DDPG_Continuous_Pendulum.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import sys\n",
    "sys.path.append('/Users/Jacklin/anaconda3/lib/python3.6/site-packages/')\n",
    "import gym"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 建立DDPG模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DDPG:\n",
    "    \n",
    "    def __init__(\n",
    "        self, \n",
    "        n_actions, \n",
    "        n_states, \n",
    "        gamma = 0.9, #遠見程度\n",
    "        epsilon = None,  #保守程度，越大就越容易用Q值大小來採取行動；越小則越容易產生隨機行動\n",
    "        epsilon_increase = None,\n",
    "        A_learning_rate = 0.001,\n",
    "        C_learning_rate = 0.001,#神經網路的更新率\n",
    "        memory_size = 500, #####\n",
    "        batch_size = 32, #####\n",
    "        nueron_num = 10\n",
    "    ):\n",
    "    \n",
    "        self.n_actions = n_actions\n",
    "        self.n_states = n_states\n",
    "        self.gamma = gamma\n",
    "        self.epsilon_max = epsilon #####\n",
    "        self.epsilon_increase = epsilon_increase #####\n",
    "        self.epsilon = 0 if epsilon_increase is not None else epsilon #####\n",
    "        self.A_lr = A_learning_rate\n",
    "        self.C_lr = C_learning_rate\n",
    "        self.memory_size = memory_size #####\n",
    "        self.memory_counter = 0 #####\n",
    "        self.batch_size = batch_size ####\n",
    "        self.nueron_num = nueron_num\n",
    "        ##### initialize memory\n",
    "        self.memory = np.zeros((self.memory_size, n_states * 2 + 2))\n",
    "        \n",
    "        \n",
    "        tf.reset_default_graph() ## 重新 build graph 需要跑這行\n",
    "        self.sess = tf.Session() #宣告session\n",
    "        #輸入狀態\n",
    "        self.state_input = tf.placeholder(shape = [None, self.n_states], \n",
    "                                          name = 'state_input',\n",
    "                                          dtype = tf.float32)\n",
    "        #輸入行動\n",
    "        self.action_input = tf.placeholder(shape = [None, self.n_actions], \n",
    "                                           name = 'action_input',\n",
    "                                           dtype = tf.float32)\n",
    "        \n",
    "        \n",
    "        self.q_target = tf.placeholder(shape= [None, 1], \n",
    "                                     name=\"q_target\",\n",
    "                                     dtype = tf.float32)\n",
    "        #搭建神經網路\n",
    "        with tf.variable_scope('Actor'):\n",
    "            self.a_eval = self.build_Anetwork(self.nueron_num, Trainable = True, \\\n",
    "                                              scope = 'net_eval')\n",
    "            self.a_target = self.build_Anetwork(self.nueron_num, Trainable = False, \\\n",
    "                                                scope = 'net_target')\n",
    "        \n",
    "        with tf.variable_scope('Critic'):\n",
    "            self.q_eval = self.build_Cnetwork(self.nueron_num,\n",
    "                                              Trainable = True, \n",
    "                                              scope = 'net_eval')\n",
    "            self.q_next = self.build_Cnetwork(self.nueron_num,\n",
    "                                              Trainable = False, \n",
    "                                              scope = 'net_target')    \n",
    "            \n",
    "        \n",
    "        #管理神經網路的parameters\n",
    "        self.Actor_eval_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, \n",
    "                                                   scope='Actor/net_eval')\n",
    "        self.Actor_target_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, \n",
    "                                                     scope='Actor/net_target')\n",
    "        self.Critic_eval_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, \n",
    "                                                    scope='Critic/net_eval')\n",
    "        self.Critic_target_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, \n",
    "                                                      scope='Critic/net_target')\n",
    "        \n",
    "        #更新舊神經網路的參數\n",
    "        self.replace_Anet_params = [tf.assign(old, new) for old, \n",
    "                                   new in zip(self.Actor_target_params, \n",
    "                                   self.Actor_eval_params)]\n",
    "        \n",
    "        self.replace_Cnet_params = [tf.assign(old, new) for old, \n",
    "                                   new in zip(self.Critic_target_params, \n",
    "                                   self.Critic_eval_params)]\n",
    "        \n",
    "        #loss\n",
    "        #q_target = self.reward + self.gamma*self.q_next\n",
    "        #Critic_eval的BP\n",
    "        self.C_loss = tf.reduce_mean(tf.squared_difference(self.q_target, self.q_eval))\n",
    "        self.C_train = tf.train.RMSPropOptimizer(self.C_lr).minimize(self.C_loss, var_list=self.Critic_eval_params)\n",
    "        \n",
    "        #Actor_eval的BP\n",
    "        \"\"\"\n",
    "        使actor往可以提高q值得方向更新，所以得求 (dq/da)*(da/da_params)。\n",
    "        \"\"\"\n",
    "        self.q_a_grad = tf.gradients(ys = self.q_eval, xs = self.action_input)[0]\n",
    "        self.policy_grad = tf.gradients(ys = self.a_eval, \n",
    "                                        xs = self.Actor_eval_params, \n",
    "                                        grad_ys = self.q_a_grad)\n",
    "\n",
    "        self.A_train = tf.train.RMSPropOptimizer(-self.A_lr).apply_gradients(zip(self.policy_grad, self.Actor_eval_params))\n",
    "    \n",
    "        self.sess.run(tf.global_variables_initializer()) #將神經網路初始化\n",
    "    \n",
    "    def write_memory(self, current_state, next_state, reward, action): #####\n",
    "        env_info = np.concatenate([current_state, next_state, [reward], [action]]) \n",
    "        save_index = self.memory_counter%self.memory_size\n",
    "        self.memory[save_index, :] = env_info\n",
    "        self.memory_counter+=1\n",
    "        if self.memory_counter%self.memory_size == 0:\n",
    "            self.memory_counter = self.memory_size\n",
    "    \n",
    "    def build_Anetwork(self, neuron_num, Trainable, scope): \n",
    "        with tf.variable_scope(scope):\n",
    "            init_w = tf.random_normal_initializer(0., 0.3)\n",
    "            init_b = tf.constant_initializer(0.1)\n",
    "            \n",
    "            x1_h1 = tf.layers.dense(inputs = self.state_input, units = neuron_num, \\\n",
    "                   activation = tf.nn.relu, kernel_initializer=init_w, \\\n",
    "                   bias_initializer=init_b, trainable=Trainable) \n",
    "            x1_h2 = tf.layers.dense(inputs = x1_h1, units = neuron_num, \\\n",
    "                   activation = tf.nn.relu, kernel_initializer=init_w, \\\n",
    "                   bias_initializer=init_b, trainable=Trainable)\n",
    "             \n",
    "             #在continuous的版本output為動作的機率分佈，這邊以機率分布的mean值和std\n",
    "             #當成神經網路的輸出值。這邊輸出值為mean1, mean2,... std1, std2,...。\n",
    "             \n",
    "            output = tf.layers.dense(inputs = x1_h2, units = self.n_actions, \\\n",
    "                   activation = tf.nn.tanh, kernel_initializer=init_w, \\\n",
    "                   bias_initializer=init_b, trainable=Trainable)*2\n",
    "\n",
    "        return output #輸出‘不同動作’對應值 \n",
    "    \n",
    "    def build_Cnetwork(self, neuron_num,  Trainable, scope): \n",
    "        with tf.variable_scope(scope):\n",
    "            init_w = tf.random_normal_initializer(0., 0.3)\n",
    "            init_b = tf.constant_initializer(0.1)\n",
    "            \n",
    "            state_action_input = tf.concat([self.state_input, self.action_input], axis = 1)\n",
    "            \n",
    "            x_h1 = tf.layers.dense(inputs = state_action_input, units = neuron_num, \\\n",
    "                   activation = tf.nn.relu, kernel_initializer=init_w, \\\n",
    "                   bias_initializer=init_b, trainable=Trainable) \n",
    "            \n",
    "            #x_h2 = tf.concat([x1_h1, x2_h1], axis = 1)\n",
    "            \n",
    "            #output1 = tf.layers.dense(inputs = x1_h2, units = 1, \\\n",
    "            #                         kernel_initializer=init_w, \\\n",
    "            #                         activation = tf.nn.tanh,\\\n",
    "            #                         bias_initializer=init_b, trainable=Trainable)\n",
    "            \n",
    "            x_h2 = tf.layers.dense(inputs = x_h1, units = neuron_num, \\\n",
    "                   activation = tf.nn.relu, kernel_initializer=init_w, \\\n",
    "                   bias_initializer=init_b, trainable=Trainable) \n",
    "            \n",
    "            #x_h2 = tf.concat([x1_h1, x2_h1], axis = 1)\n",
    "            \n",
    "            #x1_h2 = tf.layers.dense(inputs = x1_h1, units = neuron_num, \\\n",
    "            #       activation = tf.nn.relu, kernel_initializer=init_w, \\\n",
    "            #       bias_initializer=init_b, trainable=Trainable)\n",
    "            \n",
    "            output = tf.layers.dense(inputs = x_h2, units = 1, \\\n",
    "                                     kernel_initializer=init_w, \\\n",
    "                                     bias_initializer=init_b, trainable=Trainable)\n",
    "            #output = output1 + output2\n",
    "\n",
    "            \n",
    "        \n",
    "        return output #輸出Q(s,a)\n",
    "    \n",
    "    def build_dueling_network(self, neuron_num, Trainable, scope):      \n",
    "        \n",
