\n",
    "    def choose_action(self, available_actions):\n",
    "        number_of_actions = len(available_actions)\n",
    "        random_action_index = np.random.randint(0, number_of_actions)\n",
    "        return random_action_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class QLearningAgent():\n",
    "    def __init__(self, start_location):\n",
    "        # table in format: Position, Action, Q-Value\n",
    "        #self.q_table = np.zeros((1,2))\n",
    "        self.q_table = {}\n",
    "        self.position = start_location\n",
    "    \n",
    "    def choose_action(self, available_actions):\n",
    "        # check if current action/state in q-table\n",
    "        for i in range(4):\n",
    "            if  not (self.position, i) in self.q_table:\n",
    "                self.q_table[self.position, i] = 0\n",
    "        \n",
    "        # decide if random or greedy\n",
    "        # random action 5% time, greedy 95%\n",
    "        if np.random.uniform(0, 1) > 0.95:\n",
    "            # random action\n",
    "            number_of_actions = len(available_actions)\n",
    "            return np.random.randint(0, number_of_actions)\n",
    "        else:\n",
    "            # greedy action\n",
    "            best_action = 0\n",
    "            for i in range(4):\n",
    "                if self.q_table[self.position, i] > self.q_table[self.position, best_action]:\n",
    "                    best_action = i\n",
    "            return best_action\n",
    "        \n",
    "    def update_table(self, reward, new_position, action):\n",
    "        # find q value and update table with it\n",
    "        learning_rate = 0.1\n",
    "        # find new estimate\n",
    "        newEstimate = 0\n",
    "        for i in range(4):\n",
    "            if  not (new_position, i) in self.q_table:\n",
    "                self.q_table[new_position, i] = 0\n",
    "            else:\n",
    "                if self.q_table[new_position, i] > newEstimate:\n",
    "                    newEstimate = self.q_table[new_position, i]\n",
    "        q_value = round(((1-learning_rate) * self.q_table[self.position, action]) + learning_rate*(reward + newEstimate), 2)\n",
    "        self.q_table[self.position, action] = q_value\n",
    "        # update position to new position\n",
    "        self.position = new_position\n",
    "\n",
    "    def print_table(self):\n",
    "        print(self.q_table)\n",
    "        \n",
    "    def update_position(self, new_position):\n",
    "        self.position = new_position"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "env = Gridworld()\n",
    "randomAgent = RandomAgent()\n",
    "Qagent = QLearningAgent(env.get_start_location())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Available_actions = ['UP', 'RIGHT', 'DOWN', 'LEFT']\n",
      "Chosen action = 0\n"
     ]
    }
   ],
   "source": [
    "available_actions = env.get_available_actions()\n",
    "print(\"Available_actions =\", available_actions)\n",
    "chosen_action = Qagent.choose_action(available_actions)\n",
    "print(\"Chosen action =\", chosen_action)\n",
    "reward, position = env.make_step(chosen_action)\n",
    "Qagent.update_table(reward, position, chosen_action)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "episodeNumber = 500\n",
    "yValues = []\n",
    "xValues = range(episodeNumber)\n",
    "for i in range(episodeNumber):\n",
    "    #print(\"Episode: \", i)\n",
    "    totalReward = 0\n",
    "    reward = 0\n",
    "    while reward != 9 and reward != -11:\n",
    "        #print(\"Current position of the agent =\", env.agent_position)\n",
    "        available_actions = env.get_available_actions()\n",
    "        chosen_action = Qagent.choose_action(available_actions)\n",
    "        #print(\"Chosen action =\", available_actions[chosen_action])\n",
    "        reward, position = env.make_step(chosen_action)\n",
    "        #print(\"Reward obtained =\", reward)\n",
    "        Qagent.update_table(reward, position, chosen_action)\n",
    "        #print(\"Current position of the agent =\", env.agent_position)\n",
    "        totalReward+= reward\n",
    "    starting_location = env.reset()\n",
    "    Qagent.update_position(starting_location)\n",
    "    yValues.append(totalReward)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(xValues, yValues)\n",
    "plt.xlabel('Episodes')\n",
    "plt.ylabel('Reward')\n",
    "plt.title(\"Learning Curve of a Q-Learning Policy Agent\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "episodeNumber = 500\n",
    "yValues = []\n",
    "xValues = range(episodeNumber)\n",
    "for i in range(episodeNumber):\n",
    "    #print(\"Episode: \", i)\n",
    "    totalReward = 0\n",
    "    reward = 0\n",
    "    while reward != 9 and reward != -11:\n",
    "        #print(\"Current position of the agent =\", env.agent_position)\n",
    "        available_actions = env.get_available_actions()\n",
    "        chosen_action = randomAgent.choose_action(available_actions)\n",
    "        #print(\"Chosen action =\", available_actions[chosen_action])\n",