    "        state_action_input = tf.concat([self.state_input, self.action_input], axis = 1)\n",
    "        \n",
    "        with tf.variable_scope(scope):\n",
    "            init_w = tf.random_normal_initializer(0., 0.3)\n",
    "            init_b = tf.constant_initializer(0.1)\n",
    "            h1 = tf.layers.dense(inputs = state_action_input, \n",
    "                                       units = neuron_num, \n",
    "                                       activation = tf.nn.relu, \n",
    "                                       kernel_initializer=init_w, \n",
    "                                       bias_initializer=init_b, \n",
    "                                       trainable=Trainable) \n",
    "            #build state value network\n",
    "            with tf.variable_scope('state_value'):\n",
    "                \n",
    "\n",
    "                v_h2 = tf.layers.dense(inputs = h1, \n",
    "                                       units = neuron_num, \n",
    "                                       activation = tf.nn.relu, \n",
    "                                       kernel_initializer=init_w, \n",
    "                                       bias_initializer=init_b, \n",
    "                                       trainable=Trainable)  \n",
    "\n",
    "                state_value = tf.layers.dense(inputs = v_h2, \n",
    "                                              units = 1, \n",
    "                                              kernel_initializer=init_w, \n",
    "                                              bias_initializer=init_b, \n",
    "                                              trainable=Trainable)\n",
    "            \n",
    "            #build advantage network\n",
    "            with tf.variable_scope('advantage'):\n",
    "                a_h2 = tf.layers.dense(inputs = h1, \n",
    "                                       units = neuron_num, \n",
    "                                       activation = tf.nn.relu, \n",
    "                                       kernel_initializer=init_w, \n",
    "                                       bias_initializer=init_b, \n",
    "                                       trainable=Trainable)  \n",
    "\n",
    "                advantage = tf.layers.dense(inputs = a_h2, \n",
    "                                            units = self.n_actions, \n",
    "                                            kernel_initializer=init_w, \n",
    "                                            bias_initializer=init_b, \n",
    "                                            trainable=Trainable)\n",
    "            #將state value和advantage相加\n",
    "            output = state_value + (advantage - tf.reduce_mean(advantage, axis=1, keep_dims=True))\n",
    "            \n",
    "        return output #輸出‘不同動作’對應的Q值            \n",
    " \n",
    "    def write_memory(self, current_state, next_state, reward, action): #####\n",
    "        env_info = np.concatenate([current_state, next_state, [reward], action]) \n",
    "        save_index = self.memory_counter%self.memory_size\n",
    "        self.memory[save_index, :] = env_info\n",
    "        self.memory_counter+=1\n",
    "        if self.memory_counter%self.memory_size == 0:\n",
    "            self.memory_counter = self.memory_size        \n",
    "    \n",
    "    \n",
    "    def Actor_learn(self, update_params, current_state, action): #####\n",
    "        \n",
    "        if update_params: #####\n",
    "            self.sess.run(self.replace_Anet_params)\n",
    "        \n",
    "        \n",
    "        #a = self.sess.run(self.q_a_grad, feed_dict={self.state_input: current_state[np.newaxis,:],\n",
    "        #                                       self.action_input: self.action})\n",
    "        #print(a)\n",
    "        self.sess.run(self.A_train, feed_dict={self.state_input: current_state[np.newaxis,:],\n",
    "                                               self.action_input: action[np.newaxis,:]})\n",
    "    \n",
    "    def Critic_learn(self, update_params): #####\n",
    "        \n",
    "        if update_params: #####\n",
    "            self.sess.run(self.replace_Cnet_params)\n",
    "            #print('update parameters')\n",
    "        \n",
    "        #####從記憶中抽取過去的經驗來做更新\n",
    "        if self.memory_counter > self.memory_size:\n",
    "            sample_index = np.random.choice(self.memory_size, size=self.batch_size)\n",
    "        else:\n",
    "            sample_index = np.random.choice(self.memory_counter, size=self.batch_size)\n",
    "        past_experience = self.memory[sample_index, :]\n",
    "        \n",
    "        current_state = past_experience[:, 0:self.n_states]\n",
    "        next_state = past_experience[:, self.n_states:self.n_states*2]\n",
    "        reward = past_experience[:, self.n_states*2]\n",
    "        action = past_experience[:, self.n_states*2+1]\n",
    "\n",
    "        #利用actor_taget給出下一步的動作\n",
    "        next_action = self.sess.run(self.a_target, feed_dict={self.state_input: next_state})\n",
    "        q_next = self.sess.run(self.q_next, feed_dict={self.state_input: next_state,\n",
    "                                                       self.action_input: next_action})\n",
    "\n",
    "        q_target = reward[:,np.newaxis] + self.gamma*q_next\n",
    "       \n",
    "        _, self.cost = self.sess.run([self.C_train, self.C_loss], \n",
    "                                      feed_dict={self.state_input: current_state,\n",
    "                                                 self.action_input: action[:, np.newaxis],\n",
    "                                                 self.q_target: q_target})   \n",
    "        # increasing epsilon\n",
    "        #self.epsilon = self.epsilon + self.epsilon_increase if self.epsilon < self.epsilon_max else self.epsilon_max  \n",
    "    \n",
    "    def choose_action(self, current_state):\n",
    "        self.action = self.sess.run(self.a_eval, feed_dict={self.state_input: current_state[np.newaxis,:]})\n",
    "        return self.action[:,0]\n",
    "    def index_generator(self, size):\n",
    "        i = 0\n",
    "        while True:\n",
    "            if (i+1)*self.batch_size< size:\n",
    "                index = np.arange(i*self.batch_size,(i+1)*self.batch_size)\n",
    "                i+=1\n",
    "                yield index\n",
    "            else:\n",
    "                index= np.arange(i*self.batch_size,size)\n",
    "                yield index\n",
    "                break  \n",
    "            \n",
    "        \n",
    "    def model_save(self, model_name):\n",
    "        \n",
    "        saver = tf.train.Saver()\n",
    "        saver.save(self.sess, \"saved_models/{}.ckpt\".format(model_name))\n",
    "    \n",
    "    def model_restore(self, model_name):\n",
    "        \n",
    "        saver = tf.train.Saver()\n",
    "        saver.restore(self.sess, \"saved_models/{}.ckpt\".format(model_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def training(save_model, model_name):\n",
    "    update_count = 0\n",
    "    learn_count = 0\n",
    "    step_record = []\n",
    "    reward_record = []\n",
    "    var = 3\n",
    "    for episode in range(500):\n",
    "        # initial environment並給出起始的state\n",
    "        current_state = env.reset()\n",
    "        step = 0\n",
    "        total_reward = 0\n",
    "        while True:\n",
    "            # 產生環境視窗\n",
    "            if episode >= 70:\n",
    "                env.render()\n",
    "\n",
    "            # 根據現在的狀態選擇動作\n",
    "            action = RL.choose_action(current_state)\n",
    "            \n",
    "            action = np.clip(np.random.normal(action, var), -2, 2) \n",
    "\n",
    "            # 產生動作和環境互動後產生下一個狀態、獎勵值及遊戲是否結束\n",
    "            next_state, reward, done, _ = env.step(action)\n",
    "            \n",
    "            total_reward+= reward\n",
    "            \n",
    "            \n",
    "            if step == 200:\n",
    "                done = True\n",
    "            \n",
    "            if learn_count > 500 : \n",
    "                var *= .9995\n",
    "                if update_count <300:\n",
    "                    \n",
    "                    RL.Actor_learn(False, current_state, action)\n",
    "                    RL.Critic_learn(update_params = False)\n",
    "                    update_count+=1\n",
    "                else: \n",
    "                    RL.Actor_learn(True, current_state, action)\n",
    "                    RL.Critic_learn(update_params = True)\n",
    "                    update_count = 0\n",
    "            else:\n",
    "                learn_count+=1\n",
    "            \n",
    "            # 將資訊存至記憶體\n",
    "            RL.write_memory(current_state, next_state, reward/20, action)\n",
    "       \n",
    "            \n",
    "            current_state = next_state\n",
    "\n",
    "            # break while loop when end of this episode\n",
    "            if done:\n",
    "                mean_reward = total_reward/step\n",
    "                print('episode:{} steps:{} total_reward:{}'.format(episode, step, total_reward))\n",
    "                step_record.append(total_reward)\n",
    "                reward_record.append(mean_reward)\n",
    "                break\n",
    "            step += 1\n",
    "\n",
    "    # end of game\n",
    "    if save_model:\n",
    "        RL.model_save(model_name)\n",
    "    print('game over')\n",
    "    env.close()\n",
    "    return step_record, reward_record"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode:0 steps:200 total_reward:-1107.3947802371297\n",
      "episode:1 steps:200 total_reward:-1087.3822696713494\n",
      "episode:2 steps:200 total_reward:-1213.9739801890146\n",
      "episode:3 steps:200 total_reward:-1620.1186332818572\n",
      "episode:4 steps:200 total_reward:-1227.5932208378406\n",
      "episode:5 steps:200 total_reward:-1437.729067397012\n",
      "episode:6 steps:200 total_reward:-1342.7108087582437\n",
      "episode:7 steps:200 total_reward:-1217.72434150622\n",
      "episode:8 steps:200 total_reward:-1483.0471930578492\n",
      "episode:9 steps:200 total_reward:-1417.2823910399502\n",
      "episode:10 steps:200 total_reward:-1213.366547548377\n",
      "episode:11 steps:200 total_reward:-1470.5848734459607\n",
      "episode:12 steps:200 total_reward:-1345.772814296528\n",
      "episode:13 steps:200 total_reward:-1467.1432030384883\n",
      "episode:14 steps:200 total_reward:-1458.88693717992\n",
      "episode:15 steps:200 total_reward:-1337.788023479924\n",
      "episode:16 steps:200 total_reward:-1063.16705192476\n",
      "episode:17 steps:200 total_reward:-1405.6012070762854\n",
      "episode:18 steps:200 total_reward:-1407.3588249164457\n",
      "episode:19 steps:200 total_reward:-1515.1830715017563\n",
      "episode:20 steps:200 total_reward:-1548.7664144549879\n",
      "episode:21 steps:200 total_reward:-1350.3215550558439\n",
      "episode:22 steps:200 total_reward:-1225.6631164071741\n",
      "episode:23 steps:200 total_reward:-1536.7839751699703\n",
      "episode:24 steps:200 total_reward:-1281.238434181241\n",
      "episode:25 steps:200 total_reward:-1421.4324535653968\n",
      "episode:26 steps:200 total_reward:-602.2373915686425\n",
      "episode:27 steps:200 total_reward:-1285.7739503434498\n",
      "episode:28 steps:200 total_reward:-968.1044752114276\n",
      "episode:29 steps:200 total_reward:-823.2695675932474\n",
      "episode:30 steps:200 total_reward:-669.4814252226536\n",
      "episode:31 steps:200 total_reward:-891.8870251685845\n",
      "episode:32 steps:200 total_reward:-793.2862308351131\n",
      "episode:33 steps:200 total_reward:-792.9024066853646\n",
      "episode:34 steps:200 total_reward:-933.4783865130553\n",
      "episode:35 steps:200 total_reward:-1269.316369933273\n",
      "episode:36 steps:200 total_reward:-942.1622807368341\n",
      "episode:37 steps:200 total_reward:-1527.1914175487568\n",
      "episode:38 steps:200 total_reward:-734.2168095219411\n",
      "episode:39 steps:200 total_reward:-960.6406240381367\n",
      "episode:40 steps:200 total_reward:-659.3955616560038\n",
      "episode:41 steps:200 total_reward:-1526.080928191452\n",
      "episode:42 steps:200 total_reward:-134.16817157526413\n",
      "episode:43 steps:200 total_reward:-1249.8690182814294\n",
      "episode:44 steps:200 total_reward:-681.3315894758258\n",
      "episode:45 steps:200 total_reward:-1507.4506144353466\n",
      "episode:46 steps:200 total_reward:-0.20768952492201007\n",
      "episode:47 steps:200 total_reward:-1110.3301058178354\n",
      "episode:48 steps:200 total_reward:-0.28241737963679936\n",
      "episode:49 steps:200 total_reward:-391.6266487119586\n",
      "episode:50 steps:200 total_reward:-132.15300504009542\n",
      "episode:51 steps:200 total_reward:-0.16168920106447357\n",
      "episode:52 steps:200 total_reward:-1079.3303614179797\n",
      "episode:53 steps:200 total_reward:-445.9248422593595\n",
      "episode:54 steps:200 total_reward:-1501.9479099825535\n",
      "episode:55 steps:200 total_reward:-131.08597777596478\n",
      "episode:56 steps:200 total_reward:-383.6473158406118\n",
      "episode:57 steps:200 total_reward:-130.59716710760318\n",
      "episode:58 steps:200 total_reward:-3.1406869752081055\n",
      "episode:59 steps:200 total_reward:-268.8592706645082\n",
      "episode:60 steps:200 total_reward:-403.63368230048667\n",
      "episode:61 steps:200 total_reward:-272.8627780595619\n",
      "episode:62 steps:200 total_reward:-970.3446194015886\n",
      "episode:63 steps:200 total_reward:-405.9840237177132\n",
      "episode:64 steps:200 total_reward:-276.08215084802123\n",
      "episode:65 steps:200 total_reward:-268.78269865866855\n",
      "episode:66 steps:200 total_reward:-396.0673929327215\n",
      "episode:67 steps:200 total_reward:-1480.5715562868063\n",
      "episode:68 steps:200 total_reward:-1506.1859849993764\n",
      "episode:69 steps:200 total_reward:-1523.2960315659654\n",
      "episode:70 steps:200 total_reward:-1529.2630165585276\n",
      "episode:71 steps:200 total_reward:-1515.2165288081687\n",
      "episode:72 steps:200 total_reward:-1638.3500110780928\n",
      "episode:73 steps:200 total_reward:-1657.9981288915942\n",
      "episode:74 steps:200 total_reward:-1386.6764265385232\n",
      "episode:75 steps:200 total_reward:-1548.682421257977\n",
      "episode:76 steps:200 total_reward:-1500.0282909621287\n",
      "episode:77 steps:200 total_reward:-1549.3370989899684\n",
      "episode:78 steps:200 total_reward:-1546.4154480133836\n",
      "episode:79 steps:200 total_reward:-1557.6686477320145\n",
      "episode:80 steps:200 total_reward:-1470.4904550631406\n",
      "episode:81 steps:200 total_reward:-1611.723552360116\n",
      "episode:82 steps:200 total_reward:-1535.9931318450438\n",
      "episode:83 steps:200 total_reward:-1223.1647610797074\n",
      "episode:84 steps:200 total_reward:-1431.653180482147\n",
      "episode:85 steps:200 total_reward:-1481.495029684159\n",
      "episode:86 steps:200 total_reward:-1498.8117878099274\n",
      "episode:87 steps:200 total_reward:-1524.14738247747\n",
      "episode:88 steps:200 total_reward:-1520.5955819984224\n",
      "episode:89 steps:200 total_reward:-537.4185880843962\n",
      "episode:90 steps:200 total_reward:-431.51904168030126\n",
      "episode:91 steps:200 total_reward:-132.78978893107518\n",
      "episode:92 steps:200 total_reward:-692.7121265837501\n",
      "episode:93 steps:200 total_reward:-542.4436832884663\n",
      "episode:94 steps:200 total_reward:-1.4339784942046492\n",
      "episode:95 steps:200 total_reward:-406.4654150158912\n",
      "episode:96 steps:200 total_reward:-271.4395569197187\n",
      "episode:97 steps:200 total_reward:-131.96808001517297\n",
      "episode:98 steps:200 total_reward:-133.113878828633\n",
      "episode:99 steps:200 total_reward:-534.1295846940662\n",
      "episode:100 steps:200 total_reward:-409.5576664982435\n",
      "episode:101 steps:200 total_reward:-273.7955260759935\n",
      "episode:102 steps:200 total_reward:-528.5314632629136\n",
      "episode:103 steps:200 total_reward:-0.47638657947108054\n",
      "episode:104 steps:200 total_reward:-939.4172598680368\n",
      "episode:105 steps:200 total_reward:-0.5722565854511817\n",
      "episode:106 steps:200 total_reward:-2.671653410373582\n",
      "episode:107 steps:200 total_reward:-129.53277271262317\n",
      "episode:108 steps:200 total_reward:-1200.6057245374068\n",
      "episode:109 steps:200 total_reward:-134.6251895173566\n",
      "episode:110 steps:200 total_reward:-134.38175384929443\n",
      "episode:111 steps:200 total_reward:-905.9306836187393\n",
      "episode:112 steps:200 total_reward:-959.8535446906775\n",
      "episode:113 steps:200 total_reward:-405.91538308699825\n",
      "episode:114 steps:200 total_reward:-0.7165973473798619\n",
      "episode:115 steps:200 total_reward:-687.5364184017155\n",
      "episode:116 steps:200 total_reward:-1098.7445964922347\n",
      "episode:117 steps:200 total_reward:-1057.387081295719\n",
      "episode:118 steps:200 total_reward:-1.3474104152375364\n",
      "episode:119 steps:200 total_reward:-1394.6229744119694\n",
      "episode:120 steps:200 total_reward:-129.22519285549473\n",
      "episode:121 steps:200 total_reward:-1250.9440961018054\n",
      "episode:122 steps:200 total_reward:-382.07954194264994\n",
      "episode:123 steps:200 total_reward:-0.1644397514218019\n",
      "episode:124 steps:200 total_reward:-131.42881255353387\n",
      "episode:125 steps:200 total_reward:-266.8177437923393\n",
      "episode:126 steps:200 total_reward:-409.4651979166141\n",
      "episode:127 steps:200 total_reward:-417.3632750693609\n",
      "episode:128 steps:200 total_reward:-0.5223545466175834\n",
      "episode:129 steps:200 total_reward:-1.280089453652193\n",
      "episode:130 steps:200 total_reward:-0.7784311094162633\n",
      "episode:131 steps:200 total_reward:-1132.7343948006444\n",
      "episode:132 steps:200 total_reward:-959.1236928061616\n",
      "episode:133 steps:200 total_reward:-965.5394222307665\n",
      "episode:134 steps:200 total_reward:-1061.911323432408\n",
      "episode:135 steps:200 total_reward:-950.0162582873887\n",
      "episode:136 steps:200 total_reward:-439.85037615218954\n",
      "episode:137 steps:200 total_reward:-412.80469887009616\n",
      "episode:138 steps:200 total_reward:-575.7830217557652\n",
      "episode:139 steps:200 total_reward:-387.17671703065764\n",
      "episode:140 steps:200 total_reward:-269.55666531822243\n",
      "episode:141 steps:200 total_reward:-410.79161141971895\n",
      "episode:142 steps:200 total_reward:-134.82145345264215\n",
      "episode:143 steps:200 total_reward:-528.5827286151457\n",
      "episode:144 steps:200 total_reward:-265.2130766231833\n",
      "episode:145 steps:200 total_reward:-282.23057309368835\n",
      "episode:146 steps:200 total_reward:-375.5255583879841\n",
      "episode:147 steps:200 total_reward:-130.23493590891675\n",
      "episode:148 steps:200 total_reward:-415.58973631061525\n",
      "episode:149 steps:200 total_reward:-137.37384419379592\n",
      "episode:150 steps:200 total_reward:-420.44156116031655\n",