    "        reward, position = env.make_step(chosen_action)\n",
    "        #print(\"Reward obtained =\", reward)\n",
    "        #print(\"Current position of the agent =\", env.agent_position)\n",
    "        totalReward+= reward\n",
    "    starting_location = env.reset()\n",
    "    yValues.append(totalReward)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEWCAYAAACaBstRAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAABnSElEQVR4nO2debgcRfX3v6dn7pKbfQeykhCWECBAhCD7ogRRcQNxAfcIgvqqPxQUEREUEcQdQUFkEUSQHdl3MQkJhBCSQEIIZCEr2XPXmXr/6Kqe6uqq7p65M3fm3pzP89znTu/V3dV16ix1ioQQYBiGYZg0eNUuAMMwDNN9YKHBMAzDpIaFBsMwDJMaFhoMwzBMalhoMAzDMKlhocEwDMOkhoUGE4KIjiSi16tdjlqDiD5ORMuJaBsRHVjt8qSFiI4hohXVLkccRCSIaA/5+89E9ONql4lxw0KjhiCiZUR0QjXLIIR4TgixV6XOT0QnEtGzRLSViNYR0TNE9NFKXa+MXAngXCFEHyHEy505ERHdSERtUgC9R0SPEdHeZSpn1ZCN/3Z5XyuJ6NdElCnmHEKIs4QQP6tQ+W4kog4i2rUS53dcs+rfdLlhobGTUexHXOZrfwrAvwDcBGAkgOEALgLwkRLORUTUlfV3DIDXyni+K4QQfQCMALASwPVlPHc1OUDe1/EAPgvga1UuDwCAiHoD+CSAzQA+X+XidGtYaHQDiMgjovOJ6E0i2kBEdxDRIG37v4hoNRFtlr34fbVtNxLRNUT0EBFtB3Cs7P38HxHNk8f8k4ga5f4hc0bcvnL794noXSJaRURf1U0Nxj0QgF8D+JkQ4q9CiM1CiLwQ4hkhxNfkPhcT0S3aMWPl+bJy+WkiuoyI/gtgB4DziGi2cZ3vENF98ncDEV1JRO8Q0Rpp+ugV84wvJKK3iWgtEd1ERP3lObYByAB4hYjedBz/W2m+2kJEc4joSOcL1RBCNAO4A8Bk7VwnE9HL8lzLiehiyzP5gryv9UT0I217L/nONxLRAgDvM8q5j3yOm4joNV3Lk8f9iYj+I7WF/xLRLkT0G3m+RZTSNCeEWATgOQCT5Lm/RkRLyNes7iOi3WzHyTJcqi2fQkRz5bN4k4imEdGpRDTHOO67RHRvTJE+CWATgEsAfME4thcR/V3e40JZp/VvYDciuot8zfgtIvqWtu1i8r/Hm8jXnl8joily280ARgO4Xz7P76d5djWPEIL/auQPwDIAJ1jWfxvADPi98wYA1wK4Tdv+ZQB95bbfAJirbbsRfu/qcPidhEZ5nVkAdgMwCMBCAGfJ/Y8BsMIok2vfaQBWA9gXQBOAWwAIAHtY7mFvuW33mPu/GMAt2vJYeUxWLj8N4B15vSyA/gC2ApigHfMigNPl76sB3CfL3RfA/QB+4bj2lwEsATAOQB8A/wZws7bdel/a9s8DGCzL9T35XBod+94I4FL5uzeAmwG8om0/BsB+8n3tD2ANgI8Zz+QvAHoBOABAK4B95PbL4TfWgwCMAjBfvU8AdfIefwigHsBx8vntpZVrPYCDZT15EsBbAM6ELzQvBfBUzDMInhGAifIZfEVeZz2Ag+DX0d8DeNZxnP5sDoFfdz8gn8UIWY8aALyn7lnu+zKAT8aU7QkAV8DXbjsAHKxtuxzAMwAGwv/G5mnPzAMwB75GXC/rx1IAJ2p1tgXAh+Qz+gWAGUnfdHf+q3oB+E97GW6hsRDA8dryrgDaIRtTY98B8iPsL5dvBHCT5Tqf15avAPBn+fsYRIWGa98boDXCAPaAW2gcLrdZG1K5z8VIFhqXGMfcAuAi+XsC/EawCQAB2A5gvLbvYQDeclz7CQDf0Jb30p+x675i7mUjfFONbduNsqHZBCAPv2HeP+ZcvwFwtfFMRmrbZ6EgKJcCmKZtm45CA3gk/Ibc07bfBuBirVx/0bZ9E8BCbXk/AJtiyikAbJH3/iZ8IePBN71doe3XRz7bseazRVhoXKvu23KtawBcJn/vK6/Z4Nh3tHzOk+XyIwB+q20PhIBc/qr2zA4F8I5xvgsA/E2rs49r2yYCaE76prvzH5unugdjANwtTQqb4AuRHIDhRJQhosul6r4FfiUFgCHa8cst51yt/d4B/0N24dp3N+PctusoNsj/nXVCmtf4B4DPyN+fBXCPEGIHgKHwhccc7bk9LNfb2A3A29ry2/C1huFpCiVNeAulCW8TfC1oSMwhVwohBsAXAs3whZQ616FE9JQ0h2wGcJblXGnfiX5PuwFYLoTIG9tHaMtrtN/NluW4egIABwkhBgohxgshLpTXCj1bIcQ2+PVhhOskklHwhY+NvwP4rDR7ngHgDiFEq2PfM+ALv7ly+VZ5bJ1cjqvHYwDspuqQfLc/RLhemO+iUZlUeyIsNLoHywGcJIQYoP01CiFWwm8oTwFwAvyGaqw8hrTjK5XK+F346rxiVMy+r8O/j0/G7LMdfkOv2MWyj3kvjwEYSkST4QuPf8j16+E3cvtqz6y/8J20NlbBbyAUo+GbMdbYdy8g/RffB3AagIFSGGxG+B1YEUK8A9/8+FvN3/IP+Ga1UUKI/gD+nOZckncRfg+jtd+rAIyicADBaPiO+EoSerbkO6UHp7jucgDjbRuEEDMAtMHXnj4L38Tn4kwA48j3+62G71sbAt+kBMTX4+XwtVP92+srhPgQ0tHj0oiz0Kg96oioUfvLwm80LiOiMQBAREOJ6BS5f1/4Nu0N8Bvcn3dhWe8A8CXpXG0C4IyvF76u/l0APyaiLxFRP/Kdz0cQ0XVyt7kAjiKi0UTUH74ZIBYhRDv8iKxfwbfjPybX5+Hb/a8momEAQEQjiOhEx6luA/AdItqdiPrAf47/FEJ0JJUB/jvoALAOQJaILgLQL8Vx6h4eg9+wTtfO954QooWIDoHfKKblDgAXENFAIhoJ38SkmAm/J/x9IqojomPgR67dXsT5S+E2+PVkMhE1wH+2M4UQyxKOu14ed7ysKyMoHJp8E4A/AGgXQjxvOwERHQZf8BwCP9hgMnzn/D/gCxMg/MxGADhXO8UsAFuJ6AfSYZ4hoklEFAowiGENfD9Ij4GFRu3xEPwesvq7GMBv4fc8HyWirfCd4ofK/W+Cr/qvBLBAbusShBD/AfA7AE/Bd7Cqa1vNBEKIOwF8Gr7TeRX8D+pSAPfK7Y8B+Cd8R+QcAA+kLMo/4Gta/zIa+R+ocknT3ePQzEAGN8DvrT4L38fQgnCDG8cj8E1fb8B/Fy2IN9XZ+BX8xrwBwDcAXCLf9UXwG7W0/FSW4S0Aj0LrgQsh2uALiZPga2J/AnCm8COdKoYQ4nH4HYq74PfqxwM4PcVxswB8CX5Aw2b4zmpdG7wZvgC4JXp0wBcA3CuEeFUIsVr9wf+mPkx+FOIlAFbAf2aPA7gTsg4LIXIAPgxf2LwF/7n9Fb5Wn4ZfALhQmrb+L+UxNQ1JZw3DdBoi2gd+tE5Dyh46w5SMNOethe9HWVzG854NP7Dg6HKdsyfBmgbTKchPr9FARAMB/BLA/SwwmC7ibAAvdlZgENGuRHS4NIHtBT9k+u6ylLAH0mM9/EyX8XX4YZI5+OaDb1S1NMxOAREtgx8c8LEynK4efnjv7vDDoG+Hb7pjLLB5imEYhkkNm6cYhmGY1PR489SQIUPE2LFjq10MhmGYbsOcOXPWCyGsA2F7vNAYO3YsZs+enbwjwzAMAwAgordd29g8xTAMw6SGhQbDMAyTGhYaDMMwTGpYaDAMwzCpYaHBMAzDpIaFBsMwDJMaFhoMwzBMalhoMAzDxDB/5WbMXb6p2sWoGbqd0CCiaUT0OhEtIaLzq10ehmG6N08sXIOb/7fMuf3Dv38eH/vjf7uuQDVOtxIaRJQB8Ef4k8hMBPAZIppY3VIxDNNd+OtzS/H+XzwRWveVv8/Gj+99rWzX2NLSjp6cCLZbCQ34UzYuEUIslbOQ3Q5/fmymAqzYuANL1m6rdjEYpmxc+uBCrNrcUrHzv71hO/a/+FHcMsOZhaPb092ExgiEp9FcIdeFIKLpRDSbiGavW7euywrX0zjil0/hhF8/U+1iMEyXcMPzb3X6HG9v2AEAePi11Z0+V63S3YRGKoQQ1wkhpgghpgwdak3UyDA9hvtfWYUL/v1q2c9798sr8IM75zm3d+TyOOP6mZj11ntlv3Y1uOSBBVjdSS0kmyEAQEeOzVO1wkoAo7TlkXJdzbF03TZ05PLVLkZRrNi4A81tuWoXo2bJ5QWWroua68ae/yC+ddvLVSgRIITAN297GbfNeqfs5/7OP1/BP2cvd25fs7UVzy1ej2/fXp17r0Wynt+kduQLQmPO2xsx9vwH8eqKzdUqVlnpbkLjRQATiGh3IqoHcDqA+6pcpggrNzXjuKuewS8fXlTtohTFEb98CmdcP7Ms53pvexv+8+q7ZTlXrfDrx17HcVc9g2Xrt0e23ffKqk6dWwiBf81ejraO4joab64rlCWXr0zvNu84b53sVRdb5lrgX7OXW+8rZziwn3p9bVHn9fxHEuowPrFwDQDg2cU9w1TerYSGEKIDwLkAHgGwEMAdQojyhT2UiQ3bWgEA/12yocolKZ7Zb28sy3nOunkOzr71JayXz6InMGOpb4ZZV4F7enTBGpx35zxc/fgbRR23o60j+N3cXhktcWtLh32DbF+7o9A47855uO3FqHbWrt2LEAJf+tuLRZ23XZqldE2DpCBxCd/uRrcSGgAghHhICLGnEGK8EOKyapfHhuqsCADtuTz2v/gR3Du3+la05rYc2rvIZPbOe75DUDUoLe05tHZ0b9NXXr5Y1ZssFw/PX42v3zwHgB99UwytWiOnBMjJv3sOf3p6SdnKt7m53bpe9cpbtTr15RtfxEX3zi/btSvJEwujWoT+fWwvwVTbkfeP130aBL/CxImM8/71Cqbf1D0mi+t2QqM7sENWNiEEtjS3Y0tLB356/4IqlwrY56KH8dm/zOjSa6qGdu8fP4z3/+LJLr12uSlYLkqTGjOWbrD2ynUzprNX70A/X3NbDkIIvLZqC654+HXnMa+u2IxNO9oSz92rLgMgRmjInrNqaF96ZyOeXLQWN/2ve4SbvvROVKtu04TGxu3Jz8ikI9A0CudRnYx8zNiNf81ZgUcXrCn6etWAhUYF2N5a+PAzXul231tmvI13ZAhfGtpzefz+icUhk4XJi8s24tEyhAPePusdq1NYoVTydq3HtaGEj7CWUHdCJciMRau34PTrZuDSBxdACIHrn38La7b4kTr6QLDOCI3trTls2mFv4HU+8ofncfp1yZ2Hvo3+bNBbWuKFhhB+OT7xpxfSFLlm2NEa1ST0+vpeCfVVCZ2wecqvMD3EOsVCoxJsl422EIUPq61Is1BLew4X3jMfn77uf6mPuX3WO7jqsTfw52eWxu43XZpCdIoZwSqEwPn/fhWn/MGdWkG1q5Wyd9826x389P6udWeJwDxVvNTY0uzXiQWrtmDZhh342QMLcNYt/nvQn/y21uKEhm6eam7vwNqtvr8l47ChKQftotVbE8/dRwqNJE0DALY6BEsto2sDCt089V4KbSxyTqVp5KI+DVRwlPjy93bgjOtndokPkYVGkUz9+RP46t/jbY/btR6MsvsW23iqD7KY3s462WCUYjwx6/O7m5ud+6qybY1p4FTvqlJC44J/v4q//XdZRc7twmZeSOvcrM/6n1p7Lh88P1tjnNT4vr1hO9pzeQghMPb8B3HFIwXT1o62XKC9DGyqD9a/qWmExXRe+jbECw39eRQr7GoB26vT66tpnkrTsWoPNA3dPFV5TeObt72M5xavx5wyBbLEwUKjSFZvacHjC+Ntj8o8JCBg6cykQgmbYurZFmnaUGaFYjAbxMNi/A8dRdT+YjWsWkY9Ir2H3Z7yBWeVmTIntJ5n+LxAwTy1alNzJNyzuS2Ho3/1NL57xyvBO1iqhdzuaMsFmsag3nUAgEdfW43jr3oGD87zw59b24sQGo3+OdyaRrTcQPkDBboSvb6aHbY0Ic2B0Ag5wn1EEV/zhm2tRYWsqyy8DdnKN+ksNEokrtehel15EY37TktOVboiDlfX7Sc/9mIoppTqw8qmaB0qHY65o60DF/z71VSO3c5c44d3vxr4C/TGI+3YiLpMQdOIv1YOd7+8Au+//MlIuKcSUPe/ssp6nua2HNZu9TWNAVLTUGaohe9uAVAwZ6Vp2FUDtMOhRYTNU4V9+vUqvv5VC/M71kNuTaFhdpaEEPjlw4uwYNWWyD76vp582GmaAlWes295CWff+lLRfpWuGInOQqNE1m9zv0wVPdXclis5NltVOr130tKew9sbtjsFljJtNNQV/1p1TSNOIK7c1Iz9L34UQOFjsKF602k0jW2tHSU/pzvnrMBts94peSClEMJqEvrUNS8E57x91nL8Y+Y7WLnJN9nppgeb1tWey6PFGDNRCAzIw1A0Ij3Q7/zzleC3/lyE9ijbO6LX3dGWC3wn9VJImb1cFfasRi7Hoa7gejUu89T4oX2s+29r7YitW399bimO/tVTieUqJ+1GI6svL3h3i7EtXJdXbGzGNU+/iW/e9lJkH1s2CNtzNM16bYHPyb+2MjkrdrR1xHZUuiKknoVGiSyzxNPPWLoB7bl8UBG2tca/4DhUw6QfPv3mOTj6V0/jyUX2Uaqqt6d/zPNWbAoGG8ahf8utMdrBw/MLkVeZGIdwIDQs51r+3o5gVPWmHW2Y9JNH8NsnFieW0X4d/0Jrt5TmALz9xeXY7+JHI+MjZr+9Edc8/SaA6Ietv9P/vRkdwPnpa/+HvX/8cGider7tHfmgzKoBjeuBLlm3Lcg0rL/X1lw08mdHW0fwvFXjo96DOlS9W5ejXCdn6bjYtgOFDstu/Rut9WLjdv89/+4J9/iRSx9cGCT8A/yGU2lOlcJsZFdtasai1VvQ1pGP5NQye/FLZR0e2rdBO59F0zDet2LJ2m2Y9JNHcMeLhVQtLe3h7968/4kXPYLv3TE3tC4fMpeyplGzmA3JK8s34fTrZuDKR14P1PntrR3WCI002NRM1bCt3mL/kJTQUMdu2NaKj/7hv/imJS+SWYH1BsllwwYQ6kHHmafUgCab0DjyiqdwzJVP+2WU6vf9JabhUCVYX4Qaf+vMQijzYzI2fvEad/iwOdJaNZZvrtsWDMrTeemdTZF16vm25UTk2ccJjQ9e/SxO+PUzyOXDTbcttHZHWw5tUph0BEIj/I6UTyONaVGV2dUO6Q1jYBrtVWcNGNgk69S/X16ReF3FtN88hw//7vnU+9t4YuGa2ISKZv287KGFmPab57Bk7bbAYqDoyAu88Ob6oOP0lgwwGKdpVu2WkFv1qM2nsmStbzrUfaTq+1Ltht4ZUlriPXPD30qLNmi2vQtG57PQKJFWoyFR4XkLV2/Ftlb14gVuMCJ8Fq3egjOunxkxXyhmLt2AL9/4YtBT1BsYNdjKFl8OFHp76qO9V1YuW64kvVJf/dgb+PDvCx9nnB1VL3cmk8Y8FT+qNrg9uf/c5Ztw5g2zUqvZSlCuN9T4a55+E1c/Fk3J0dqRw4/uno9P/fkFPDBvVaC1uZz781ZsCjQOhRIaxfhR1Dtpz+WDRriYPuGzi9eFNQ3NoU3km6O2tRY0jYj9XR0nG5i4dxccY3HU6+jlUe+hX6NdaCj/iKvem6ixTmu3Fq9BqkGOgD/B0mnXusPWWxxZClToap32nNpyeXz2LzNx1i1zcOkDC3Dlo379GtqnoGkoYa1rYa40IrbxGyphqOr46fevOnOmIqcLt1I7qcXAQsPBC0vWh0IVTUwTjuq55fJ5tOXyOGDUAIwY0AtztV7nttYOfOPWl/Dc4vWYvcweGveNW1/Ck4vWBrZMvZo1KqHhSG8QaBqyFqpUHo31mci+eg/rT08vCUXhxA0wCwkNrfZu2tEWMoOlH6chQvt/9465ePaNdViwakts2G+j9NuoD2nD9las39YahEn+8uFFVpOXas/Wb2vFuf8oaGAd+TxWb24JDcwEgEsfWBg5x1f+PhuPLViTyrG5o60Dqze3FMxTuXzQqL69YQfGnv9g4CuxoR7xNU+9iSmXPh6sX7i6YG+vz3jYd0Q/PP362uB5m7Z60zyVxqdRGLyXbJ7a0tyO+oyHhjoPAr4Zc+z5DwbvUJloWjRht3jNVow9/0G88OZ6o6wCKzY2y3IWF4p179yV2Oeih/GrR9wj4nW2OzpgSmjoAuFFTWP56/NvBdqV/nTatOe+aUcb1m9rdXYSPDOMDr4Q29HWEXzDazSrgvJXNcl24L9L1mPs+Q/iNc0R/4O7XsW1z4Q7OeWGhYaDL//9RfzzRXdaaLPHpBrQXF4gl8+jziM01HmhnsyknzwSNM4bHb1UFWWjTCL696rqmGvEtzIBbGnuwP2vrAo+6hUbmyNZWPWevNkmmI5hvdHQP3rdET75ksdwsNaoucZpJJlmVA/r43/6b2zYb59gDEFbUK4plz6OA3/2GJ55w51NVF3PVCx2tOUw9RdPREY1jxva23qef7+Uzsxy2rX/w9RfPBESGsX4udRxs5aFTSzf1+a5qM96+PiBI/DGmm1YIjs6Hbk8NmxrDcxvD8xbhVWbmjVHuLsx3rTDz1BcME/Zy6v3nK99dimaGjLwiJAXwF3y+cyT6cDVOfTvZoZshP81e0WofuZFocOz24BeznLaWC6Ps/kcbZidBIUSGsP6NQbrTKd0gPZ8dAf45Esew5RLHw+NnNdRb8DUNPRvTM+ioDpIveqzuHnG2/jK3/3ouv8uCQvdPz5VvrxjNlhoOOjTkI1UKP0jaTHi3TNeQWh05AQyHqE+4znj4l2mDTUIrMWiTahKt90iNPJ5EVTOXz68CN+87WW8vNzXZto68pH5HvTG3GwSTH+NXqnT+zR8TI3MFU1FmtA1rzl3+abIh9BU7x549oUbZjnL5WoAlXb1+pqtIcG2S/9G6/5E6cxL81duCV23PSdicxCVQkM2g8G9/R6x6o125AXOvGFWEL+/YmMz3n/5k4GWGucI/9btc3H2rS8FvX1Xcc1w8kG96/3nIkTIHHXds28GAQN6fVBFuPvllaH6mcuLoPHfbYD9+btQ9cb05bhQ39KPPrRPaP2qTS2oz3gY0FQIH3Yl3NTrqs3MWajT4W1K2dMDW1ra8yETkz6/zRZZ19/b3oof3zM/aINM/2elpycvfhTYTkJTfVRo6AO5IpqG/AI68gIZ8hv/uozntOFutDgyt7V2BDbUs299KbJdNWY2n4YtP1BcRFFchJRpnsrlfeftYZc/GeptxabTcITcRswmxmE2m+zH/uinKznn2D2CdU318cn0XLga7C3aefQP3xX3/tCrqyMdBxNd+OjXLbfZuSHrBQ2w0lDbc3kseDfa214lTWHZGJ/GGjl7nerYuBQjU2O68YuH4OL7X0NeiMCU2tqex88fsodDu+pPXojg2kljjr77z7loy+Xxh88eFBwLpM+KoMxTowc3hdav3NSMwX3qQyZY1zejhyjYzLGqPm1r7cAZ18/E+KF9cPFH97UKtpb2XKjO6YJKfeN1GS9UlpxRocrdKTFhTcNB74Zs4NBW6B+JWYHUpnxe+ILDI9RlyCk0TGfzo6+txqSfPBKaVEexYVsrFr67JaiaNp+GbdxIXGqHuPETptDIC4EN29si6nlcw+PyaZgDxVT9VvvHDU7Sn38w8KzI9NWuBlu/Z910157LoyHrYe9d+kaOcYU+K/Ty6neV5qMeN6RgFuufMFiuPusFpkKlofqpRqL7qvuM0zSUD0ylBk8TcgsAw/s3wCP/nSqfk83R/MKS9RBCOAcY5uQ3BCQ/q3+/vBIPzCuMnC424lR1DBuyXkhzXrnRFxr7jxwQrHN9y3oRbZ0e1ajfOWcFnlu8Hje+sAyAXbC1tOdCz1W/ZsE8FfZRmmG2lY66ZaHhoHd9JuI70HvJZgUK8uhLM1HWo0iPQEc3T721frs1LFZx4m+exUm/fS74gGzmKVuisrgGNS46yfRp5PLC6hyPa3iCRI3G/ZsaltnwxKUo0c+lrp3U2zdxahraPesD59pzAvUZL9BsikG/F13rSJMl4LwT9wp+D+5TH7On7whXPWKlabiEr3qPNtNiS3sOf3hycWSb0zxlvKuGbAYkfRoN2UxwTpPP/nUm7ntllVPTUN+Q7RpJKBNy2qPUt6QsA4q31m/H8L6NOPe4PfDdD+wJwF3X9CLaBl26PjWbptHcngvVG7392Cy/ncZsuC7+Y2Z4MinWNKpEb4tPI9wDCJuqVAcjJyt8xvNQn/WcjeB7WuP50d8/H2suUlqEusbrq7fi1D+/EDLNbIgZoW4jLqop6tMQVh9M3OC+YJCZcR3zPOqZqlOZjYTe2NpsymbocxJun4YmNPJhTaMu6wU+lGLQBbN+W2lGvw/rV4jaGaJF8Nioz3oh86h5bR0lHDOW6Kl/zHwHVz76RiTpnSt6yvYsSe4faBqOhnbNlhan0Pj0tf8LxqEUk+csVCbjsMVrtlqnMlam3vqMFwqvbW7PYeJu/ZDxCAePGQjA7dPQNTFbLjLTfAQA5/7jJTxvmf61pT0f2l8XuluMkHoX7NOoEn0aspFQSF31VBXomqffxC8fXoTD9xgs9/F9GlmPkNN6bH0bs6He+lvrC+G8cdlidVRdWLu1FWu3tuKR+atx2vtGAfCdYzZ612esM5DFCg3TPJW3DyaL0zRU6KFpBttk+CBygQ1aNXpRs199xkNbLh8qs2pLip3i1NUGKQcyAKzeXAhzbM/lUZehVJrG+m2toQZe7+3rgiJNOzisb8EBPCRJ09DMU8G1HReJ0zRc6Wf0U3395tlYvGYbbps+NUiQqeMR+eapGE0D8N+pK+p30eqtQaip3kC+tX47dh9ij2Yzy2qa1C55YAGeW7w+sr/qINVlvCAIRbHvbv0BFOq5s2Onm6csGp6KBNPRTWo6bR25wKLRkPVCQlfV9aTUPMUkRiwF1jQcNNVnIppGR8g85b+4xWv8UZ1qPnDfHptHJkMhdVcNzFMsf6+56DkI4vL2uBpPV/K4uIpnNgbb2jpw7bPR2G+b0JixdAOWrd8e9HSTNY3wdvOjm7t8U1BW/aNVzyJOQwMQyRTr6qXpPiZ9oGNbLo+6jIfeDcn9K30cBRDudRaT5LBXXSaUqVhFRrnQzVMKt3nKXecG97YLJ70ReuS1NVi6fjsO/fkTQdjvVacegKf/7xgAfkRQyBHueD/12UxsIIU6Tt3HU4vW4tgrn07MHKDer/ma6zP2pk5946Z5CgAm7toPQGGAnysSUq9TNg3vkdfSz8i3vS2Hf832w5X7NmZDQleZvpLGPrXnBO6as6Ji0yuz0HCQaJ6SL8Q05XTk8wWfhtZzMZ1XALDw3eSJcHTi1E4zKknhij7Z0tyOC+951Rp1ta01vO6sm+fgRctgRFtv9fTrZuCYK58uCA1T04j4NPz/qv0we8ifvKYwbkL/CNRuSSOMzUyxLqGxwaGpKZ9G74YSfBraO9GdleZgNpN+vbJBowsk+zQa6ryIU9mVsl1pGnbB5Ypmir08Dhw9AGOlBuD7NETBMR/zfuLCYpXQUO9Ljfd4Y038NxP4NET4Hk0tQqG08LpMVGgoE6EaCOkaPa5XKdd3mJY/PLkEN/z3LQB+G6QLXfVObYL4wpPDIcPf+9cruHNO+pQtxcBCw0HvBt+sI4TAV258EYf94olQzidlS4/Y//MIoqfqYzQNAJi/cnNRk9fE2TJV78P8OFxza/ztv8twy4x3rAOBzDK9unKz9RyFlM/Rcrk0Dd0RfuQVTxbl6GwNmaf848xG6cP77xo5Lo1pyJU6pb3D1zQGNcU33DZ0oaFrVL9/Mn7wVb/GutC8CIMTfBpN9ZmIecpVVeKEhutd6O930oh+ke366HKS11bP3OXTaGnPxfrEctrxUy59HPfMXQkgrDEcf9XTkeN085Te6zcFgiLQNAyfBlDIwKCiBJ9+3T5oNOQI72SWWd1i0FQf1jRUfbK9p5EDmyLr+qTQjkuBfRoOejdkkcsLtHbk8YQMrdQzcLZ25LF03TYsXhtONdKRz8MjivTC9Z5j1iMM6l2PX/xnIS55YEHqMsU1r23S9p71CHrz5+phuRp1IBpyS2RvhNRHb08Pblel1QhuwDfRFTNJU6vFp2Fe+g+fPQhz3n4C72p+iY68QL18Hy4ntEuY+I5w/30VS9ihrvV6pY/GRb9edaFe+H4j+sdep7EukyprLVBIdqn7jlZuakZ7R94aLgqE372t86P7JjwiCGhC3dE7T+uLem+7n4pDRQc21HlYtn47iGANT9fNU7YQbRPdPFWftWuTSSlXdPNdRz7v/F6Kpa/UNIQQIKJYgVSfjb7/tHWiWFjTcNBbRsvoJiq9AXxu8Xocd9UzWLe1NdQwqxjzjOc5fRoCwKQR/YtWZeM0DdUjNitK1tHDUt+TGrinYzrCXZf1khyEiJqnzDBg1wQ/NnSbcpx/x7SV50KaRuG36lnGzXSofBoDSxAauqah/x7YO37cRT9ZHo+A4/cehgEJ4zR61YX9A2aPWUfVYf2ZHH75kzjmyqedmsbTr68LTGq2ffQ655H/jAPzoSPsu7U9V9IEZfUZD8dc+TSO/tXT1u2B0ED4mbvNU8oRTqh3PLe48UhA1DzVu4RIOxvKJKq+r7j2wqZJlTotQxIsNBwox6ee0KzN0WsaqKUaMMdpKEyfxqTdomp+IpY6sGZLC868YRbWSuFlCo06R29DfVw2LUFFc+ljBT49ZVTgGFQobSou7NXUNExzkjKFpUn7EPZpuD8I8xmoxmn9tlac+udCxtOGbAaLLzsJ155xsPNcHTmBOs9LrWmYAwNtv5M+ZhW8sOhnJ+EvZ06x+sN0ehmaRhqnve29uyKuVm9pwWf/4oer2nbRzUye9Gmo9+MaK9Tcnm6CMrND0mDRdHTUdR9bsAYX3Tc/WO8UGirk1uIIV7g6KMOlzyM0BicvEt9XWtR7PPuWOXh+8fpYTcNW9s76V1yw0HCgKpneU3aZFAb0KjQouZxARy7v+zS0itpohDPum2BysGE2lAICv39yMZ59Yx3ue2WVr2kYja9LRVXnylkqlmro9R54XZYi51LLcaaWNqleK0x/idLkVHx/HLoAimt4TdOgusd/zHwnlGq6sc5vKIbG+AyUeWpgSp9GyAatpyNx/LahghdUKK05mMukV30m9N7T9HSL8Wno5IXAMXsNxR7DCnNIhPwp5Pv1lEBwmaGajZHPLsyG0mVmmrl0A8ae/2BoXpR75ybP0bJdC7l1CY1xQ/pE6n5D1sPMH56AAU11ob5cR15YTXg2ksxH6j0+9fo6fP76mRHz4aG7Dwp+2zWNMuerkbDQcBBMnBIKp7NX8v6appETBU1Db7wajYqUZKe2YfvG1sj8Un0asn7opalpOD4E9S3GNWB6Y5WRPUidIH49ZlR2mzZ/BAC8tz0cPaWiV4iS48/DIbfu/aJjFgqj9XXUqOWGmEa5XZqn0moazSHHZd762yaoFY11HkYMDGd2bayP/0wb6zKhORbSjCkpRtNQCKlBZIhCQsrUNIBCXTUburOOHg9ADmJLYZ4yhYZLY1BZcmc6Jlxy9dK36Y5wx7k9j/BjIzpJ3acal6LI5fOphUaS0OxjmE3XGLnkvnLE7sFvm0mSNY0uRlUKvWKrRuuGL04J7au/sA7l0zDGaZgN0679G3HAqAFFlck2aGetdGwSwWqectlj1dq43og+2IuMjwPQzFNxPo2OfEjYbDbGaeiaR1IlP/+ueYGJKs48FdE0gpQU9l5rnGWsVfqK0moaugBtd4TcxjXOj/6/o/Glw8eG1rnGGChM81RTgnmqjwzyMLniP/HzrG9t7UAu79cFXTB7Fp+G+m7M8SJTxgzEuKG9U5unTPOma2yHq3PkOo9ia0sHsp5/P7pP4wfT9g7tZ/oGVTEIhbr4x6eWYP7KLdb5a0rBNDNuNCL8dCHHPo0aQFVOvdIXzDZ1obh4c5/2XB4ZCpunzIaMiHDvOYcXVSZb+656H81tOdRl7CakX3xiP+c5XXMKmyHDGS+qaSBmSldFW0d4/ggz91RgnqLkQUvb23J4cuFanHbt/7BsQ3SUrSLiCFemOOP0yj5uaiY6Le051Gc89KrP4PNTR8eWD/A1jUJKD7umETe72rB+DZEORpK/p1e94dNIaLSG9WsIlUeRlJlg7ZbWINGg/siyIaHh1xPh8JkR+UIurSPcPN7VWXBpIApXh6S5PRc0uOr/GVPH4Oxjxof2M79fVceIgFtnvoMf3zM/mPipKaWmkUQfY2yQGdWYCQU/FO7/AxOHAyg+BUtaWGg4UHVEV2tVo5bxwo1ztGJDOsI1Fb4M4W/RvEzAmq0qhFJYo6fqPA+fOSTa2KkzuUwlDUZqCnv5/ZxUc+W8HTb8kMHCsmnjDmsayTbYXvWZ2DmfbWVVQt1scJSmEfdq/EbF3+HSj7mFb7B/W2H8gcv5Hfcxl1JPzOipJPPIkD4NJfVC125t8c1TRv3XfxP59V+f3lbHI0JjXSa1pmHiOiTpucV1SNT7VQ2v7VwRTUP9l8/95hlvB9tcjvCRA4ubUMrMd2aab/W+hN7B++3pkwHA2jEoBzxOw4GqOHoPRc13Xed58iP1tx2/z7BIkjcz5FaviLqf5H8XHIdcXuCIXz6VWCZbY6O3g/aQW/vHpCqUqwFryHrWqBidvPBHbNvi5QG/8Vq5qRkX3jPfuh1whzS7iJ3DQxKJnnJkTC0IDfc5d7TlEk0fOi3tOX/cQs4wT2m/Y/0xKScP0mmsC7/3JPPIkD71JfVCN25vRy4v4BGFyqn/VmZM1V61m+9Uaho72jpKElyuYInmhBT5Nn9ZQ9bPQq3GZ6j3bMt0YPoMdPOUSbmip1xpXRShMOtsVIjvNJoGEV1MRCuJaK78+5C27QIiWkJErxPRiZUsh2fpLd4yw09BnPEKH83HDxyBD07cJdhHqclZw6fh6gnt2r8XRg5swpEThgTrxg6Oju4Eoj1lU12tz0ajp1yz66mPyOXTMAeMeRTt5eWFcAoMoOCQvStmalQV8vjetjacdu3/nPsp0gwKc4XcRoSG7JHrj+wbhlmipT3ndJC6yhcMegyN00jX63N1mJ/43tHOY7Je+L0nTVw0uHdDSQ1KTggI4Zvz9HKa9URo5qk2Q5NVmkZLe96pNcSRNNLdha1DooJTlC9DDZDLWDpaZp1SGoZNyLsivIplUkKwjH5lva1RgxHj5qbpDDUnNCRXCyEmy7+HAICIJgI4HcC+AKYB+BMRlUekW1B1wdZDyWq+g8a6TKgXoibMMX0CST3Im79yKD550MjYfc1Gz8zCa4uecg3uUz3AjnzYva6Ob8h6ocY041Gkl5f00ZsRYzaUeWrV5pbQKG4XcbmMJsgwUJemYfoSVE9Of95mlJTKPZWWlvZ8cP2l6wrhn2kbaZf/wlaGEXL+7GyGQu8qyQwysKnOOqgzCSEd3B5FBUXhd3ichvnMCb5m1NKeK2nehzTzodiwmT5V4646BXGahjkqXO1ie111CSPIXZhmxaTvR68rdRkPt371UFzxyf2R8fz6wCG3wCkAbhdCtAoh3gKwBMAhlbpYYJ6y9FCyWk/Lo3DDXK9VPF1l1Cuiq2FQp3HJF9MGrOZwVrgc4TaUMOzIiVDvTTnxGrKmplGInvr7l/3HntTopEnyZyaFTMIlNG7+yiG4bfpUANF5Pgo5e8LH7Crn/04S6HFzodvKp57bVY+9UShDJz9gm7NeNaCmWXLUILumqlD1tegJjrSQ20LIabg++9FEhQ6F+f145Gvg+kRLxZXBvj5p2t+2jnxk5H9B0wgLDdtcI1HzlHSEW64VF1gRhxkqnfUIz5x3DK741P7W/U2fxuF7DAmmSsh65Axy6Sy1KjTOJaJ5RHQDEQ2U60YAWK7ts0Kui0BE04loNhHNXrfOnmQsiYJ5Kvrgs17BSZzxKDTqWvkQMh6FeidpKpL68ONmNNMbsBUbwxFE9caUlYC7YS+kJgh/1Moe21gXNnlkPAoiSg4eMxAHjxmYmF+nl+HIswmwYoWGzXZNBBw5YWgwl4X5zavGyRS6BaFRWGd79nHmqcmjBuAzh4wKlnWhoaMEV6kBEbbDlPaSNRzTfRNCbku1efuD9mTIrXxOZg/c92kUBIL5/RD5zziXFyWlEfnN429Y12+2zPei057L46DRA/ErrQFWA27N6ClbJ8F8bwVNw6aVEMYMbiraTGX6QrIZwpjBvUOma10biUsdk/W8nhVyS0SPE9F8y98pAK4BMB7AZADvAriq2PMLIa4TQkwRQkwZOnRoiWX0/9vU2kymMLjJIwppGqpymWlE+sXkN1KoSuBqWHJ5EXJsbzDjtjPRyXhcpnR1X+b9qd5OQzYTiZ46dcooLLv8ZPRpyAbx+HGYoYe2AUhps/yqEcgtFs3PbOhdIbdmIzm8ny809A/fI+DOsw4LfZxxjvCGrIcB2hiOZofQaO+k0CBLn1alr2mqz4buOekaqlHM5UVianmdnNI0PK2DYzwapZEWfBrRiJ+M51+7lOgpU7tWbGqOn7nSd3h7oWywgaYhG/f6jPv7i9YBqWlYHnXGIzxz3rH42ccmhY9IePVm2LASyHpd1LURl1/JPzY+wWFnqIrQEEKcIISYZPm7VwixRgiRE0LkAfwFBRPUSgCjtNOMlOsqghIKNp9GnUchR5jekGeCht8LZZ48/ZDRmDJmIOJQL95lvsoJEerZbWvpCFUW22Q8LrumcgyaA/NUmF+DEZFjFknNmxCHqW7bbL22WQVtqA/HpmnEpdgGCs/ALK8SRCFNwyNMGTsIvzq10CN1JbLzj6VwOG1OWMujypBUVvd1ouuuPPUAXHLKvpg0ol/oXSWZ03RNY+8fPwzAnvLcRGkQnja4z7wf1ZlwyQOCrxX5mRMSLxmgxh64SJruuD2X92cL1IrbEAgL+T/r1jSi4zT8/7bXmXU8myTFyvw+dL+pQv+tn95sM7IZ6lmaRhxEpE+I8HEAKl7zPgCnE1EDEe0OYAKAWZUqh/oobFEXfpy6+h2uUC5Noy5DmH7UuPhranZiGzk5T4eiuT0XyoBqGxHuMgEENmfjy1V+iIasF+69GpVS2a7jsKnbJmkrdsYjf/pLS9LIJL+jbR6CC0/eB+OGKqERtskDYbNLXAJAzwMO36NgPtAnINJR9uWSNQ1L6zSwqR5nHjZWmosK6zMeYeYPj8etXz00tP8j/+8ozPzh8SFNQ5FmDnTlq9Cjp8x7VeM0XPVOmafyRZqnpu27S+z2JFNbe05EpsVVDbDyPQbmKYtmadZd9Tps5kwVfZWUHdekzkhvrt6THpSiayNxgz4znrfzhNwCuIKIXiWieQCOBfAdABBCvAbgDgALADwM4BwhRGXmM4Q95FaR1Xr0/twZUfNUxqOQWpnRNBKXnyHJpwFEezz6dK62cRpJvTnT5lzwQ7gHcAVlTKiTEU0jZRTS3rv0xTF7hc2KKmLn2meWRvY3n5f5MeUskTx6ChdznIFf1sK6940tJIazXfvoPYfilZ98MLiW3aeRD+6jFGzHuVJ5ZDOE4f0aMW5oeD7tsUOaMLxfIzLyPeh1O02+KuUI90jXqKP1QsAdmeVRQdMoxjyVpgEe4whVB/zOX10mbORTI+9NR3i66ClpFbCVNcV3bMP8PtQ7JSp0QHU/SdzZ6zK08wzuE0KcEbPtMgCXdUU5VL2xCg3dPOWFR36rCpPNUEiV1Md2uCjYiWOEhvHx9G7IIOuRc0R4UtidqUkVUlCIWAex50Vj8E3MEMK0QuOazx8MIQSefv2Z0PXNNNtEvsqfbJ4S+PlDC/HQq6u184XPo18HCL8DMyW8jtpf3asQdm1CaTuuEOgkbHUnpF2EfBpeZB1QMH+oOqoHIaQTGihETzlMML7Z0p7yxt/uP59crjhNI03dmTSif2iiNJ3Wjhzqs16oQ9FgOsKzcSPCo5o2YO/tZzy78EmSIXH3qCZX0zWNuPYkI9uESlCLmkZNEBs9pYW2miGHelRV2P4YDYd1XTNuN7PH07s+G1Roq3kqobOhT2ELFExKQoQrpVkmj6joyKe06ro/A6G9Z6ejsvAmRaZtae7Adc+GNZSw81vXNPz/KkHh148aFzr/Q9860ihX+L+y+fdpyOKAkYXBWeoDLmXEt16u0Dqtr2nzaUS0AGO9HoTQqy65/6h8Gro5zHz2HhXGc9hQWoqpaagxJy7ShD3HDWpsafcTT+qnUfWnznCEp9E0gpBbS7FsWphrwK5f7qy8vrs5VufS94mrSnUZb6cb3Fd1grkijJ74Tz+6rx9ZJF9YZP4KKlQ806af1CNW9cEjwicOskYTWzSNbNCDrM9Ep5ktdgCV6nHmDTOLTeAlRT6ZV06radRnPacNWUeFTJpFM3e1xfBnHAJR/T5gZH88c94xuOBD4ZTYE43Js8yIt9tnvYMla7fhmL2GYpiMzgIKprESFQ1rj9alCWYcQkOh3vE6bW6RNGNq8nkRaFK67y5cznDuKcud+JqGMU7Dlnpev+U0HY64Z9siNQ39Oan60xAZp5FC05CLVvOUEYXVtzGLh759pGVPYGjfhsAnFnePgX+jLr2msdM4wmsFV8jtbrJHVMhyaQiN4IP10GiE0CX1iAt+EuDXp03Gnz53UHQf4xy9G7JBzfV7UuHtxaqoyiEqjGtFe5SULDSMS6cdJFefiY43sadr8Bu6JA1uR1u0nHG5k9T/MYN7R44zIaMerNrcEizrZssg5LaM0VN6i5VG01CosFN9fvs0+ZLyQmlSulYcfU9CCGekkEd+XdIjrKYfNc7aYOr2e9uAu9HGIMa4ubyF8IWDXl4l2IbL8ToFR3i0LGZkU6wj3Hj+g3vXOwMNMkT44Yf2wXF7D4t19iuzZtrsBD0u5LY7UEhYmDfW2/czlzNe9ENMNE8pxxfcZiqz8vauzwRd+jqth67qcrGpBBrr7OYpW2ilaypPF8VoGlHTSnS/hkDTiNdKbOUMhytqv1OVUCuXo+H2KNyIqV5fqaOF7T4Nu7ZU0ALsz3uUTDPyxpqtwbqmFOapwBGu+edsOZmUcLFB5I9xas8JvLx8I/o1ZvHDD+1jnZ1QP4U5bfEvPrFfxKSVZPprrA9PVjVvxWYAwPvHD/avETMi3JaPyr+f6DpTaMdHORFGDWrCDV98HwY0uc1rauCqnjY/NmBmZwq5rRVcPo1ozyp8nF5RzA8h0REut6tsULb9zV5QU30WrVKw9a4vpMjeXyY705MppkFV+DSaRlJW2vdp01EC4Yiku7/xfpx77B7W4+oyntWnsefwPqF1jak1DUuYrkW7MNenwdVwm+N3VOej9BHh0eP0Nfo9qGfn6ngP6l2P3vWZkNBIZZ4Koqco0psulNP/H+vTkDu9/M4mbJGJBm2ajl6/zAAClV9JJ8mE1cuY4VCN/Zgyxq+nalyVNcut473ZBl2azyauVPrzi5tBcvSgQp6x4NoEzPzh8Xj2vGMj+2c9b6dLI1J11Ls0B/e5nIuKoNG1xOtntAbZes3gWFUGi9AwztmnIRN8XL0bssH2CcP7YtHPpuFjB9p9I7Z7CZVRCMPuH+1Rmvzm05Ox9y59g+UpYwZi0c+mBbZj/cM/cPRADO5jT/2c9SjSsyOiiBPapWmYNFvMU67Gu9hcc2Z+LgVRWCvsbBoR2y267jsYJ+C4GSK/d/vGmuLNU3kZreYK2lCNqGueFoI9ICQp5YYpEMzvIE10YqMx78i3j5+ART+bFtx7MfNpKGI1jRQdkGxIaLifgTKV6n4oj/zQ6tEWJ3vWI05Y2NUEmobRm3Y5voNlTyWDi54z7UjdoAyWt2NWXn1qz6b6bCB4skb0lvV6CSYPfbP5zdi+h48dOAIP/7+jCmWVZVDnNNOI9O9lV8eJog59j6L3HsyHYZbNOJ9t1LnrVRSraei7hwcJhjUN5QgvJvmh6zq26+kUxgm4zze4T32oJ5+mgcsHI8IL53ZpGi5fGpG9UU6qqzYhYW5Psn6ak1URhb+RYnJP6eeI7mv4RmIerd6xjJt9UKVJ1028ce83m6GKRU/V3DiNWkFVLlPTCF4yhfdTqHdqk/LpzVPu/SOOcK2H2EfTNNL0aNVkQTrKuTh13GBnL9pfTjx9ZLCiaSt2CQ3APb2mjvrgkxo8myM8aQrVOKaOG4QZS9+LlMv0aegfeGdDbq3HOU6VxpZu8yEkkdf8Mi6/ifo+XL1cIldQQ4KmYVwn64VT9/tCI/4cjXWeVfgqhkjN1xbJFclyq3JPWc5TyBbhOfcJ9tUKFGee+sj+u4Lgf+PPL1nvnzdOaHheqrlnSoE1DQfqo7h37qrQ+uggN3sPyKZppHWEq9G0acxTekVrasgEg8zS9Ghtje34Yb3x3PePxdlHj08eEZ5A1vhozBxOQ/s2uMvm6MHquGbeMxtLu0/Dft0093Xjlw7BvjL0Vj+PqXXo76DzWW5tWqF9X9u7n3PhCaFls2efpk+qCz5VV21JKPV9TXx/SHR90hzf5nUyHoX8CXqWBhe+1uvevsewvnju+8fiwNEDI9uKMWe6BlfaCPk0tHDaUYPCTn4iwkcO2C20T1ynIJvhwX1djut9mBU+6gj3d7DNn5CkPqvvohhNQ89X06chG4RTbmtN7mXYInk8ae/WI2RsZUnTuGaMSC6ztzisb6N5SID5Qdg+EDXzXlJEkk1oOBuBFPfVWJcJMuS6NA0yzGmdd4RH17nnZYmuH9wnLKD1xgdInlccKAg+3TxlmgzV83BHT5WoaWRMTcP4DjKU0hEev49rLhLXxEo2R7grekrtO7xfA47ac2hoH6DwDA4bNxjPff846/VsedJsZL3KpRFhoeHA1XgE6+U3YX6gyhE8vF+0QUxqkCIRSpa3Y4at1mfCaZNVD+XdzfYU0jq2xkVfpffuImVP0faZH4/5UQ9xOMJt2C6nPrK05qmLPjwxWOd6F2mtR4GV0iFYTb+M6vXpz/zCk8MDB+NIGtynEzdeQaFrGt/7wJ44ef9dY/b26dDNU6TMU3aNME7TsGlCpmnm4o9MxM9O2TdYjvg0jNkKs150jJKJ6QgvBlfHxHY6V/SUioq89auHYjc5NiSUpVqbKtpZjpSRfjy4rwq4eoS2uHSds48ejzvPOgxTxw1OfU5Fmt68eQ5dre/TkA16Sqs2pRAatjBObV3cHOepNA2jp2UOTDJ7jx/ef1c86ZgL23a5RoemYe6qNA29YXQVP70ioBoG/Vj9gw4/M9XrU898QFMdTj9kdNqLOUqQXtMwUT4NIuDc4/ZINYamkHRRN08VqWk4ymdqGl88fPeQ+TJN9FSiI7w+3jxVCq5JmML7hLdnPC8iWICC4IwzLbvMoZEyZLyKDe5jR7gDd6Oiupj+P9UIfOrgkajLUDAfQ+yxDgrhruFlHdO2qy831ReERpp5KlzmqcK5daFh7pd4+tAIdyA5jn70oKYgXXkaCj6N+P2Wrtse2t8/xtVzTNeqqGuGzVP69nBoqRrvoyf662wD5ipqGn+WCoNuSmGyUQTaEmkz9znSa7ijp8ha70xzGeA2/dmW/amOExzhJTj/k7A9uUJoffJYlbAj3D24MLheSk0jW0FNg4WGA2cMfKTy+v+vPPWAxHMWUqPbt9tGXUfOYVQoXdOoz3qoz3r43gf2xHH7DEssT4YoyBRru2aceUpf/sG0vXHI7lFBqad2BpIzvMbNW2HDlUbE9S3VZ92akyKt+cKWRsYMV9Yv0WFMwuQZjtxScJpQUwgN5cMwp+SNQ/VcSWv0XIn84qKnbBquLXLINrlZsM0yYVFSeqrGeg+t7eXtfVu/UeM7L0RaReuK1TyVUtOII+t51mSr5YCFhgOXnTzJPGXy5PeODhLmJdneg+ipmHNHo6eiDfE3j58Qex2FGhClj97VK3Zcr0ZfOuvocbHPQW2x5c15/LtHYfpNc7B0/fZU6bl1GlMO7lOkyRCathlXbZb+Okx7s/5McoZPw6P0/hMXnTk+mOq0iImC9Htwjb1Rz2P+yi3Wc+imLR1bPdZ73DZNwzSlZhI6Jb3qMmVvSO2+puROjM3Pl/V87dOVssQ8d+wUCuzT6HoSHeGSJEEwbmifIIQvMWGhehvCndzONAekzedkw/OiPRd31JhdWDYYcxRYr+NwmgJ+mOOew/3ggd5Gr1c/rU07U73T6Gnt5clmks1TaUeEF/KD6R+xtp3C22zmqc4KjVKdugCso/STCO5BM625oqdcEOz1wGae0gVSZDpTm08jlSM8dpeE46NljDVPxbTZah9z+oSGbMaZsgQoInqKQ267HlfjYb7PYtJOJGoaFNY07I7wqHnq16cdgHOOHZ++IFp5TBOJS7C5BvclhUrq+7oaqFY5hWuTkf8oyTavPuK05qlwmew7pbXvq91cmoafLqOwTTmR1T2R5dkXS2eOVuHKxUxJqkxO+uA+s4FLenymr0dhmwtDN1kl+TSyGS/REW7LAl0Mi352Ei6REV1x+eHUe50wvA9O2Gd4YLpWgwZ1Tc3UrhvqvASfRuF33L385CP74oXz7WG7nYXNUw7S+jSKqYTJmobZgEf3Mc0B9RkPnzhoZOoymOUxi+8qouu+61M4FwOfhuPkLdLObH5A/jXcvaV6x+C+NLgEeGrzlOabsG83NI1gYJy8vuXZF0ua4/9y5hQM7xcdRNloGQT6588fDCLg6zfPsZ5Lvwe3IzxB0yD7+zp2r2H40Yf2Qe+GLA7Z3dfMQ6nRTQ3fC4vcNCPCgc4JWuvxMSesy3j46xemBMvXnXEwHlmwBiMHNgUCzqzz/XvVoU9M8kj9McQ96qTBkp2BhYaDtOapYhqspFDIII1ITMLCaNRI6ZVDj4JRuO/bWJaXTaNpqFO6ytqiNA3DPJX1CK22AyQZIhk1YzRciSUCyKlJpnuftuip0HYv6tMgKjSqZqNXCmm0IpXJ1UTN9aK/k2mTdsHqzS3W/YFw2LDNkQskO2p1J3roOI/wtaPGhdaFot0suc/020+Te8o/rjySOi6NiIth/RpxxtQxAArfulnnr//CFAxoco9fcuWG60rYPOXA1WlJUpPjSO8ILyaNSCeEhhcN+3QVMdqj9pdttujIsY5eqUJpGqZPI82zLdXkkLZT4CKYI96xuxk9paaBVetcPe6uItA0Ij4J9zEdmk9DHWY2wok+DUtHxUV9jKZhBmd95wN7pjpvZ8Oc03wvrlDb0HnkicyR+HsM6xvMnWG/Pll/dyUsNBxUxjzl/3c6mw1Nw9Zzsvk0SsUrQtNwhQOnmUlM7evKU9QqE6v1qo+GUcYhIIVGCS2B65C0pyr4NAoH6I5Pm8fCo0IPtRyO8M6g6k2c09VEHxGunnmkEU04hz5GIYk4n0ZeiOBa139hCqaOGxw7Et72vkrBfKtFzqYcoARwsZ0+/TFUq/qw0HDg0gpKyfaqUJXaOU4jRThvxKfRSU3DvITrdlzCsiFFzqJCeKbDPCWFhplEL8lGnRfCn+UvRe/v7m+831qmCCnfp22cRnh7VDvzB7b5v82Q0a5GZayNvJM4TUM5wjUtKdqZSNA0YE8jYkOv27ZnFZ1q2X2uvnIMUGeFRpqipxEkKvgjjaau05kJw8oFCw0HTm2gE+Yptet3P7CnfbsRqmcTXOX0aZh2d7+MxQnLYnwaruipFjmvgyk09MbF9h0K4Ws6SR/Pp6eMimQu7bR5KthfK49WSpsWR9Bs4dVUM1BwapsmQ5Wu3pxKFQhn6s04hGZy9FRyQIgirkM0amA0sWBcJ6OfSsPfycee5rWlUT5aZZ1PO+e3IqRpVKkKsSPcQVwPMryc/s0REZZdfrJze5r5NEwhUeqkPv710t9P1BHp72er9Jd/Yj8s37gjck5XUf9y5sG44fllGGQ4ANP5NKKDxfR5Og4Y2R+//NT+kePc5ql0z7NwT/b9bT4LTzNJdULWl4VDdx+E4/cehh8aSRMbshksu/xk3DbrHVzw71dD25R5irT7KDowhJJ9e4Wy2B/S7dOnon9TXfCe64P0G+5zqZDezvo0zHbBJiDSCAI1Mj2Npq5jJsWsBiw0UvDN4/bA759cAiDaSyrFnu5CNcxqPg1bnYgmQyv++lnPH/hjmyLTFVVkfuiFiKjo9c1EfEn25IPHDMLBY6JpSJLGEAj4Pcg+RvqRH394Il5fvRWvrNjsPLbTWW7ViHDH+yeKBhl4pDvQq6tpNNZlcP0X3+fcfurBI7FpRzuefn0tZr7lTzil557Ki3AIsSJZ06DUY5tcHSL17H56yr6YuGs/HLHHkNB6G30by2OeSjr6wpP3wdRx9txzOso8ZRswGAdHT3UTTpsyKvidJj9Uqdhi0SP7lOGC+ox6ESHo1DTs+6UxjwX2/xLL6UIIgd+efiB+MG3v0Pq+jXX4+tHxgx2dUWKpCxmvPZnjNPxrFgSJ7TmPscz1XC2yGQ9nHzM+ZDIsZLn15woHig8MIYTr+fihvd37Jmj7/Rrr8LWjCils4hzhfQNNo7M+DffxdRnCV4+MT6mjUOapuNn6bOinrlbHgzWNFMTNYJdW1S72OkA681RnrmMbK+Ac3OcQlmnKE0xCU+SjSjK9CQHsPsTe6CSPF3BrCOmIn75VN0UVzl0QnOa7/vVpB+CjB+yW8trVoUNLhZLTTFU6aTQNXYN8RJtTPi3OAZUxVXFQ77pU5UsicrxmnyqmQ1cQGkVqGto1qqWrstBIgd54RWy4ZTVPSZ9GMLgvuk85NQ2bearYQY2uMFodCv4XV/ak6CkRG6ZS2nMqtvcWcoSHQm6jc03rznHzNTZkM0Xlgeoq9Cfcni/MpyEc5qnkcRpa5JVHJd2z6xouTePMw8bgOyfsGVy/M8TdX5rJrxRB9FSRmob+vDl6qobp7FzZqa8TOMLDye10OuP4Dq4TEoLpjol8D0VoGsGxRRZdP7VNPsSJjOC+inw/xT5dpwnFIxy39zD89vTJGC3nOPGocIE0mVCL5T/fPhIPfPOIzp/IQU5LI+IyTyX6wakwgr/UW3Zqw5aqSARccsokDJQCvNM+jZjDi+nQvU/OuTNyYDRKLQ72aXQT9B6EWS/K2TlM0+svh2ajDyJMG3LrEpZpeopfPnx3AMDYwW77tf2aSZqGe1upjUPa4+JSvQAqzQXhlMkjgnlC9NHQxabfSMM+u/bDpBH9O38iDV2bS2OeKkbTKLXR60y4dLl753qYdTEdum8eNwHPff9YjHWYV13oxd+phAYRnUpErxFRnoimGNsuIKIlRPQ6EZ2orZ8m1y0hovO7srzhlNdmz6qc5qnwsnU61jJcR9mUVT6kUBmK9GmkmY/hEweNxLLLT46YaxLLaSmMXoxYTaPEmp32dcaN2gfCprh6bYZBlyO82uM20tCRyjwVfw49y22p9+w6LE1Pv7PCOU7oFNOhy3gUzLJZ6vV3NvPUfACfAPCsvpKIJgI4HcC+AKYB+BMRZYgoA+CPAE4CMBHAZ+