      "episode:151 steps:200 total_reward:-521.7165233063138\n",
      "episode:152 steps:200 total_reward:-410.09406346531256\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode:153 steps:200 total_reward:-0.32196446905369397\n",
      "episode:154 steps:200 total_reward:-987.1682006033841\n",
      "episode:155 steps:200 total_reward:-129.60879945020253\n",
      "episode:156 steps:200 total_reward:-970.1743370722751\n",
      "episode:157 steps:200 total_reward:-654.789902141506\n",
      "episode:158 steps:200 total_reward:-270.78664776580325\n",
      "episode:159 steps:200 total_reward:-750.7325153525244\n",
      "episode:160 steps:200 total_reward:-273.7736690675931\n",
      "episode:161 steps:200 total_reward:-422.3095303995601\n",
      "episode:162 steps:200 total_reward:-266.18684134565865\n",
      "episode:163 steps:200 total_reward:-1.9593904596603988\n",
      "episode:164 steps:200 total_reward:-133.9181039702078\n",
      "episode:165 steps:200 total_reward:-128.781751654639\n",
      "episode:166 steps:200 total_reward:-810.3083780280856\n",
      "episode:167 steps:200 total_reward:-273.5348953083545\n",
      "episode:168 steps:200 total_reward:-274.3869183768352\n",
      "episode:169 steps:200 total_reward:-1272.9585746279722\n",
      "episode:170 steps:200 total_reward:-280.15460843290435\n",
      "episode:171 steps:200 total_reward:-1125.801964446057\n",
      "episode:172 steps:200 total_reward:-130.12484230871914\n",
      "episode:173 steps:200 total_reward:-133.6748121503362\n",
      "episode:174 steps:200 total_reward:-132.48355031995433\n",
      "episode:175 steps:200 total_reward:-133.3758026699036\n",
      "episode:176 steps:200 total_reward:-280.9245542144458\n",
      "episode:177 steps:200 total_reward:-274.2481802117269\n",
      "episode:178 steps:200 total_reward:-271.28823637093114\n",
      "episode:179 steps:200 total_reward:-1141.6050819311386\n",
      "episode:180 steps:200 total_reward:-1255.6801590118926\n",
      "episode:181 steps:200 total_reward:-1094.463725402063\n",
      "episode:182 steps:200 total_reward:-269.3066213008572\n",
      "episode:183 steps:200 total_reward:-266.64400550344993\n",
      "episode:184 steps:200 total_reward:-533.035010889385\n",
      "episode:185 steps:200 total_reward:-133.27468531394598\n",
      "episode:186 steps:200 total_reward:-131.09839327776638\n",
      "episode:187 steps:200 total_reward:-133.2713065249887\n",
      "episode:188 steps:200 total_reward:-130.95623260821125\n",
      "episode:189 steps:200 total_reward:-400.0672752328545\n",
      "episode:190 steps:200 total_reward:-426.1608827086122\n",
      "episode:191 steps:200 total_reward:-496.3177373338237\n",
      "episode:192 steps:200 total_reward:-405.1763654970369\n",
      "episode:193 steps:200 total_reward:-2.1849002845481325\n",
      "episode:194 steps:200 total_reward:-509.4654345377353\n",
      "episode:195 steps:200 total_reward:-416.4096150890552\n",
      "episode:196 steps:200 total_reward:-276.06176369854654\n",
      "episode:197 steps:200 total_reward:-0.5655466594388692\n",
      "episode:198 steps:200 total_reward:-264.1104729162097\n",
      "episode:199 steps:200 total_reward:-265.12339632543507\n",
      "episode:200 steps:200 total_reward:-133.57035608182645\n",
      "episode:201 steps:200 total_reward:-0.838156330635117\n",
      "episode:202 steps:200 total_reward:-137.4734615976936\n",
      "episode:203 steps:200 total_reward:-282.14740238012615\n",
      "episode:204 steps:200 total_reward:-270.7090647796741\n",
      "episode:205 steps:200 total_reward:-132.11817579385098\n",
      "episode:206 steps:200 total_reward:-556.3441004553151\n",
      "episode:207 steps:200 total_reward:-135.38527357691527\n",
      "episode:208 steps:200 total_reward:-402.4751038033301\n",
      "episode:209 steps:200 total_reward:-1.337416666712674\n",
      "episode:210 steps:200 total_reward:-403.4535943226303\n",
      "episode:211 steps:200 total_reward:-2.0468008478507675\n",
      "episode:212 steps:200 total_reward:-135.70120300446672\n",
      "episode:213 steps:200 total_reward:-274.64215955383537\n",
      "episode:214 steps:200 total_reward:-539.9133147155823\n",
      "episode:215 steps:200 total_reward:-273.52016690667153\n",
      "episode:216 steps:200 total_reward:-272.640079969218\n",
      "episode:217 steps:200 total_reward:-137.41233195510927\n",
      "episode:218 steps:200 total_reward:-410.49661683780243\n",
      "episode:219 steps:200 total_reward:-713.1307360179115\n",
      "episode:220 steps:200 total_reward:-700.6041414720645\n",
      "episode:221 steps:200 total_reward:-928.5519950063879\n",
      "episode:222 steps:200 total_reward:-940.6160983637217\n",
      "episode:223 steps:200 total_reward:-811.9795193351969\n",
      "episode:224 steps:200 total_reward:-808.3619480917873\n",
      "episode:225 steps:200 total_reward:-544.7513195292081\n",
      "episode:226 steps:200 total_reward:-132.8236382740416\n",
      "episode:227 steps:200 total_reward:-1251.1705696975705\n",
      "episode:228 steps:200 total_reward:-1.7614912935284741\n",
      "episode:229 steps:200 total_reward:-135.73252867591816\n",
      "episode:230 steps:200 total_reward:-0.7232890431364647\n",
      "episode:231 steps:200 total_reward:-807.6842701482623\n",
      "episode:232 steps:200 total_reward:-668.7062680617667\n",
      "episode:233 steps:200 total_reward:-399.0518296057735\n",
      "episode:234 steps:200 total_reward:-1076.0714011744203\n",
      "episode:235 steps:200 total_reward:-1133.9236889826489\n",
      "episode:236 steps:200 total_reward:-576.3549558041121\n",
      "episode:237 steps:200 total_reward:-0.23290709454243774\n",
      "episode:238 steps:200 total_reward:-653.8967394396398\n",
      "episode:239 steps:200 total_reward:-269.05309020367315\n",
      "episode:240 steps:200 total_reward:-543.5846179109557\n",
      "episode:241 steps:200 total_reward:-1075.8849473021792\n",
      "episode:242 steps:200 total_reward:-783.6552607290117\n",
      "episode:243 steps:200 total_reward:-2.292884131351099\n",
      "episode:244 steps:200 total_reward:-4.624510340217558\n",
      "episode:245 steps:200 total_reward:-133.88182288624787\n",
      "episode:246 steps:200 total_reward:-4.258638173096404\n",
      "episode:247 steps:200 total_reward:-127.71922941158535\n",
      "episode:248 steps:200 total_reward:-3.4138564957076274\n",
      "episode:249 steps:200 total_reward:-299.04820368442466\n",
      "episode:250 steps:200 total_reward:-382.7396721889747\n",
      "episode:251 steps:200 total_reward:-421.5005108659404\n",
      "episode:252 steps:200 total_reward:-121.98211798778974\n",
      "episode:253 steps:200 total_reward:-433.0654083437584\n",
      "episode:254 steps:200 total_reward:-263.2938082776178\n",
      "episode:255 steps:200 total_reward:-1071.1985650529834\n",
      "episode:256 steps:200 total_reward:-1037.019363831135\n",
      "episode:257 steps:200 total_reward:-1026.104817169867\n",
      "episode:258 steps:200 total_reward:-1016.5568578672268\n",
      "episode:259 steps:200 total_reward:-945.3197423615178\n",
      "episode:260 steps:200 total_reward:-954.1979613358809\n",
      "episode:261 steps:200 total_reward:-6.121711093092703\n",
      "episode:262 steps:200 total_reward:-275.9504757023254\n",
      "episode:263 steps:200 total_reward:-137.36692783337605\n",
      "episode:264 steps:200 total_reward:-278.6541755344665\n",
      "episode:265 steps:200 total_reward:-5.629415347742028\n",
      "episode:266 steps:200 total_reward:-732.0344946869297\n",
      "episode:267 steps:200 total_reward:-1515.5545242388587\n",
      "episode:268 steps:200 total_reward:-569.904968795957\n",
      "episode:269 steps:200 total_reward:-279.6466260310772\n",
      "episode:270 steps:200 total_reward:-6.893176354298975\n",
      "episode:271 steps:200 total_reward:-1502.2598660469362\n",
      "episode:272 steps:200 total_reward:-8.14203897375704\n",
      "episode:273 steps:200 total_reward:-550.1507858272895\n",
      "episode:274 steps:200 total_reward:-276.4689251493521\n",
      "episode:275 steps:200 total_reward:-832.1943079778423\n",
      "episode:276 steps:200 total_reward:-140.12726227151546\n",
      "episode:277 steps:200 total_reward:-666.555102403651\n",
      "episode:278 steps:200 total_reward:-9.767563116203519\n",
      "episode:279 steps:200 total_reward:-684.9231371189283\n",
      "episode:280 steps:200 total_reward:-1499.4465688346897\n",
      "episode:281 steps:200 total_reward:-131.74495209685242\n",
      "episode:282 steps:200 total_reward:-1285.9907346189966\n",
      "episode:283 steps:200 total_reward:-421.4660536377018\n",
      "episode:284 steps:200 total_reward:-417.1735080791483\n",
      "episode:285 steps:200 total_reward:-948.0859602675052\n",
      "episode:286 steps:200 total_reward:-800.7911617261601\n",
      "episode:287 steps:200 total_reward:-1507.4430658486915\n",
      "episode:288 steps:200 total_reward:-136.72957980164688\n",
      "episode:289 steps:200 total_reward:-6.430740804163348\n",
      "episode:290 steps:200 total_reward:-550.2405259438702\n",
      "episode:291 steps:200 total_reward:-6.0814334452838885\n",
      "episode:292 steps:200 total_reward:-307.1232990563035\n",
      "episode:293 steps:200 total_reward:-1087.3458488474373\n",
      "episode:294 steps:200 total_reward:-138.57571105790637\n",
      "episode:295 steps:200 total_reward:-290.71283865407264\n",
      "episode:296 steps:200 total_reward:-397.0392090693711\n",
      "episode:297 steps:200 total_reward:-268.5099340986485\n",
      "episode:298 steps:200 total_reward:-6.921888412735283\n",
      "episode:299 steps:200 total_reward:-1508.667738499471\n",
      "episode:300 steps:200 total_reward:-283.4068558563959\n",
      "episode:301 steps:200 total_reward:-1038.6266438741382\n",
      "episode:302 steps:200 total_reward:-670.9375114025129\n",
      "episode:303 steps:200 total_reward:-416.47290680292895\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode:304 steps:200 total_reward:-1502.911336313096\n",
      "episode:305 steps:200 total_reward:-133.27297679920667\n",
      "episode:306 steps:200 total_reward:-3.4270523004896956\n",
      "episode:307 steps:200 total_reward:-1347.4834248986915\n",
      "episode:308 steps:200 total_reward:-1156.6452300980152\n",
      "episode:309 steps:200 total_reward:-1201.9739546477097\n",
      "episode:310 steps:200 total_reward:-300.770565700266\n",
      "episode:311 steps:200 total_reward:-1087.3899257722296\n",
      "episode:312 steps:200 total_reward:-1508.4171192077763\n",
      "episode:313 steps:200 total_reward:-1140.663305384268\n",
      "episode:314 steps:200 total_reward:-1056.1469677774664\n",
      "episode:315 steps:200 total_reward:-1509.32161551157\n",
      "episode:316 steps:200 total_reward:-657.2395165261529\n",
      "episode:317 steps:200 total_reward:-1516.8338722802291\n",
      "episode:318 steps:200 total_reward:-1504.8794080448213\n",
      "episode:319 steps:200 total_reward:-1648.1139282062077\n",
      "episode:320 steps:200 total_reward:-1531.2761306008065\n",
      "episode:321 steps:200 total_reward:-1533.4227190982663\n",
      "episode:322 steps:200 total_reward:-1645.1987209640267\n",
      "episode:323 steps:200 total_reward:-1510.3324381281134\n",
      "episode:324 steps:200 total_reward:-1642.5976104475576\n",
      "episode:325 steps:200 total_reward:-1543.2628803148662\n",
      "episode:326 steps:200 total_reward:-1488.8425686805847\n",
      "episode:327 steps:200 total_reward:-1616.6310437984275\n",
      "episode:328 steps:200 total_reward:-1573.0867233765862\n",
      "episode:329 steps:200 total_reward:-1655.9500066681487\n",
      "episode:330 steps:200 total_reward:-1514.9635597521174\n",
      "episode:331 steps:200 total_reward:-1542.2830452894777\n",
      "episode:332 steps:200 total_reward:-1567.5279199537829\n",
      "episode:333 steps:200 total_reward:-1413.6399718917146\n",
      "episode:334 steps:200 total_reward:-1514.7811037120387\n",
      "episode:335 steps:200 total_reward:-1550.6080621381925\n",
      "episode:336 steps:200 total_reward:-1652.295377758995\n",
      "episode:337 steps:200 total_reward:-1535.9508427660326\n",
      "episode:338 steps:200 total_reward:-1382.513432257689\n",
      "episode:339 steps:200 total_reward:-1606.745587191732\n",
      "episode:340 steps:200 total_reward:-1533.0067313293093\n",
      "episode:341 steps:200 total_reward:-1520.5120981754492\n",
      "episode:342 steps:200 total_reward:-1465.5257163018568\n",
      "episode:343 steps:200 total_reward:-1451.2227614081835\n",
      "episode:344 steps:200 total_reward:-1219.7235757556805\n",
      "episode:345 steps:200 total_reward:-405.9704378901821\n",
      "episode:346 steps:200 total_reward:-1307.6097211600795\n",
      "episode:347 steps:200 total_reward:-1184.351826896739\n",
      "episode:348 steps:200 total_reward:-1419.8617008489807\n",
      "episode:349 steps:200 total_reward:-691.2077869763347\n",
      "episode:350 steps:200 total_reward:-1396.4242081406987\n",
      "episode:351 steps:200 total_reward:-139.4308199574946\n",
      "episode:352 steps:200 total_reward:-6.096481787490108\n",
      "episode:353 steps:200 total_reward:-918.1071335706748\n",
      "episode:354 steps:200 total_reward:-1369.3402646083368\n",
      "episode:355 steps:200 total_reward:-1502.640765044162\n",
      "episode:356 steps:200 total_reward:-137.66347703509186\n",
      "episode:357 steps:200 total_reward:-131.42665163226283\n",
      "episode:358 steps:200 total_reward:-779.1603599499032\n",
      "episode:359 steps:200 total_reward:-668.2358709555459\n",
      "episode:360 steps:200 total_reward:-117.19849851322222\n",
      "episode:361 steps:200 total_reward:-932.8319951847935\n",
      "episode:362 steps:200 total_reward:-1477.9412918271669\n",
      "episode:363 steps:200 total_reward:-130.40624260684277\n",
      "episode:364 steps:200 total_reward:-275.6830525251936\n",
      "episode:365 steps:200 total_reward:-1521.9443101272016\n",
      "episode:366 steps:200 total_reward:-0.09926932450636736\n",
      "episode:367 steps:200 total_reward:-1216.445821640056\n",
      "episode:368 steps:200 total_reward:-3.135779610842097\n",
      "episode:369 steps:200 total_reward:-801.7361264615855\n",
      "episode:370 steps:200 total_reward:-1464.1851585455\n",
      "episode:371 steps:200 total_reward:-5.299052636693963\n",
      "episode:372 steps:200 total_reward:-1341.298177014325\n",
      "episode:373 steps:200 total_reward:-129.57460437547607\n",
      "episode:374 steps:200 total_reward:-3.9238465058200247\n",
      "episode:375 steps:200 total_reward:-1368.9445381340774\n",
      "episode:376 steps:200 total_reward:-5.3007761232414925\n",
      "episode:377 steps:200 total_reward:-1332.50139824217\n",
      "episode:378 steps:200 total_reward:-257.31084197028144\n",
      "episode:379 steps:200 total_reward:-442.6379539742842\n",
      "episode:380 steps:200 total_reward:-282.92162444583505\n",
      "episode:381 steps:200 total_reward:-404.6810995158603\n",
      "episode:382 steps:200 total_reward:-1228.173913205903\n",
      "episode:383 steps:200 total_reward:-131.939030877572\n",
      "episode:384 steps:200 total_reward:-0.6245276237488097\n",
      "episode:385 steps:200 total_reward:-267.36599642648724\n",
      "episode:386 steps:200 total_reward:-133.7106854922737\n",
      "episode:387 steps:200 total_reward:-1185.5374659957815\n",
      "episode:388 steps:200 total_reward:-528.0911346903323\n",
      "episode:389 steps:200 total_reward:-314.39589373523916\n",
      "episode:390 steps:200 total_reward:-272.39484511957136\n",
      "episode:391 steps:200 total_reward:-133.39569628136846\n",
      "episode:392 steps:200 total_reward:-453.15884707491557\n",
      "episode:393 steps:200 total_reward:-549.3733004120681\n",
      "episode:394 steps:200 total_reward:-986.2060433268272\n",
      "episode:395 steps:200 total_reward:-1166.4740020554686\n",
      "episode:396 steps:200 total_reward:-1156.117944114189\n",
      "episode:397 steps:200 total_reward:-1153.6845927896484\n",
      "episode:398 steps:200 total_reward:-2.610769774843758\n",
      "episode:399 steps:200 total_reward:-133.76121347893587\n",
      "episode:400 steps:200 total_reward:-130.56400359509215\n",
      "episode:401 steps:200 total_reward:-132.55752951679\n",
      "episode:402 steps:200 total_reward:-132.96305616687772\n",
      "episode:403 steps:200 total_reward:-0.3904519351320058\n",
      "episode:404 steps:200 total_reward:-128.85199068780636\n",
      "episode:405 steps:200 total_reward:-1502.1072182823445\n",
      "episode:406 steps:200 total_reward:-1006.247840122112\n",
      "episode:407 steps:200 total_reward:-1504.0518234212038\n",
      "episode:408 steps:200 total_reward:-137.5115887513639\n",
      "episode:409 steps:200 total_reward:-128.63164073967442\n",
      "episode:410 steps:200 total_reward:-1509.735997623727\n",
      "episode:411 steps:200 total_reward:-1312.8807955533748\n",
      "episode:412 steps:200 total_reward:-1121.864748405177\n",
      "episode:413 steps:200 total_reward:-1455.7980825876407\n",
      "episode:414 steps:200 total_reward:-132.60093503840136\n",
      "episode:415 steps:200 total_reward:-1209.041779072182\n",
      "episode:416 steps:200 total_reward:-0.8890274053557465\n",
      "episode:417 steps:200 total_reward:-1502.3407613426855\n",
      "episode:418 steps:200 total_reward:-134.0060634828879\n",
      "episode:419 steps:200 total_reward:-1409.1359258953319\n",
      "episode:420 steps:200 total_reward:-980.4733374231799\n",
      "episode:421 steps:200 total_reward:-132.25454232927711\n",
      "episode:422 steps:200 total_reward:-1529.345324331712\n",
      "episode:423 steps:200 total_reward:-1516.083914259917\n",
      "episode:424 steps:200 total_reward:-1496.4191956279283\n",
      "episode:425 steps:200 total_reward:-982.3658049895106\n",
      "episode:426 steps:200 total_reward:-554.0152766633776\n",
      "episode:427 steps:200 total_reward:-1231.1094488327121\n",
      "episode:428 steps:200 total_reward:-269.27067271792754\n",
      "episode:429 steps:200 total_reward:-270.059360149065\n",
      "episode:430 steps:200 total_reward:-3.614758948185645\n",
      "episode:431 steps:200 total_reward:-418.1806285451701\n",
      "episode:432 steps:200 total_reward:-404.3922847451704\n",
      "episode:433 steps:200 total_reward:-413.5689258958033\n",
      "episode:434 steps:200 total_reward:-140.25083365349164\n",
      "episode:435 steps:200 total_reward:-548.3471636843417\n",
      "episode:436 steps:200 total_reward:-140.6652528693597\n",
      "episode:437 steps:200 total_reward:-274.902379784097\n",
      "episode:438 steps:200 total_reward:-417.57877322203075\n",
      "episode:439 steps:200 total_reward:-416.6984802508629\n",
      "episode:440 steps:200 total_reward:-433.7979666985714\n",
      "episode:441 steps:200 total_reward:-395.00641940055687\n",
      "episode:442 steps:200 total_reward:-283.15778089806327\n",
      "episode:443 steps:200 total_reward:-7.955351688294166\n",
      "episode:444 steps:200 total_reward:-7.989249184827093\n",
      "episode:445 steps:200 total_reward:-937.4780485325961\n",
      "episode:446 steps:200 total_reward:-5.056010654714408\n",