S+XUKcrbISjvBgxi/LZfWK8rcvuUMm4xg3xJ9S9fU1W1Ob21zRU8XYpIv1xyQJzTifRuB8L+qKxffeXGlE9FttkM9IT4eeNmqtlsiFQm4R/NZJzHILdNo81ZlZFzsrnOMOL4fpOImQeariV3OUoRoXFUIsFEK8btl0CoDbhRCtQoi3ACwBcIj8WyKEWCqEaANwu9y3S7BVUrWmnB+7Olfc4D5dfh27V/KUrjbU1KzvbW+LVDx3KKrROMj/xURzFStgrc895TkqbZ5SuMNCHZqGGt9RZA+9FihMwoTCOA3Tp5FwDl3TKJVqCljz2npHoRxBKsnXd5elq6g1n8YIAMu15RVynWu9FSKaTkSziWj2unXrOl0oWw8iaNjL+ATNq9h72p2vKP161eHoPYfigpP2TjSTqMVIWYqInlIU+6y+fUJ02tqwphFzrTLby4u9jn4aJTRIJhIBis/ZVAsE4zS8whTBUfNmsk+jkIKktHKkeeaVIu4SXaFp6N9nj/NpENHjRDTf8ldxDUEIcZ0QYooQYsrQoUM7fb44W2UlehdxI8LLUVEIwN+/fAi+fvT4xPOpDzzS4KspaYu04xbD1HGDcd0ZB6sLAgjffz5GapSuaaTbLy59vYlKKxHyaRQZdVQt9EdcmISJnPnRkjoGRLp5qlSfhv24cnSoSr020PWaRk2mESGig+K2CyFeitl2QgnlWQlglLY8Uq5DzPqqUs4eojqV+k7jzDP77tav5OvoTW1S+T0ZX2mWRfU0i+ldleL/UferGqkPTByOh15d7a+LO67oK6nrFbd/mtsPNA3SczaZ56lRqaER+DS8QpbcYh36hEIWgjS33Kchi22tHaF1nU0B0xnKFXJbKrVQT5IG910l/zcCmALgFfjf4/4AZgM4rMzluQ/AP4jo1wB2AzABwCx5zQlEtDt8YXE6gM+W+dolUd56EpYaVp8GATN/eHyQS6ezJFVCV0Ovz62Q+lolPCzziKs/PRkZbx7uf2VVrHmq1FHoaXtvQbBCGk0jEBrhgW06tdAYJKGH2brMU8k+DU17TXHPL1xwHNo6wjMuOafq7YJHGPFpaL/TTDdb7utXg9iWRwhxLAAQ0b8BHCSEeFUuTwJwcakXJaKPA/g9gKEAHiSiuUKIE4UQrxHRHQAWAOgAcI4QIiePORfAIwAyAG4QQrxW6vXLQSVeXZrwVwJheL/Gzl0n5pom139xCm564W00GiNXc7mu0TRMGrIZXHjyPlizuQWnTXHPjd5VPg298dejuXSBps9lrvbuzLws1SZDFMxt3dtIFpns0yhucJ/KThs+h33frmhQq+/TqPglEknbXd1LCQwAEELMJ6J94g6IQwhxN4C7HdsuA3CZZf1DAB4q9ZrdCdXe2Hqx5aiXeu9ovxH9sWj1Vue+7x8/BO8fPySyXmkaxaROKFfKleH9GnHHWfFKbqnXSntUIcLNvl33t+g+DVeW21odp2GbutQjwjePm4Cm+ixOPXhkZJti7136RuqWR5034zgzC3fqrGmv7d62s5in0n7xrxLRX4noGPn3FwDzKlmwnRFVHeLHH5T3mj/72CTcdfb7k3c0yJfgCO/K3nSplyo65DbFlRp0TUPuXokR4V2F5/mp1c85do/IOB11f/0as7h9+tTIsUSFCZxKfUmuZzW0r3tu7XIRne61tBHhpVIL9SSt0PgigNcAfFv+LQDwpQqVqVtQ4tTAsew2oBf6NWZx/kluJa7cPdLGugwOHjOw6OOCuRUqGD3VGVzt0jnHjsf7xrrvt+jH69hfrx/BDHSaT6OSiS8rTZxgVZ2J/Ub2x4AmewYApb2UescuM+eApnosuOREfH5qYT6Xsj9VVa8sJ+6Kd1gLmkaieUqOxv6P9G9cXfki7bw01mUw7+ITY/ephUoDlObT6EzZyyWkzztx79jt5frwQz6NTGFu+IJPI7x/bbzVKDalN+495oyoqnvPORzPLV6HKx99I9hHdR5cae2TiOs4NdVnE+vZjz60Dw4cPaCka8dnuWWfBgBACJGTOaL6CyE2d0WhugPVeneVuu6kEf0wf+WW1Pt3labR1R9Jap+Gip7S14V+az4NqWm05/KBUIpmh62B1iAlcUkAzESOB4wagANGDQgJjb6NdfjLmVNwUMkNd/z2cKBHdOevHTWupOua5wbM6KmuEBrVrydpHeHb4Ps1HgOwXa0UQnyrIqVinFQqqu+Orx+GrS0dyTtKcjJ5XdKUrDrlzNOVROxUGzGkznIbDDa0769fX6Va6cgVEkRGs8MWWdAuolhNI62v6wMTh5dcpjQRWpWi2ppGLZBWaPxb/jGSq06bjKsefR1jikz33VnKMerVdoam+iya6tOP/cglpAa3UQ27fbENSGeK6BJUStPoyOeD92f21mvF7OhCn0MjLmLONE9VpixJQqNil04Y3FdrWZkqQ6pWQgjx90oXpLsxedQA3PyVQ7v8uuX4IMrhHwg0jZ72oZSpwVEjpoGC0GjryAdCqbvknlIaVdbz0BbknnLvX9A0KlemxFQlFTQem68pnLCwYpetKVLdJhFNIKI7iWgBES1Vf5UuXC0xalBxM2xVilqwaQKFCXmKMU91B1I33kVI3l51/kC4lo680zxVI6/Vid5Qx5mepFzpuZpGjEDqcR0oB2ntEX8D8BP40VPHwg+33TmekOSJ7x4TmyCvqyiHhacc35Q+t0JXED8fePkofpyGHb20Q/r44wd8TcPuCK9VTSMwp2nli3vnKr1IJU2RiUKjYleO//66MqS8mqQVGr2EEE8QEQkh3gZwMRHNAXBRBctWUwSx9lWmHKp3WcxTygxRpKZx3ol74f3jB5ehBPGUeo/ps9za1tnTiAzrVxh0pk4fERq1Ub0i2LL5xgmEfBd0JpJOXVmfRvjkej1gR3iYViLyACyWOaBWAuhTuWIxLkqtl8fvPQxPLFpbtnKUqmmcc+weRe3f2Qag+LF6ZRqnoTUnw7SRyq5EirWqaSh0QRH3zkvJFFB0WaoYPVXtLLe1QNr+zbcBNAH4FoCDAXwewBcqVahaYb8R/bHn8NqSjaV+EH/83EH4oAxzLEfV7ihhcF93wDbFrg3bnCe6dqH/7qMl9QvSiHSzkFu9QYzTNA4aPRAA8LEDnXOkdZqkxpnNU5UlrabxnhBiG/zxGl+qYHlqivu/eUS1ixCh1E5UY10Guw3wnfnlME9d9JGJ+On9r+GgElKQdAWVHqehcO0uQvtoDW43S1io0Msb1ziOHdIbyy4/ucJlSdihoo/SHT7FQiPMDUQ0EsCLAJ4D8Kye9ZbpOmrFjLHPrv1w+/RyT6cSZXBv37QzYXjfko4vej6Nkq4SxeW4V+c3e+u18l5NbNl8u3KQpo00Ez1Viji50NO0bhdpx2kcTUT1AN4H4Bj4c2D0EUIMqmThmCjl+F67U9U+YNQA3D59aklJFUuhUo33s+cdCwGBB+a9C6D7mKcUei+61nvUXekI1+HBfRpEdASAI+XfAAAPwNc4mC6mM99rMPlNbX/zEaaOKz7aau9d+6IuQzj3uAlFHVeuYRqmojF6cFPo/N1mulfLyP+aFxoVPLd566HoqR42ZslFWvPU0wDmAPgFgIeEEG0VKxGTQOkV81vHT0Bzew6nv2908s7dnH6NdVh82YeKPi517qkgYWFhf1fCQh21f3dKhQ6EQ4JrvejVGtxXs4K/zKQVGkMAHA7gKADfIqI8gP8JIX5csZIxVjrzwfbvVYeff3y/8hWmB1K0D8TlCHeoIkEakcg4jdpucAoO/Np32ndlGhGdncWnkcoIJ4TYBGApgLcAvAtgPHwBwnQxtf7BdnfSZ7mN8nUt5bbLfOU2T6W6bNVQPphaN00B1UxYWPvPphykzT21FMBVAAYBuAb+nOFHV7JgjJ2dpF5Wjc48368eOQ7fn7YXgDhNwx5ym0ZYXXLKvvjS4WNLL2AnKAi72q+A5PhdDsz719/zzqJppDVP7SGEyFe0JIyV678wBb0bsjj9uhkAusdH253prCY3cqDv8E5KcBkdp5F87jMPGwsA+Nt/l5VStJIw04h0i950lUaE17qJsVykFhpEdA2A4UKISUS0P4CPCiEurWDZGADH71P6ZDVM1/OR/XfFoKZ6HL6HPeKruyUsVKjyVnuMRhpCmkaZixuf5bb2n005SBtY/BcAFwBoBwAhxDwAp1eqUIybWm9cdhYKaUSiI7uPmDDEqbEEZp5uIjSCKDFlVusGDWMlH2U05HbnGxGeVmg0CSFmGevSzw3KlI0abVt2OlRTUezr0KOQwus7XaSKoiYY6g4NI0dPVZa0QmM9EY2H/FaI6FPwo6iYLqZWe6TdnUN375rkBq6EhbUeFdedfBpVGxG+k0zdl9ancQ6A6wDsTUQr4Yfefq5ipWKc1Hjb0m258UuHYMP21tT7f+LAEXhw3rs4YOSAoq7jMvPUalusNKo62SB2B59GJZ+leerQdK/d4NmUg7S5p5YCOIGIesPXTnbA92m8XcGyMRZqtXHp7vSqz2BkfVPq/Y/fZ3hJ2VzV6ysl5LYa9G30mwg1ZW330DQqaZ7ikNtYoUFE/eBrGSMA3Avgcbn8PQDzANxa6QIyPkSqgu4cFbOnUjDz2NfXGledegDunLMCze05PL9kfVXLct+5h+PtDTuqWoY4uoNALQdJRribAewF4FUAXwPwFIBTAXxcCHFKhcvGaOzSr7HaRWDKgGpXShmnUQ0G92nA148eH5inumqudhv7jxyAjxywW9WunwQnLPQZJ4TYDwCI6K/wnd+jhRAtFS8ZE2LEgF54d3MLNje3V7soTCdwjayuVU1DUScbxHz1ZEbNw5qGT9BCCSFyAFaUQ2AQ0alE9BoR5YloirZ+LBE1E9Fc+fdnbdvBRPQqES0hot9RrYeblJnffeZAfO7Q0ThgZP9qF4XpBOSIQqr12pyVaW5zVdQ0ah12hPscQERb5G8C0EsuEwAhhOhX4nXnA/gEgGst294UQky2rL8GvolsJoCHAEwD8J8Sr9/t2G1AL1zGGWq7PQVHeHh9zWsa2eqbp2oR/WkMaKqvWjm6klihIYTIVOKiQoiFQPooByLaFUA/IcQMuXwTgI9hJxIaTM/AnbCwGqVJT53H5ikd83X99vTJmDpu55jINO04ja5kdyJ6GcAWABcKIZ6DH721QttnhVxnhYimA5gOAKNH9/wJh5juQzC4L2KeSi81vvj+sRg9KH14cDlQjvAcSw0rewzrU/MDNMtFxYQGET0OYBfLph8JIe51HKYc7RuI6GAA9xDRvsVeWwhxHfzBiJgyZQrXcqZmcGkaxXDxR4v+JDpNNnCEl/Y53T59Kra2dE3moa40oe2M5rqKCQ0hxAklHNMKoFX+nkNEbwLYE8BKACO1XUfKdQzTrXAlLKx16oOQ29KOL2We9+5EJfNdmTRkPXRUUeOrKfMUEQ0F8J4QIkdE4wBMALBUCPEeEW0hoqnwHeFnAvh9NcvKMKUQRE91M1NGVgqNUjUNpny88pMPVvX6VcmwRUQfJ6IVAA4D8CARPSI3HQVgHhHNBXAngLOEEO/Jbd8A8FcASwC8CXaCM90QV/RUraPMU93Np9GVGkBX0ViXQWNdRWKUUlEVTUMIcTeAuy3r7wJwl+OY2QAmVbhoDFNRPEfCwlqns+apnkI/mYvrsPFhc1s3Uxw7RU2Zpximp+NKI1LrZL3OOcJ7CoP7NOCp/zsGIwbET+fbk2GhwTBdSCHktrrlKBY1uK/bjQivgGzefUjv8p+0G9HNqi7DdG+oDCG31aDO6z7mqa4sY3d4HuWGhQbDdCGu+TRqnZ0lg2uxnLSfPxRtaN+GKpek62DzFMN0IROG98Wew/tgVBeP6O4sdd3NntZFfOu4CfjS4bujf6+6ahely2ChwTBdyO5DeuPR7xxd7WIUTR1rGlY8j3YqgQGweYphmBSwpsEouCYwDJNId/JpdDN3UbeDhQbDMInUs6bBSLgmMAyTSLYbCY2dMQy2K+k+NYFhmKqR7WZpT5jKwUKDYZhE2BHOKLgmMAyTiDnTILPzwuM0GKaKPPado7Bua2u1i8EwqWGhwTBVZMLwvpgwvG+1i5GazxwyutpFYKoMCw2GYVKx7PKTq10EpgZgnwbDMD0KPeKWPTHlh4UGwzAMkxoWGgzDMExqWGgwDNNj4TxU5YeFBsMwDJMaFhoMwzBMalhoMAzDMKlhocEwTI+Cs9xWFhYaDMMwTGpYaDAMwzCpYaHBMEyPgsNsKwsLDYZhGCY1LDQYhmGY1LDQYBiGYVJTFaFBRL8iokVENI+I7iaiAdq2C4hoCRG9TkQnauunyXVLiOj8apSbYZjah0NuK0u1NI3HAEwSQuwP4A0AFwAAEU0EcDqAfQFMA/AnIsoQUQbAHwGcBGAigM/IfRmGYZgupCpCQwjxqBCiQy7OADBS/j4FwO1CiFYhxFsAlgA4RP4tEUIsFUK0Abhd7sswDMN0IbXg0/gygP/I3yMALNe2rZDrXOutENF0IppNRLPXrVtX5uIyDMPsvFRsulciehzALpZNPxJC3Cv3+RGADgC3lvPaQojrAFwHAFOmTGELJ8MwTJmomNAQQpwQt52IvgjgwwCOFyJwXa0EMErbbaRch5j1DMMwVognfC071Yqemgbg+wA+KoTYoW26D8DpRNRARLsDmABgFoAXAUwgot2JqB6+s/y+ri43wzDMzk7FNI0E/gCgAcBj5I/5nyGEOEsI8RoR3QFgAXyz1TlCiBwAENG5AB4BkAFwgxDiteoUnWGYWkagYJHmlCLlpypCQwixR8y2ywBcZln/EICHKlkuhmEYJp5aiJ5iGIZhugksNBiGYZjUsNBgGIZhUsNCg2EYhkkNCw2GYXoUp00ZlbwTUzIsNBiG6VHsNqAXnj3v2GoXo8fCQoNhmB4Hj8+oHCw0GIZhmNSw0GAYpsfBmkblYKHBMEyPg1hqVAwWGgzD9DhYZFQOFhoMw/Q4WNGoHCw0GIbpcfA8GpWDhQbDMD0O1jQqBwsNhmF6HGT8Z8oHCw2GYXoeLC0qBgsNhmF6HOzTqBwsNBiG6bHweI3yw0KDYZgeB8uKysFCg2GYHgfLjMrBQoNhmB4Hm6UqBwsNhmF6HCwyKgcLDYZhehysaFQOFhoMw/Q4OOS2crDQYBim58Eyo2Kw0GAYpsfB5qnKwUKDYZgeB8uMysFCg2EYhkkNCw2GYRgmNSw0GIZhmNRURWgQ0a+IaBERzSOiu4logFw/loiaiWiu/PuzdszBRPQqES0hot8RD/lkGIbpcqqlaTwGYJIQYn8AbwC4QNv2phBisvw7S1t/DYCvAZgg/6Z1WWkZhulWiGoXoAdTFaEhhHhUCNEhF2cAGBm3PxHtCqCfEGKGEEIAuAnAxypbSoZhuivKDNGrPlPVcvREstUuAIAvA/intrw7Eb0MYAuAC4UQzwEYAWCFts8Kuc4KEU0HMB0ARo8eXfYCMwxT2/RtrMMPpu2NE/cdXu2i9DgqJjSI6HEAu1g2/UgIca/c50cAOgDcKre9C2C0EGIDER0M4B4i2rfYawshrgNwHQBMmTKFNVWG2Qk5+5jx1S5Cj6RiQkMIcULcdiL6IoAPAzhempwghGgF0Cp/zyGiNwHsCWAlwiaskXIdwzAM04VUK3pqGoDvA/ioEGKHtn4oEWXk73HwHd5LhRDvAthCRFNl1NSZAO6tQtEZhmF2aqrl0/gDgAYAj8nI2RkyUuooAJcQUTuAPICzhBDvyWO+AeBGAL0A/Ef+MQzDMF1IVYSGEGIPx/q7ANzl2DYbwKRKlothGIaJh0eEMwzDMKlhocEwDMOkhoUGwzAMkxoWGgzDMExqSA6R6LEQ0ToAb5d4+BAA68tYnO4A3/POAd/zzkGp9zxGCDHUtqHHC43OQESzhRBTql2OroTveeeA73nnoBL3zOYphmEYJjUsNBiGYZjUsNCI57pqF6AK8D3vHPA97xyU/Z7Zp8EwDMOkhjUNhmEYJjUsNBiGYZjUsNCwQETTiOh1IlpCROdXuzzlgohuIKK1RDRfWzeIiB4josXy/0C5nojod/IZzCOig6pX8tIholFE9BQRLSCi14jo23J9j71vImokollE9Iq855/K9bsT0Ux5b/8konq5vkEuL5Hbx1b1BjoBEWWI6GUiekAu9+h7JqJlRPQqEc0lotlyXUXrNgsNAzmfxx8BnARgIoDPENHE6paqbNwIYJqx7nwATwghJgB4Qi4D/v1PkH/TAVzTRWUsNx0AvieEmAhgKoBz5PvsyffdCuA4IcQBACYDmEZEUwH8EsDVMsv0RgBfkft/BcBGuf5quV935dsAFmrLO8M9HyuEmKyNx6hs3RZC8J/2B+AwAI9oyxcAuKDa5Srj/Y0FMF9bfh3ArvL3rgBel7+vBfAZ237d+Q/+5F0f2FnuG0ATgJcAHAp/ZHBWrg/qOYBHABwmf2flflTtspdwryNlI3kcgAcA0E5wz8sADDHWVbRus6YRZQSA5dryCrmupzJc+DMjAsBqAMPl7x73HKQJ4kAAM9HD71uaaeYCWAvgMQBvAtgkhOiQu+j3Fdyz3L4ZwOAuLXB5+A38GUHzcnkwev49CwCPEtEcIpou11W0bldr5j6mBhFCCCLqkTHYRNQH/gRf/08IsUXOGAmgZ963ECIHYDIRDQBwN4C9q1uiykJEHwawVggxh4iOqXJxupIjhBAriWgY/JlQF+kbK1G3WdOIshLAKG15pFzXU1lDRLsCgPy/Vq7vMc+BiOrgC4xbhRD/lqt7/H0DgBBiE4Cn4JtmBhCR6ijq9xXcs9zeH8CGri1ppzkcwEeJaBmA2+GbqH6Lnn3PEEKslP/Xwu8cHIIK120WGlFeBDBBRl3UAzgdwH1VLlMluQ/AF+TvL8C3+av1Z8qIi6kANmsqb7eBfJXiegALhRC/1jb12PsmoqFSwwAR9YLvw1kIX3h8Su5m3rN6Fp8C8KSQRu/ughDiAiHESCHEWPjf7JNCiM+hB98zEfUmor7qN4APApiPStftajtyavEPwIcAvAHfDvyjapenjPd1G4B3AbTDt2d+Bb4d9wkAiwE8DmCQ3JfgR5G9CeBVAFOqXf4S7/kI+HbfeQDmyr8P9eT7BrA/gJflPc8HcJFcPw7ALABLAPwLQINc3yiXl8jt46p9D528/2MAPNDT71ne2yvy7zXVVlW6bnMaEYZhGCY1bJ5iGIZhUsNCg2EYhkkNCw2GYRgmNSw0GIZhmNSw0GAYhmFSw0KDYVJARDmZSVT9xWY/JqKziOjMMlx3GREN6ex5GKZccMgtw6SAiLYJIfpU4brL4MfTr+/qazOMDdY0GKYTSE3gCjmnwSwi2kOuv5iI/k/+/hb583nMI6Lb5bpBRHSPXDeDiPaX6wcT0aPkz4PxV/gDstS1Pi+vMZeIrpVJCTNEdCMRzZdl+E4VHgOzE8FCg2HS0cswT31a27ZZCLEfgD/Az7Rqcj6AA4UQ+wM4S677KYCX5bofArhJrv8JgOeFEPvCzyU0GgCIaB8AnwZwuBBiMoAcgM/Bny9jhBBikizD38p1wwxjg7PcMkw6mmVjbeM27f/Vlu3zANxKRPcAuEeuOwLAJwFACPGk1DD6ATgKwCfk+geJaKPc/3gABwN4UWbo7QU/Ed39AMYR0e8BPAjg0RLvj2FSwZoGw3Qe4fitOBl+zp+D4Df6pXTWCMDfhT9D22QhxF5CiIuFEBsBHADgafhazF9LODfDpIaFBsN0nk9r//+nbyAiD8AoIcRTAH4APwV3HwDPwTcvQc7/sF4IsQXAswA+K9efBGCgPNUTAD4l501QPpExMrLKE0LcBeBC+IKJYSoGm6cYJh295Ex4ioeFECrsdiARzYM/N/dnjOMyAG4hov7wtYXfCSE2EdHFAG6Qx+1AIZX1TwHcRkSvAXgBwDsAIIRYQEQXwp+lzYOfqfgcAM0A/ibXAf70xAxTMTjklmE6AYfEMjsbbJ5iGIZhUsOaBsMwDJMa1jQYhmGY1LDQYBiGYVLDQoNhGIZJDQsNhmEYJjUsNBiGYZjU/H8OhnHtUcQSbgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(xValues, yValues)\n",
    "plt.xlabel('Episodes')\n",
    "plt.ylabel('Reward')\n",
    "plt.title(\"Learning Curve of a Random Policy Agent\")\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}