      "episode:447 steps:200 total_reward:-1188.0819655760401\n",
      "episode:448 steps:200 total_reward:-4.09379261533382\n",
      "episode:449 steps:200 total_reward:-1222.5068606815769\n",
      "episode:450 steps:200 total_reward:-1.508623905806145\n",
      "episode:451 steps:200 total_reward:-1376.3389446282888\n",
      "episode:452 steps:200 total_reward:-1239.6832279208668\n",
      "episode:453 steps:200 total_reward:-1071.2481980768123\n",
      "episode:454 steps:200 total_reward:-133.8414584146928\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode:455 steps:200 total_reward:-131.8888163368479\n",
      "episode:456 steps:200 total_reward:-1.846242286327329\n",
      "episode:457 steps:200 total_reward:-1.2915321536647058\n",
      "episode:458 steps:200 total_reward:-1470.6146741663972\n",
      "episode:459 steps:200 total_reward:-1278.412504539712\n",
      "episode:460 steps:200 total_reward:-1643.0267629234627\n",
      "episode:461 steps:200 total_reward:-1609.4477981701225\n",
      "episode:462 steps:200 total_reward:-1552.6281484734338\n",
      "episode:463 steps:200 total_reward:-1643.1308233234042\n",
      "episode:464 steps:200 total_reward:-1576.4092951869238\n",
      "episode:465 steps:200 total_reward:-1578.9011462568114\n",
      "episode:466 steps:200 total_reward:-1325.2112779115062\n",
      "episode:467 steps:200 total_reward:-1427.7803525877769\n",
      "episode:468 steps:200 total_reward:-954.7959518635145\n",
      "episode:469 steps:200 total_reward:-1277.2059508434525\n",
      "episode:470 steps:200 total_reward:-809.2826513032632\n",
      "episode:471 steps:200 total_reward:-138.2369860927046\n",
      "episode:472 steps:200 total_reward:-943.7397792421099\n",
      "episode:473 steps:200 total_reward:-146.62275514645174\n",
      "episode:474 steps:200 total_reward:-20.670365640074305\n",
      "episode:475 steps:200 total_reward:-815.7207806013834\n",
      "episode:476 steps:200 total_reward:-679.5182808152172\n",
      "episode:477 steps:200 total_reward:-1063.5436855688652\n",
      "episode:478 steps:200 total_reward:-1149.6600404655014\n",
      "episode:479 steps:200 total_reward:-1246.6091549965986\n",
      "episode:480 steps:200 total_reward:-1237.4214842811903\n",
      "episode:481 steps:200 total_reward:-1260.1863412748671\n",
      "episode:482 steps:200 total_reward:-1283.2988677039623\n",
      "episode:483 steps:200 total_reward:-1302.080397171655\n",
      "episode:484 steps:200 total_reward:-1267.1657284612888\n",
      "episode:485 steps:200 total_reward:-1442.1033461735542\n",
      "episode:486 steps:200 total_reward:-1443.7660941672068\n",
      "episode:487 steps:200 total_reward:-147.13043036542587\n",
      "episode:488 steps:200 total_reward:-1368.1640826715375\n",
      "episode:489 steps:200 total_reward:-1376.1893261624634\n",
      "episode:490 steps:200 total_reward:-1496.5070222816307\n",
      "episode:491 steps:200 total_reward:-1488.1134132609134\n",
      "episode:492 steps:200 total_reward:-1502.4570938498953\n",
      "episode:493 steps:200 total_reward:-1522.2030064075605\n",
      "episode:494 steps:200 total_reward:-1399.130700601119\n",
      "episode:495 steps:200 total_reward:-4.133655657371589\n",
      "episode:496 steps:200 total_reward:-1359.0917604562983\n",
      "episode:497 steps:200 total_reward:-1387.2213906207246\n",
      "episode:498 steps:200 total_reward:-1499.1404221375599\n",
      "episode:499 steps:200 total_reward:-1518.8136544187566\n",
      "game over\n"
     ]
    }
   ],
   "source": [
    "step_result = []\n",
    "reward_result = []\n",
    "env = gym.make('Pendulum-v0')\n",
    "env.seed(1)     # reproducible, general Policy gradient has high variance\n",
    "env = env.unwrapped\n",
    "RL = DDPG(n_actions = 1, \n",
    "          n_states = 3,\n",
    "          gamma = 0.95,\n",
    "          memory_size = 10000,\n",
    "          batch_size = 32,\n",
    "          nueron_num = 10)\n",
    "step_record, reward_record = training(save_model = True, model_name='DDPG_Cartpole')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1126a1f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "reward_result = pd.DataFrame(reward_record, columns=['DDPG'])\n",
    "reward_result.plot()\n",
    "plt.xlabel('episode')\n",
    "plt.ylabel('reward')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "reward_result.to_csv('Pendulum_DDPG.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "reward_result = pd.concat([pd.read_csv('Pendulum_DDPG.csv'), pd.read_csv('Pendulum_DDPG_dueling.csv')])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsfXe8JUWV/7eq+977wuQAQxoGJCqC\nBMUEEhRUVEAxsurqKq5x18TimsaAIiZU0F1QUfyBiKKIgMCIIEoeQGBgCJOYxOSZl27s7vr9UaGr\nqqu6+743MrLc4wfnvns7VIc653y/JxRhjKEnPelJT3rSE7qjB9CTnvSkJz3555CeQehJT3rSk54A\n6BmEnvSkJz3piZCeQehJT3rSk54A6BmEnvSkJz3piZCeQehJT3rSk54A6BmEnvSkJz3piZCeQehJ\nT3rSk54A6BmEnvSkJz3piZBwRw+gG5k1axabN2/ejh5GT3rSk548o+Tee+/dxBibXbTdM8ogzJs3\nDwsXLtzRw+hJT3rSk2eUEEKeLLNdjzLqSU960pOeAOgZhJ70pCc96YmQnkHoSU960pOeAOgZhJ70\npCc96YmQnkHoSU960pOeANjBBoEQ8mpCyGOEkCWEkLN25Fh60pOe9OTZLjvMIBBCAgAXAHgNgOcC\neDsh5Lk7ajw96UlPevJslx2JEF4EYAljbBljrA3gcgAn78DxeOWvT2zEHUs3446lm7Fy/RbcuegJ\n4/dWFCNO/nFLkT411MCabY3M90s2jGLhii246v412DjSMn6L4gR/emQ9GGPYNNrCso2jAIBOnODX\nC1chEeP92xObcOldT+KuZZvHPb5rHlyLZifOfN+OEqzYNDbu49rCGMOTm/nxVm2p4++rtgEAHn74\nQWzabI5/69atWLZsCdbcfwNQcpnYp4YaWL5hCHc/ssz4/sHl6/DYmo3q76F6BxuGG0BzGKw59A99\n9hMRxtj4xsYYELW3/4AsiROGKxauQqMt5k/UAuIod58bH16HVpR913QZa0UYa/HjjDQ7uPHhdbnb\nr9pSx+PrR7obfEl5cvMYtoyl93LFpjHc8PA6tKPkH3K+icqOLEzbDcAq7e/VAI60NyKEnAHgDACY\nO3fu0zMyS2752Rfx3vB67EY2Y2GyH15MHwcOGlK//3z+O9HZ6SB8+GP/3f3B774IuO5TwGfXA5U+\n5yZf/cbXUEGE8772deP713/nBgRIMIoB7DGjH3898zj12w/+vATfu+kJXPyeF+IDv7iXK+dzTsI9\nS9fj3t+dh71nnoXD95qN7176O8xor8V1fXNw6Rc+0PXw71mxBR+57H6cfuRcnH3q843fvvubBWAP\n/RZnfOb7mDGp1vWxbbnq1nsR/umz2PyvP8XGn70LJwYLgflDeN6vj8LawecCn75Dbbv++8fhAMYV\n+6oNX8EeJ34se8CR9cD5RwCtYWCfV6L1+EPYi67HXgCiz25EWKkCAA7++f7YyKYAX+Kv64MXnI6j\nxm4AABAAAYCtH1uG6TNmOsd9xcJV2DjSwoeOeQ7OP/9beMFLTsBRRxyCRxdcjMrSG7HT0IMY/NDN\noJN3AksSLHpwIZ7/ghc5j3XJdX/BxiZFnABnnnaU+v7uK76JwTn74HlHn6q+O/sPD+HK2x/GfV9/\nGwghxTf4/kuB338IOPazwM1nA59aAkwqLG51ytUPrMWS9SP4xAn7e7f51T2r8N+/ewgrf/clxDsf\njP/a8nmsnXYYjtl0JtpRghfsMQ1Xffhlavs7lm7GGb+4Fx94xd74zGsO9B73eV+8AZQAy75+Ej71\n6wdww8PrccunjsG8WYPO7Y8692YAwIpzTvIe8y+Pb8S7f3o3bvjPo7H/nMnObR5ZO4xf37sKI80I\nVz+wVin9SbUQi750IgDg1B/ehq31Di7+1xfi2AN28p5vovLen92DJRtGceuZx3a13z99pTJj7EIA\nFwLAEUccsUNcsc9XLlWfj6CPZ34/I7wW2HItgHEYhFu/yf9tbAUquzg3uaD6ffHJNAi31z6G6WQU\n18RH4pptLwGQGoRt61figsp5GNp2AebjQhxVfQjASehfewfOqfwY9699FbDXq3E++zp2qW7G5ngq\ngO4NwrZ6BwCwbqiZ+e2Ux8/C/pVlWLnuo5ixz0FdH9uWufedg8ODO3HXQ7/jxgAASxIQALuOPWJs\nK40BACQbH3Mf8IkbuTEAgCV/wjwNLydRBxAGAQBmk2H1ed+x+zKHGn5qidcgnPmbBwEAHzhqHj66\n+atY84f/AY5YigNu+0+1zQM3XYpDTvk4bv/dBXjZQ5/DvZsvwuHHvyVzrHfd/Yb0j9NSp+RFj3wV\neASAZhAOuue/8Lm+2xEnb0EQBO57oMudP+T/3vot/u/W5VmDsHkpsOERYL9Xc4/+gV8CL3wfYBmc\n711+DXYi24ATzvSebmude86fqvwa2PJrAMCu2+5TilQiQCkbRvg7tnZb9l2zRQKj5QKhNhwIthu5\nfhFHGfc+udVrEC6/ZyUuvWMZYpj3erSVop6tYr64EPX2lD8/umFc++1Ig7AGwB7a37uL754ZksQA\nLTHJioSIY7DuX5DphNNArwvuwvH0fgBfUr+dtOFCvCi4G/euuwGHh39Od2rXAQBxnSuTGeCKbiZJ\nlUs3kgg6htKsB1oFv6akvX1oozDhCiShKZJqNuroL9iPtbN0GwAg9tMiSdQG4PYoKyQGLNckDIrZ\n1yiKEALYjWzK/NZpcSUXrOfGo7Pu0cLjFckpwe0AuNFEGYNAhTqgARAD6Dju24IvAI9eA5z4NWDD\nYuD+XwCz9gX2PsbY7Kbap8Unv0GYNfY4VvS9o3hcQqKY3/RKUALtCJmTrMep4dWI45d5tzmELMFk\n0gDgRwgk6eAU+jdQ+B2b5278I5b2fRXXxEfi1uRgMBBsZNOwytFCqKxn+/Jv/BkzBqu4+iMvL7nH\nxGRHGoR7AOxLCNkL3BC8DUD5t2NHS6cO1CYjipOJ3URpVJKJeQz9xFRu8oUjsCZPzD2UqMWNSQSK\nmv5bUOnqvEwaBMccbRKuuONmvatj+iRk/BrjIKWfGmNDhQYBkc8gdPz75BiLChw8d1TstbajCG5S\nECAsnzsHgDhOMB4XhJVVP/LZS8PguqaOeJZjm4Ch1eIE4+PD99z8t66278T8PBVaPvT52fo3sH+4\nBIs2fhjY/SXObX5f+4L45E90PGbDJTihejFuW78vgPc6t5k79hAA7qC9LrjL+vUMNVeA0mEtrN7a\nwOqtnvf3HyA7LKjMGIsAfATADQAWA7iCMfbwjhpP1yI87fpEoZ8yCMUKoUgSI4AoPltQnglFx1rc\naw+RoM3EGMbhyctTUgdH3aFccbeb2wchVBIeOGeBhhDq3LC1iD9GwTot9w85Sp/Fnn0ABA6DkLSL\njV4U+Z8xSUzjxGxDDqBeNwOfcSzevTzDBoEQykggKDL5LF0IQWqy9mj6e1hokreLKIMQlkcIVSYo\nmvbE5tf09lP83LH/XaZJ/nNoaYHk0kb6aZYdWofAGLuOMbYfY+w5jLGzd+RYupYOfzHqrQkaBEUZ\nTTzroNlJX3ovQhDZI4kwCAFijGCA/9bp3pNXlJHDILSFko6a2yeDoyIQAqMpJmvXOeXVhN8g0Njj\nvedSRv7JXWXZ/crQYnEXBsElzTHzPnbawmgVnJuVfbfkfZVK34UQpOPSHkvfl8rTZRD4uMIuEIKU\niXL2gXw+1I+gSZKfmdVop2MoixCebulVKo9XJEKYoOexvSgjAKg3NIXOMh+4SG+yPYYkThCSBMNs\nQH3XrSRuIAIA6AiuP9pelJGYcESLtzSF15yHELx0To5nzSIPQojaTsqItYqvsZ1jZIq8SyC9VjUU\niXwKDDkrq30kQpDvotMgiN9aIylCIE+PGpEIoRqWO59+3RM2CILSY2HVuw0tQPk6m/BPag96BmHc\n0pEGYTshhBIKoUiao2lgWFIOGW5fesWdMcRiciuEMA6DwHIQgjQIcWs7BZWFZ06hGQThNTfY9kUI\nzKe8O55rKYGu4o5fYUiDkEeGtCyDkHT4PqzouZU2CML7lYii47hv0hi3x1KDMW50W576ATTKqGRQ\nWadomp2JzS/KJELwRwyLjHpDcx7LGOnShnw7Ss8gFInvoYhJOGGDIBFCjnIqK616NlMoM3XEeUin\njlgovZEJIQR/UFkahDL8ehmRlBHRFFCrIQwC/J5b4IsH5N1z32+eaylUygDinKIrUhAHAIBOw6KM\nhEKOGqO5+5WOISjKSCIERwxBUUajqRHMMwg5yDdgxddsevn8PBka1CP63FT02jhFUkY051qLrmes\nS3pZXu/TKT2DUCS+F0AZBG2Sl514uki4XVChWUbamgcpJ5LtuEueM9ANwkRiCOKSXQghCji3XEZZ\nlhEZVNYNgowh1FmOQUi6NwiJT0H7rqWE0WvrFbbW86aKf/Z7hR1L8ccdM2PMJ0lZD15SRnkIwYgh\nCINhOU2GZ5tzj8O4OHtGP5RU8J2S80yfm+32xByuUCj7PKVPiiijLmMIYxOlo8chPYNQJL7J5KKM\nfOmNeUK3H2XUHhvOfmkrajFBaVRXHusoEfn27XzF4pJYGZ6sQZDpoWw7IQQXZRQ3OCoaS/IMQnfe\nPgAw2brB9nA9lBEp8ezbOm1hHaeMcuw0zecTCa83Kcji6pp6kIrNiRAcMQRrjhiebV76bhmDoH1u\nCMpN1iMUSaMdq/2jiSIEYQjygv9FCKFhJH0UX0NjouzDOKRnEIrEB3ldlNF4FJ+MIZSgDIpEz+ZR\nWUaWMpDURBjXVSZNahC6H39eHULC+JdkHMjDJRWWRQgQBqHOqt6+PRUfQsgxgCrt1H4uPoRQ4hqb\nbe1Y1r0OY/Nv3ehJSSwkIFOIixBCaYNgK7sihCDfMssgGKg5570OfbEd/XTa2BVCiMsihFjtP2HK\nSNaJ5F1PgUHoIYT/C1KAEPRAkTfgmCfbsQ4hNgyCUMaW5y49nDBuIBYvd51MPMvIRRkxYUzJeO6L\nQ2S6J4H2TETriQaq3kwSaUgykqPEiQwq2x6ux2jSEgah1dKO1bENgkm/uDJWYkvxx5GsKckahCjS\nlU9Jg2DTlq5rkg6SYUzN4xtOUh5CSIrvmZMyKokQxtqRej8jXy1KSVHef45BCAqKC7s1CBOOT45D\negahSLwxBP4yjxkIYTwGQQTytkNQOWk6YgjWNkScpxo3kIi8+DqZxH8ch+JOW1fk/DYeKi1HqIba\ngja/5hiBdwK56gYA5D4v6X1nFMAEKKOWThlZ565E/H2SzytI2sDWFeaYrNTWuCNrSrKKtd5Mx1O6\nDsF+B50GQdJJmndvHd/oG6Tdv2YnxvdvekL1KipHGaWas9ElQmi0Y1WsGU/QICjvP2eeFlJGukEo\ncc4J1ziNQ3oGoUh8k+nx64EkRr3loAGG1gDzpwJrsk3QMq6BCipPnDJKWqOc273t+yobQj/b1rG2\nCipXk6YKnDZpH6d3xkEZSQ8sThgie6IKxd2pD/PU03bd9EK3PukOpm99MqUrmsPAtZ8EVt2T/q55\nYkFHGkHmRQihg34BUM4g2DSK2KfFzPRDGtXxxh/ehuO/fQvw6LXA4msyx2wZaNJCCAm/XioyfHbd\ncjfwvUOAe3+ejskabxx1gCcWoLLmDtjS1OJJpEwL7MXXAEsWmN+53gcXhWobhLbbIFx46zJ8Z8Hj\n+MWdTwIYT1BZxhDKGYThZieNIQhjnCRMNbzrRsISMYSwK4RQ/EwmXOM0DukZBI88tHoI1z30lN8g\nrFmIj33zf3HBn7Xupz95JfDr9wBP8NbIuPdn5j6XnMz/08WmjFbeCfzu34EkwRX3rMKv7lmpNr3o\n1mU49Ye3YdEaM700Flz9Uxs24d4ffwRY8Hkc0uS9VHQO9tCvLMDWYT4ZaqyhWh8wWkEdNa+CXLOt\noXrKP7l5DPPOuhbX3X4f6kOb0hf7/l9g/rnf4J83PQHc9BXlmUbNUeAbewJf2wX4ykw8dOtV+Nnn\n3gx872DgKzPBvncIr6CePxX463f492fvjC3z5+KJc14K3PNj3lBNiqaUKh1OXVAwLNs0htf94K/4\n40NPZS/itu8DPzgC9as+zvePO8CIv08+HV0H3HcJ8G2tffOf5gM3fh4AUBddif4cvwBLk11AOw3c\nt3Ibpm26D7j8HcCvTudB104DR9MHADCs3abRLNa9nj36GHDNx1FNREuRJm+Al1x/lnefcNty4NLT\nMHnJ1XxbluJBvYiN6RRb1AIuPx3423eBv5zLr/Hy0/l4bXn8j8Bv3msabRe1mYkhmJTRnx5ZjxO+\n+xe1Nsa9T27BvLOu7bqCXR73qr+vxW1L+P1ZsmEE31nwuKFga6Jw7eO/ekAhrkQUGr7/koU49lu3\nYOtYG4wxfPKKB9R+5//5CSzZMIo3/eh2o0MpUBIhuHpcCWm0Y/z89hXq7yJzECdMjaFWshBve8g/\nffvrHSWvP5833lr++RdnaJflyc7Yi67H0NA2UOxm/vjwb4E9RC/70GpltuwW/u/QamDq7vyzHlSO\n2sBPed90nPg1nHkl73z5VnGYqx9Yi4fWDOFvSzbhoN2mqsO2UUE/2li1aRsGyRogAAJVzWv2NxoZ\nqwMEqLEmGpInpwEaqGGwPebM8D7p+3/FtnoHK845SbUkPvGG4xDcyJC86u8AgHMrFwEtAPgMcOlp\nwNYVmFbjvdhDJGnpP4Bd7v0mDqYp5UC2ruDIBgBu/pr6fgaGMAPC+GketZ5XL71MAoYf3bIEi9YM\n44OX3ocVdhe5BVyRD2x+AncMvhQvmTEGjPoNwrRbv5j98m/fVR/btA9gozjouLdj+c0XY56gfPal\nWsPes+cAAC6pAv8bnYQr7j4a75P1c2vvzx5/4U8xKZkDUKDRbAFExCa2LAMuexuO3LzU2Hz09osA\nAK0pe6E2vByUMO5QDMxSrUkAoHbX+cDdPwC+sAVYdTc3rrqBrU3x3gcsuhI4/gvA9Hn87yQCqpOB\ntqbMM5SRHlRu432X8FblU/oqCBDj7sfXAqCImmOFtWm6I82RAQNAcPqP78KKc07C2y+6CxtHWvi3\nl+2FqQO8sG6nKTWs2tLAKw/cCYOrA6DD+f1mJ8ZNoi10tOxW1EmEK++L8W3xrnzrxsfx+1vuwIr2\nVPzlsY046eBdgAVfBPY7EdNjvgATSTq4+Tc/wqblD+C6We/Fd9/6Akwb4BlueQhh6cZRrBtu4oA5\nk/HoupFCi/C8L16vsrX6Ktuhq3JJ6SGEAnEV9fw85ko7QAwKB4KQ/GroqZ5deWf6WY8h1LW2yA5I\nKblXO5uGgSBiFAEShMJLiYWh0Q9DwRCAK+Z+1lRpp4yEGGN9XoTw1tZvsaj2XuN4AeEfZm57KLuD\nOK7kVANi0gyExaja3pTi4D0zpaH1xtdjCCKDiIKptRkyw2Gm1lmzdi2weYmzKds1yUvd59ckAQEV\n/Xv6+/rQYYFqp+Gjp94Y/A2B/q789dtosbQvzlXPPQ8AUCPi+UDjvO/4IbDpMQQswkY2Bae2eJvz\nKe2N2MwmY9Gbbsa5nbfybX96InD+4Ug0T77v7h/wD1HT3WZi7ouBg95kfledBLz5Z/zzhsXAVR/i\nTguLgarVFtwyCEZaqEYZxYzh4sq5WEj+BSv63oG9iQPJWaLHEOY3vobHau/GQWQZAAYwhre3fo3p\nGAYZWw/8/ZcAgLdFf8C5c+/Cj9/9QuwxnY81YJGqXK6hjdlXvgmDv3krdsYWdfyPh7/GAvpRnB78\nSZycAbedB1z8GrVN2B7BsYvOwpvHfonoiZsQ//IdHNmuvlfNPZfsseAMvJQuwulHzs1cl0v01N2+\nil9NNzsx5p11LS68dal3m26khxAKxBWQa4qq2BAxiOvByj44PoOgw069DoGZ3rwtMljnSq+MQREi\nRiiUjlykIzEMQoIK6wAEGCCtNNeehmig5g2YfqbCJ1qSMPUix4wgIAyT6isB7GXuIK4plBlNlpKk\nSdtUeEDKVzvudx19GGhsVX8zTdlVkyZAOUIYbrgNQgtVDGjn60eLxygqfZlc+w4pbv/dQB+YyKoK\nwioYiHpPApeDAKDBqqbz0Knjr8mheGXAkUIifLOqMNh9+v1pphRhhBAd8WynkDpG0Y9GO0bLnsqu\nuEESuQ0CrQCVAfM7QtPvfvVO/n4e+k5+jL5p5rZ2YZr+h/auxwnD0YHpQPwmPhoECd4UuFth64d+\neXQXQIBrap/Dg8lewMqZ+AS9HAdVnsDAb34ErH8A2OeV+HD7p8AGAPiy2jdgHTWwPqRjmkTS538w\n4YsqzSDDvHzHosdiRjBnaxrL+kX1nHTNx8eudbdFFzJ1xfV4SzCMUfK2zHUVSZCz2p10gn7yt+U4\n4+jnlD+oR3oIoUgcCqrBUoPgVAAug+Cr3tQrlfVzqe2zedhugxCAaghBKhj9mDqCAICoIQKPNMQY\n+goriuMkSRthCgVEXcFwgXqqjE82G0UFcRuDxMpBl8bIcb+HMchXlBOSaJy2nNwEwDavQeBKvk25\nV19jTe4tOxBCB2UMQk3Vd9AKNwjyxijjZ/W8aaOSeVc2s5T2SwR3UoMDIWjPJWIBOuLeT0EdDVbD\nSLOjvpPibFcR+wxCwBGBLoSkXUwl3RdUOTrLIATLICRmDEF9dLy3Q2wQq1m6lGSHmfSIvkesqauD\n6XL163QyAjKyVuzgSyyIlDOjo1PdOAwQfs8D+TS0Boerw7kYwQBCXzZZu14YVK4gQiAKdnrN7Z6h\nwhxZFSlCSEBzEYJGZOtperoSNRCCPomFwtGOPyoymlwTKwJFCN69lO8nxq1NEAKGivbStseEkqUh\n6qxWmGUUx7Ga+9JLdbb8Fcqwxvg1h7ZBSFqGx84H4z/3CAaARgrt9XUF+khamyANpm2ApEFohXzp\nwz5lELIIrk38Fc9SGuhT3DdVCIHfmEDed8uLJmCZd2UL0qUYZUBYGgRJyfFBpXx9GyEice8pYWig\nipFmlDEIcFFXcdtjEMKsktcRgtouEDGEfMrIaPKmedmu97aDEG0tYyuxVJIeLLZ/k602uGcui+Tc\nqjZgkfpJ9+T193AAYtU6+f5oxowjs9DfBqUzlosQAL6CoPT2e+2vn6Hipoy4IgkRmUVSUlQMQTMI\nusLToageVHYgBJ1ukbxi7HibYgQIEKvtVVBZM2gUDBWSnrszJnh5GvCsmYLWFbHWAVQqIGeHR2EQ\n+pj0uEzlFCStLGWkIQBbtrFJpRCCFHtiSoXTCgaQMIIaawmDkF2/LCpBGfEMI0EZVWoCIfBno4xf\n31RjH24QbISQBnMlBVUjDqWiIwQEaGvKv4EahksjhLa7T3lQKWcQWFzKIBiN+jSlmrjiYgiUgQO4\nY6OLbkOye/NrqUJb0tSTFhqwSG2izwGJCoDUOISI+W3S4x+E33dfGxTWGkWFxIo9cEmISN3+bhbI\nKbPl9jIwPYNQIK6J1RQPPSAFCEGnDXR+3kAIYfqdAyHYyhQoiiHwl11RQ3oAFomhLBMZqKUhTzst\nqLaNIi2vOxch8N/6BEKQXq+UajTG1yXWZWyj97zbmEkZ6V1D+yH7G6X3xI5ZtASii1FBA1VBGbV4\nDMGSBEGGtrCloymwIKzwM9sG3DIIAZIMZbSF6Qgh54SWQYg0j7rOahhpRoaXDWRblvCTRBnlDUAg\nBJsyCrIL38QdD2VkHpNp77eKU8GDbFloGLPYVkmGQbCMmVD+FaTKHpH1Pgrnq8I6Cm3oc0B3THTK\nCCCA1iU3RoCIBQg9CIGJuZRFaqkYlNF2UuA54YVxSc8gFAhzcJJNQUGEiD0GQdJDOuzXEYLDi7G/\nUxREdgLbzb0Y+ETiMQQR3JQLeiQ6ZZSghghNkd3CRB8gGoRosBpIQQwhiWM1qRRt4crLlghBQPBJ\nVryAurjWsezC81KGmOmpJnF6TTIrRw/u2whBUkYR4YavmjR5jYADIYCQ3EmtNpMfgioIocrj4xlV\nJONdUzAExDIISBFCnJd/qT2XjuVRN1BzUkbO6uS47a4joAFQzQkqS4maAFjWeNgIQXMSmIEQsqfu\naEFywEEZac81YxDi1CCkS3vq6bBMzasAaaO7qkEZNTOfAyQZhCBjN95Gicog+J2JColVi5ftxRht\nb+qpZxCKxIEQGooy8gWVxUumxx90ZasX+sjJ5EUI2eO7oHeMgMcQhIeqlCKzKCNEGILw8FoyqBxg\nDH2FCCGOOuoFlJ6cc5UoYRAGBEIYRInWFTkIYZSZitvVmlpfM6WSQQjCIIAbviBqcITgygIjNHdS\nA9ZkphVO9yjKKHYuokJIAWXE8gxCSuV1EGYoI1dQmbhWO4s7HoPgiyFYCEF2N7W3tdSbXs1bhBDa\nCIyxR9a911/1TAxBGBudAjKox7itlLo3hqBRRoOQMa843V9IR4zTmWYOAM2tYju/M1FFBCq7QHah\nycuAgO2FFHoGoUjGk3YqXyR9X50y0tGA3MaG87qCsSQSRkqvTI0RICBxdnsry6iCiHPyAIjwpohI\nOyWdeu6aDkncyfSi596gdQ+kQRCTbZB4AnG65BiEhrVeso4QpIQ0ByEIRNRBiDr6EMR1b5YRCDUU\nrlN4bRSXoMIbCOqILsjGIWgRZVQSIUQsNBECcweVqSse5DUIFQdl5EAIag1lO7ZgvxNam+cCg9CB\neT02ZZSbiC3OU9WDynq9SntMzcWQdRTa8FFGMiFDZRkZQeWC90KkBndYPmWk7IH/SBl5OuPPPYNQ\nIK4Ygp52mluYZiAEzfuOHQbBE1R2HV+2ctEDcInMMrINgjZ+GVSWCCEQCIEEFYzJJShzUEIUddCJ\nzPHQpJ3xyFXmVDeSYxDq9vKYDsot1Fyk0IpPtAVCiEmAJqoCIbizjDhCyDcIxgQNqoCddkrDjMtG\nrSwjRkOeTisvKW/WOzxVKXX0YaTVySgr2na0hYjb7l5ENMzSZ4QCQZgumgNoCKGAMtLGy7Rn5UK2\nEUIjZpMxCNo+zIcQdMpIRwiduhshaO9HJtsN/P0hhBjxiA4LMuhFFyLOm2c0wnFmGbnumy29oPLT\nJK4YQqts2qkPITgNQttZmGanbAK8HgBIi88AbhwCJBllmEk7RYRh8AldiQRCCALVmyfPILA4Uq2H\npaGiSSdboZmz7qxXcmIIamznkkseAAAgAElEQVRCSOSYxHkIQdYhMJ5eG8TSIGRjCIzQXC9PjSHN\nOwUhRHmfIWIwh0EMkBjGPa5ONpRHLkLQJEJgUFoq7dQKhNO2Y7GkpOMxCEEW1cj2tTptpAxCPkJ4\nxS1a1bPuZTsRgmngsjEE92d+bEdQWTcI7TG1TYjUIOgxhH4Heg2clFG+oyCZgqKgMlEGobwGz3MW\nekHlp1mcWUbCIASIef8YW9Ti4x6E4KSM3AjBbvsApAjBDMaZaafpcfQYQoIqIjQCTlVUI9EYLqik\nXrhONSSJQVfEcaQoI/kekqSTRQglhek58XkIAcUIoUwMoc0C1FHjC7PISmVbCHHTgPbYte1BqFq0\nxxtDADMoo7gy2VAemYCpRyIEhqfcYDUMN7IxhMCLEByUUVDJGgT5bCpavEAahMBKrdSVG2NG3yqd\nMkpcMQSLAouZjRCyw0031oLKUpoWZSSr5VlUSBlJcVJGLMhkcrkkL/5UHWdhWtcr3k1AegahSBwx\nBOnZVYiPMhIvme6NSc+bBBZCkFFaM4aQ1wohThIkCTMRAgmclBGx0k5r6KAe8GBmTRgEQsPUC9cN\n1y9OBr62q/oziTroRHy8UmkGSdtBU+UU6Mx5fvpZVzZa4ZktNmVEHJlNgbZkmy+G0GK8iV8lkUHl\nrEEgoG4jrwuzAn1EL0xLwEhWKdgxhDgcMLxhu99SdmB8W5uSqIsso6gUZRT5g8o1M002NQgOhGAj\nIH2OWOhNT0F1UR92kNw2xnn5+olc24PEbspIiyEEKC5Mk5JmGemUUTGVKK/HJxWixRC6ooz8v/Wy\njJ5mcVUqAwSMhsXN7fSJIj3vvqmlEIKkZtxBZYaYsUwMIUCcyfk3gsokQR/pIKpMQotV0C9aLZOg\nknrhEhH89gxg+a3GoeI4UouTyInLEYKlZHwtgv/1OuC4L6R/zznIOU5bxmAGf13VoqGmxKvWPWgT\niRAo6qyPL8ziiyFQiruSA7xjAYBhmJQJ0QxCiBiRRd9sxDQRQ9AMQsCv6Zudt+D1ra/6cldSEZXP\nNo8t005tyqk28mT2GN600xCYtQ/w5p+nTe6UQdCuVXdqdDFakloZZQWtKyIEhjGzjfGLzr4Jdy7j\nnUbtudZuO94zI6g8miZn6IVpBZRRiBiThp8A6qmT0mFBKYOwls3y/lbRsoy2F0LopsCtjPQMQoEw\n31SlobcOQaZFNvTFc6TnVJ3kSTttG0oxbvE1a22EMDVsgyUx4oRpHiYRlcpJFgJrBu3W2sf5V2E/\nmqiin4lVukQdAgCgvhn4w38AD/4qc10f+sXduPoB3jNGXvdgZyumkBRVLN04ipZn/drOwGzD40ze\n+BN8a9KnndvqMmzVIVBHLrhOGZ1d+anxW0I5xTHcpmigyg1C3PJkGRF8p/NmtAZ5W3NGK/hR9Hpj\nk090Pogb9v0CsPexwIy9ufJUFF+CtSMRNu0jFOvHH8EC8jJQK94UC3RyQXwKHmJ7GxljThGFbjYS\naLAa2nGSoZymP3l99hhXvJMv3mOL9PifdwrQP0PcB2kQNBSlEIKlGI1YmZnuyrSVytxpp6ERSNYR\nwmHkccwjT+Gyu/iaIPZcY2v/rv3hyDLSPvMYAt+mqqWpDiKbnltBhMMXvBm48wL1XT0iuUFlKb+K\nj8H8zrucv1WMoPL2iSH0EMLTLZ7VmZgwCC6+eesof8lWbtZge9zhEymouBFCrKXOAej/8cvxoeD3\nhkH4XuV83Be+F4dsuwnf+MaXMB38+NezlyAGRQWRSp1Lj59FGB1aE9vzcdBAm5QP/DJd2MdKLwyR\nYOUWqfz5WPcevRcLameqbY7/9l+w9Cl3G4qjv3Mnbl2e3hM6ZRf8pvUi57Zq+CBp3QR4qm3giCHE\nmuHbnZjxCEY5Qhjt8LYj/YlAa2ENOPkCjNR21ramWIPZeOC0vwGfWY11H3oMf0kOUb/WD3kPNrBp\nWDflYOBdV/HnSah6jhVEiEGxZs9TgS9uA6buBkaIaD2ePpuImnTVdYs25N4HaUjt4HFDxLPKBqXx\n8G+z31EtfiANgUQBRj+uAoPQ2Abc8N/mb2KfnbEFx7M7YUsHIYZZ+nx1pf/b2nzcUvtkBpVK6X/w\nEm0MQskPrUq/u+rf1cdQZBnNJevxufD/qe+PDdIFcgBe8zJImgijOjCavkd2yxCfNFkVV8Uvc/7G\n0067jwLnZRmVyUDqRnoGoUh8VEYOQpC8aazny8dtHowLKiWzjIA3BHcYrStODm5HgAQz68vwxeh7\nqJIYt8bPx5fwPsQIVNVu0fg7tA8JCVTwNQi0wJ6eUjhlV2M/HbLnvda+Jl9NVHDOguXpMSjFbjMm\nZdNKLdEVRgNVd/8krTtsP1p4ONkz/UkovDZCxLryC/uAQ/8FPzv89+m2QiEmjAG1yYiDfkRMp+b4\nlWshC4MyCpAgQoBKQFUKCCEUFAwzB7WYT2AahMJpLegtSVucH52MK+OjcI+gt8oGpZ2iK3iJFqRh\nOOErwAln889FMYSbvqwMzopkZ2xkU1VM6sPh7/H94LuwpcNCPMzm4WXN7+Gy6FhnzOyPi9Zh/tUP\nO9u4SKlEgurc9iTWYDaWTT7C/B28vfyttY9jGvFX5DdQxWRZSKklVPT31dBfK2582KKDRmzPHEOk\nkre60eN52/YQwtMsrrRTAAAJvJXKUiHG9tKDVGRzJA7KqNNwKm9X2qkeuNuMKYgQIiE0w50DZpGQ\n2p32I9F4YI4QxN8aRcCmmKvB6fGMvEwc30IhbVRUhpaUH/3LYSBV7v0yVydOmJx9AzUnZbTfzpNw\n6fuORA0dBIRhVIs7SCNQqVRQrWk0kbhW/QlLD056XoyZ6b2SptM9Pb0wLUSMGAGq2rKHMSgqlOGL\nJ6WxiZhaxXZFU1HQW9Jwfyt6Kz7Z+aCqZSiNEFyiGwSFEMTxdj0UeMmHARC+UI69PaC9wyl1+OXo\nnZzqE98NEne1uryeNZiNBNTdLBLAz25fUSr7CwAexTxcsu/3gXf8GnjuKdgUzkGVtVWtQJ60UMVk\nSYFqGXcJCZzJArZE1cFMgz4pIUm0tc67oYx6WUb/NOLsGgnuddq55VKk5623aUbc4YU+tGIGXdVk\nGnMaBGdhmp65IRRBjCC7ChncfYOioA8MpkFQL7HGASeTTYQQlDQImcZ1QjoIVcaP9Lp3mtwHFnKF\nn1TdSzmOaXUIvo6TfQHBS58zU8VQxrR2F0woMEoIEl2ZCTrEyPARLpxK/kqYwR3L+62rX6OXERJE\noAaCSBjhik6jtTrENAj63Wy7muuF3JAWtdUYlwQug6CpBkIAMGD9Iv53XpaRkCaqomEi97J9qclm\nLQbVEHf2/XLW/DhkKduDD3m/E4C3/Bwr+/bHvORJzLjpPwv3bbMQk2HSooCM3RSrS1KbbMR5PtP5\nN+N36VT1EMIzVTyUEaMBKsQdQ1DLWOoGIemkCMFFGbXHnE/XmWWkH1czCJkMI3gQQtCHRJvUNKyk\nMQTNy2sPzDHHosUn8ianjzLSG7Otwex0PML7TewK2PQq1KeYBU4kBDAQQlT3U92IJCLLiBsEDaEI\nGsZ4wjplJP6NXAhB0/hBQJXjIBGCHghMQHj3UQ1tRsT0snWE0LTrLgBlvHyBzUKEkSd5lFHR9gBS\nxZnekxaroI4+EEEZ+d8JszhPvlfZdzm7noRPVmLntHAQfP3rKiIMLF9QuG8LFUxy9N7iCKH4Hgd9\nUwyEYD8vuYRtN3o8DyH0soyeZnGnnQKgFZF2mn0gVeEhR0YMocNjCNSmjMT+bTdCsDtkArxiWH0W\nL37iiSG4+PaY9hnw16CMtKZojf6djf1KIwTP5GegWI/p+EF0Ct4bnZWORyGEyc79jLHDTY3Jezcg\nOquOspQaooHozEqBRKdqpCHSL4WYCME2CLLRrB4brARBmt6IWHjxWkYRE1SInkVmdyfVPtu0GgBV\nDOaLFUxILRiUkbjWvPYjPspIuylNVHkH3U55g8BAFCIesLJ/yhoDABhh/QZC61BHAaImN8RpvKGF\nirk4kZAE5Sijvr5+o3DQLrQLxNzf3ghhe5mFnkEoEi9CCFHx1SEIMWIIkjIKQgshCCXrMwiO4+uK\nWaYrJqCoIUuluAxCFJoGIQgqGmWUekdNYuX/lw4q51UuE3w7egtWYRf1jczJjytlDYJDuYiZoSgj\nDSFI5RAQYgWVBULQLYKFEOLE7K8j6SU9hlAJqbofIYkRsyxCoGAWZWQbBE2ZMsciPUJB+w3CPyio\n7Nw+AD5wK69bADSNZRoEfY0N5zODXW2fUkZ2wVjePNvITKpxhPUbBrtN8pMWNrB0dbuWyxgDiEkI\nRovV5aQ+87navZkCJhHC9okhbO/oQs8gFIgvhgDCW0X4FlUHzJW9FGVEPWmn7VFniqiLMtKpG6kI\nYuKLIWQNQhL0G0qABiESZhmE/3gAzcSGu9n23C4pWkoQMD3sRFFGxQYh8RkES5HoxWyByvYhYAZl\nxI2GPuGI1XwsYcxQWpEyCOlhKkEIQswYgn5MpRQ0ZBhZazczS5lmRDwvHzLTg8qs215SrqBy0fa7\nHALsdbQ4YXYOtFBNO+jC7yTofaPkmh5AujSqlLx5tp7NMP4eQb96jkAxQhjRkhZ87SkSEmQL8hxS\nC837Zxtque7y9mput73bWvQMQqH4g8ohklzqxKSMorRnjKswjSWGdw5wFWen2iUgRhaPPDsPKrso\no6zyjIM+NfEjRlEJaUqLyDFUBtBMzNdDN07jiSHoohsESRnFZSmjnPTafqFI9DUUZFuLgBIkes+e\nijQI+rhshMAQM5dBsBGCmWWk+xFKWeutEDIxhPR4drtvPrD8qWoonhKKyxAXZZSnaOQ2aj1I2X43\npXmarIo6q4GKOoTQtTQo/DEEGyHkzbN1bLrx9yjrN9RwoUHQCh/bcKAzCIRQwljadQb2PHE5aEWS\nW5jW9dHypWcQCsSfZRSA5+bk9Vqx6xAqfPK5EAIAtMwe9gRZzyhGaChchRAQqNiFLi7KKAlThBAj\nQBhQLYYgG5hV0IjMl9ukjNzXTZG4OVirElefOHEoKSNfUBm47phr8b7K1/0IIUMZaTEEmnr1LNBj\nCFmEkFJG6WFdlJEulTDALAzh/cE1vHWFjRDkPhpV2OkWIRQZBCNTagIIoQQtoraXY9ITI4Q0wduh\nkKjbGIJpEGQ2Wt4822AZhO4RQvqutHwGAbQcerJeD4l4YoFM1UqGXXr2vu3l99ur6ek4+hQ/u4R5\nGrXxXkbtfIOgB6SLsoyAdAUzTTIGgYQG/FZBZc/L6vJIYo0yikARUpJNOw2qqMemp2nWIbjFN/Ht\nPHn9r6QEQtjWPxcPoIkI1JlNpYLKIhg5phW7ScqIEmJW5QqDYMw1FVQWCIHpLULcCKEaBqiRCJ+t\nXAYAWB2nGVQA0uUxDcrIvLf6/WnlxBB8YiCEbtejCByVynmSMQjiBmot3iVlRKMmCBJUfZSRldIr\n6xBkjyGpoPMoo3U2ZWTFEIoRQrFBSEhYCnnZCEH+ldAagqStGQT/MVzKnzF3q+te2unTLIkPIQjF\nTB1ZQGpfJ2VUzTEI2Q6VdgwhJqHxnZ5l5BJXmwcSVEGE0ohBeTqmnXZKK2jENvwtRghlDYI+cZKw\nOKjciRPVvylDjQU1SPAsKSMTIchiMiAJCmII1EQICWMmQkjk+NPDVELTr7IRgorPaAbBVjx6Zoob\nIeQro2Q8lJFU7LoBUfvmaBo78OxACB0EqgK9H+2SaadUKX6J9OR98hWsAcBGYhqEMVYz3rY2zQ8q\n6zSRby2MsoVp+nkfTvZU+iEW751E7Hl63NXzyRdH8FdtjE92CEIghHwTwOsBtAEsBfAexti2/L12\nkHgNAl+yMi/7IUmsoHLY56aMghpvtmYZBIZsdkXWIIjDE9PTksjFhRCCgBqUEaWaxyp54KCCsaj7\nGIIrCM7HafkejhhC5DIIwnBIgxCDZvs1hX0ZykivVJYxBEqJkzLSJyCxEEKSmF1lFYVBTYSgi6sO\ngf+gUUZW8Zk+38dFGVlB5XIUgtjKlWWU53qq7a0Ygt46HUS1VB9Ay2sQ7BoPSTdKykjeizwkvomY\nlBED6Qoh6MbUzgpS35Ow8BlECJWjcEjzQjRRxSnBbXx/YZRUllHO7XXFDHxxhP8rCGEBgIMYYwcD\neBzAZ3bQOArFhxCSnF5GUlipXkYMqAlF6FgH125dEZPQ6NYoPXv9RdbT7FxBZUoISJBSRgREK0xr\ncFqFENRj8/UYIC38/L2yGR3DH+MXZo7tm/j2RNMRAsujjMQbHyVMGYSMTN5Zo4yylcpUo4xIQVBZ\nahL5ndlVFohUHYLfIJSLIeQUprGJGYTSlFG/SLc0mtuV2Fdmw+UgBACqg24/aaLiCSrrnkEaZ2Ip\nZcSKKaMt1Gw5nTBmvF9FBoGVMAgM1LkSnpSv9n8a8+ddot6LIUxCC1WFpJVBSIrTTl1owI8Q8i1C\nt7GKHWIQGGM3MqZ6KtwJYPcdMY4ywnx1CMJT322qH46yRCzcsepuYNsqEVR2pJ1Kg3DX/2SOYa+Y\nFpOKEyHocFavHwgshLCazUJAASpe7gghCNE9NaY45bpFGX21cjGO7nCPh4BhOdsF18Zmt9LD6BOO\nO+GIIegIQXbydCCE+HXf479FSUY5400/Af7jAWDGc5BSRi20WGgo3EDsQomFEMRn5qCMoiTBSLOD\nxAoqd5Js2mmtYiEEFhiem7p20U75xvAYLJr6CmOfwsK0LmIIhdRG/wxg/lC68lm3aacyE802CB3T\nIMg1NjhCKF5VTzUOFA0KgbQuIM/xepLujoUzT1Z/M2ZSN3Zn2Svjo7B19+PU3/rz9VWCE0IUenTJ\n7fGBaA3unuH55bgTYXRpqRhCue+KjlPmd1v+GWII7wXwx6f1jEkCrFtUalNfUJkjhAQv32eG83e+\nbwLc82PgJ68C6pug2l/baac1P3due0aMBFbaaZplJKVJNf7c6mX08tb3TYTAAhBCTM9IGIRmnJ0c\n5Ale/k8AkXRrvkL/Hl6DpckuuCU+xPg+L4YwMusw3B4/F61Jpl9w+2HfRnD4u0AI0IxikQKqnW/X\nQ4Hp84z1CPrRQgM143xPTDsKAHDf9BNBQk0JintgZhnxa55/9cN4/vwb0YxiJ0Iwm9uZ9yACNYxM\nZGUmndP38QwK0M/R6CKGMLnGr6Gvqv1ehBCkAZBj1LdXWUaWJvn0UuDgt/LPYm0GlZGVJLj3ya0K\nIaye8WLjOuwYwiN0P+ewVFsQJArppTGEHM0WVPHb3a11NbTnoxuEITaAT3Y+iMeO+zHWH3U2zove\naK5cl9MrKg8hbB6LMGNS1ZF2amYZEVenXktcaMCHBAoNQuHZTPmHGQRCyJ8IIYsc/52sbfNZABGA\nS3OOcwYhZCEhZOHGjf51d7uSv34b+J+XAU89ULytFyGItNOcJzKNjACbl6Rf+NJOp831HsM2CHrb\nakAvTNMUCknzql1BZRDevwjgzcVE2VaaZx1U8adH1uORDY4OlUKJUvBFWWyIPQVjWMz2zKwqZvfa\n0adNY8YBeEfnc4istFMa8AXs9541iEfWDqOTJOb55IpnWrfRCmJ0EOKg3dLq05GBuZjXvAzrB/ZP\nvWJtcRx9yQup3LfW+X1rdRILIdDM+G23MEJoxRDMa08YQ2RRkYUIgcjzmu/b/nMmY8U5J+GFe6W0\nCSPZ0OBPotfggugNxrGUlKlDGJwFnPxD4P03AzsdaBzn7uWb8aYf3Q4Wt4GjPoXrD/0RgDRYW0Fk\nGISH6AF4feur+K/O+41TyHc5QNJVYVpICeLYHK/RukK0Gu8MzsHhrf9Rl7ftee/GedFpBrrydSql\nBNhW96OcVsyw96zBTOzmFrGWxpK5b+HHKRFUdlNG7m2fMZQRY+yVjLGDHP/9HgAIIf8K4HUATmc5\no2aMXcgYO4IxdsTs2bN9m3UnK+/g/46sL9zUW4dAQm+3Uykvp4tMGsi3HkJYA17zTecxMmvMksBZ\nh5BoFEkjhzICuHc7bTBtlqbqi4QiYbSC912yEA+tq2f2BQ1x/AE7gYC/1LZB6CNt1Fkt830eZaRa\nTlu8usyEOnTudPxtyaZMTYBS6rIbJ7ihSkDwyRPTVtMVsZxaJaSpEdGWzzQpI3OcdpaRCyHYSbix\nnWUEUxIGtCPruWrnaDliCKv3OAkJCH4bH2V8r4ahjWfjWBbV1lFL15WQBXJyn7Jpp0EI7HZY5uSb\nR5sYRAOEJQBN6TKZsVMhpkGIGcFDbG/8Kj7WOHzaSTZdTIha/7okDAg6SQKcuRzxmSvEMbIUGqMV\n1YmUif8BVp1JDkLY2vQbBAaCeTMHMwhhNdsJ85qXYeP0F/DrkA5a3qI3jkv1xhCeKQghTwghrwZw\nJoA3MMYcWucfLHJChO6+Jbr4g8oBppAxHPvoF8qfl4r212ApbcQSPgk9BUGZZQPFOgzqb0cdgmkQ\nXEFlIBQIYdcZk9PJI8Yg+Xd7bWC+TQUXnH4YKGFgIBnvdwAtNFDNGIAMQtApF/FvbCkjyee/YI9p\nao1pwyCotRuIQnIUPM5AtZbOFRFEqFACogxCSiPERusKi/5JmKGsZQzBtAc2QgisoPLEEcLqYDe8\nZ+6NWMrMNSrS+5ieY6iVVQMxaJp5FVnLRrqyjMqoEnHulwz9EQ/3iTbPJDWG8j2qooMa0WswPJk8\nijJi6h2X7z9xFDsOsQHccOLNHCEkDBiYAVabqg8NANAKuCEc20lbNIelirdMlhEhBG950Tznb/xw\nBHvNHnTWCgBpd1uSdDig9R7JQxl57OH/lRjC+QAmA1hACPk7ISQbTf1HivTQg2KDYC+Qs/BlFwLg\nhSqzyTBqkcgMmlIiLh5UUsWvt6wg1MsRuxGC1tzOUYdg9PFxGATCOSMAwOTBAQWvJUKQPYyci4rT\nAH2iXwsDyVQgD6KBBvowuc/PkfMxaJ9l/yC7VbDYaM+ZKf2U+BACSxFCDKqC5gD3IAEgoBREPnNt\nIaBITzu1eOLIWkJVbmt4gpYRiREYM96ubmYMiCyKwyhMcxRHJSxwKpEUIOQrtYgFaQFWx/LBnHUI\nJYVQzIh0pJ2+sfL9sdtQuKq9AT2onKiWLRIZuCijJ9nOSCbtgjCg6n7Kc+tnqFem41+Cb2D1Ueeo\n7xhSusWMIfhV4m7TB72/MRDMnlQznoMusWjBTpIOR9e5aafbMcuoS4ywo7KM9mGM7cEYe4H479+L\n99qOErsXgXeKpiz+MPv92Lwrzw556ClrGb63X4bzZn85PYVrYtEK0ldVHFcZhJIIgQYIiQshpOer\na50+3ZQR0olPK6lXIxSD7GHkXIwlqKi3mbFsDKFKYtRZLY1HiOyKvKCyNEi20ZAVxtMHUuMizxdT\nzbgSCkUZEQbGCK+1EFLRPpOKyCzSKCOz26nl7VvkrcoyMoZqIwRqcL6xjZYYU+sESzF57Ox9j0Gc\nfLC8j6zAIMQIjNoM8yAuhFBS7PeW0LSjuzAIg8REJBFzv+tMQwgpZcSMf83tCSqBqLQX7r6Kk2vU\nHwGwGM9Boi1byli6baKNx4mKIe5zDp32s387UmQiuX+XDhtNOrzJYm7aafY739b/VxDCjhWJEOJs\nu2hb9CwjIqp6AWDNkKVowz6MhmmBjL1EIoB0QXYgfVKMCcrIhxDsLCNqxRDE+UgXCIGkCIHzx1Kp\nyNoE819DaKjO6qKMAJ5dkgao3QbBcLCJDIyb55NKd+akrEEw0kcNyojHEHSEIGMIABAGISJGwQI3\nQqAWdWd78m1H6wpbC8QaZcQYMxQO/y5raPT741oO0zYq6XjFEAp48AjUWCPCEGN1NOv9LJKMQUiV\nnTQI/TZC8KgddwyB4c2H7+6MITAA1ZAiDIh6Ti5PWlI0zNhXj/HoxjRvHQj/b4fP5dmG1GMQOl0g\nBJfhL6pU3l7yLDcIxV059aAyI6kvmPHiCEVHMwIxcS1yUkmVh6SMkph/VxYhEHdzO53nNoqyHBOJ\nEKQvNw01ykh430IxOykjLcXTFVQGRLfODELwF6YphEDc25gIgY9b9/b0MVHwIHAQBJnjAEAYUHQQ\ngmkxhMRRqSzF5vojWZOVE1TuMN0gZBW8CyHYx3h16xx8rP1h9XfMqFOJyPgPo7pScyOEESvzy3BK\n1AE9aac+yUEIMqg8aBkEH0KQ9ylAosUQEhx/4E6eOgSCakjRFwZodExq17DXAl3pitZACCWyjIiO\nqr0bmMFsXeScInGZGILrO19QOX3P3Md6BlBGO1wkZVQiJ1gvTCNIXzSXQRgLpqo/IxdCoCH8lJEH\nqtqvjjeonO4/ak98+5iEpAYhqKYcvmx4pwyCY0xxR42d51hlJ0Cd1VIDpXL9Pa4TchCC+L5PK/xS\n/eormrdLTITAQNQqabZUAoI2QgNhRHkGwUIIHbk4mHkBxjYxUr7fZTSTghhCwigeZXOxBenCL3Hi\n5oPTUxfEEECNrp5eKdO6whhAFiFIUTGEDGXkiyFolBFJnyclxIMQuEGY3BdipGkWfOmK2Y0Q0vup\n03W+pUgJkE+nifvgQwhy/QuStIWB8h8qdvzoLUzzHyZ3P588Sw1Cx/w3R2yDIJVUBloSikaYTuDE\n1YI4iTVIrgeVAw9CIFmEQMsYhPxmXpTAoIxSnSIa3inF7Hi747YaO4N7ApVCCDpLoWIINq2UPf8w\neGCPVPTqUwIz7dRECLqElKCNChLNIJjN7aygsoUQOqUoo5TvTxgzFA7AvTobIej3R96H2PrOiRDk\nubX3x8WDxwiylNHrvgNM2xOYpC2V2m1Q2X5HCFWIS8UQRAda2b+pyCAQJIoyCpAgoNl5AAjKKKCY\n3FfBSNNsCaErZknR6PePIwb+ObdSee9jAADLagfkxhDkffAFlRMSIGGEr3FO8oO9iQMiPF1pp8/O\n9tddxBD0pGAOG/nnjPdMqLler/NYHY0yshCCJ+00G0MIjBQ++cANg5Dk924xYgg01OoQZDsLdwtg\nAPzeqbcwG1QGuEEojFiZc38AACAASURBVCFofyvlWuLtHRK59NTOnbcoI+rx5sKAoo3QMAiGt25N\n+o4dQ4jlmM2r0UVfIIdTRtm004416fW/5L0yPFdNgemSjqOIMqLZoPL+r+H/GZfSpY9or/JHUtXd\nsQxCG7xDcBFlxIPKgjIiDJSWRwjyttoxKm6gNcpI+8uVZZSQkFf5H/gGnDr0CUyp9gNkuf8+qFX5\n3D8zcJ0hYwh577rrOfsXySnIMupRRiVEUUZlYgj6C5+qscykI9TwGp2rK8VRlqPNpYxY1jOyFJ18\nmfW00xFWZBCQjiNIy+1TyijHS4zbauyuSmXAooxke4NMDCH72faCXO/ykEAINNGMuZF2yj3yMIcy\n+mV0HMb2OSm9JEcvIymZoHLi8AStZ52AqGtJGHPEELLprPr9UahP86TjhDm9SuL45DMIziQBW7qp\nQwCQSZDXYwhWlpH82xcg1w2CbOpIwBAQH0IgGkIwF56xuXwG631i6bZmDEHEqGTDv6CCmAQCVZeh\njNzXxhi//nHXIXgUu8sAGvvlnMclz1KDUJ4y0l94QtIHnumbTqjxUJwNxpIO1MTN1CFkHwUlxFmH\nYPwtD619P5IUUUbEoIzUkBRCyFEcGkLgXlb2TWyglvk+jw4iSgGaxxmZtr/6/KU3PA/vesmeCiGQ\nSAtU6mmnIssoCNzXEFKK8+NTMbq3ZhC0EweWQYhtysixHkKWMjLTTp2tK2L/VE2rz8sgBGnMc2gP\n+NMpM9ItZZSpmEqzjCSKlllGspVFJ3GrHdW6gqR1CCll5K7MkgihHSdoduK0DsFm9Jj5funm1YkQ\n5NrbtIKEMf6+lqGMPL/ytblDkFjGEHIoo25iCI68gDL7+eRZahCEd1kqqGwhBPHEM14YDXLbGfCd\nXJQRgy/tlLo8o4xBEApB+364iDICTMpIHkt2QCV5lFELWPQbcU4PQoCGEBzFP4BjwsL0guY1L0Nj\nYFf197tfOg9fPvkghRDMals9qMwpI9vTlyKL1HQO31iQxKaMLCvVjiVH7X/W/Fr9CIEx8FYLmX3M\nz8Z3zM06qxCC51hS8gqunAcsq0gSB2Wk9iWIaUUhhLZKZ3aPRVZ06zGECiKE0ZizMI2BG4Qpffy4\nw81OGlS2HA6OEDTKSIspGAhBGM5Fz/8vYN8TgQNfB8ZQHiF4osqcMgrTLCM5lOYwcNHxwIZH1ba1\nlbeiH02cG/4v9iTr+BgLsoy80jMIJURSRWWCykYGClPerCvLKM9rVOe1vYw8hEAJqF2ybwWr0+Z2\n6fdFBsHMMqqkWUZlEMKj1wJXf1Sd26VomqyappBKu5eJIWifFWVkHsf1rp/7L6/gHzqaQSDpxhIh\n+Lw5WaSmxwbinDoEe/WqtkQI+mbWs+YV3Ok1uNJObYRQFEOIEzgntzOo7HgmhpHom5b5XUm3hWkZ\nh8V0YRJSUcuaOpcGdYxRb10xk4zgyMsPziBlfmYZQ+DHHWlGGmWkDYmIILK+r0EZZRHCyOS9gdOv\nAGqTxbMk+ehJpZ16ro0xbhBlHYL8YelNwJqFwC1f439vWY7d/vB23Fz7JN4S/gVfD38s9ncft0jf\nPyMqlf9ppBRlZHpAUum7DYLmlbiaj8Q+ysj9sgU0XWM2HYBd5MSPpy+KM1RAGfE6hHRd3LQOIduy\n4tp5ZwHv/kO6s9b24LTD98CujnL+ut5+eta+2Ehn46vR6dYYskFl29txeUWzZ4uMmEjrxKpRRgES\njk4cBoGBIRQXqyvkKA8hWFy/iiHoUz/TuoJaMYT094U7nSYMQokYgk4ZFcYQ8gvT1LHOWgl8/OHM\n7+kBu4whZPYnhiVPaAUzCF8JsJ2XrKCNUa9UluJGCDKGwN9XbhDSYaghIRtDMIPK6cZDmISEEbSq\nM9NtGRPV/TnqUgWV/TGEiImgMtEyxuTaErINi0Bcc8hWAGlfq3FnGfUQQoHYC98XiB5UpiBqBroM\nAiEEb2zNx6W1t4LohkSuBFbp6yrLyEUZBaG9Fq80CFrP90KDoCEEQpVyk5XKLe1Y9886GdjraCza\n90OZ4+y38xS8+Yg9M983dMqoOoAzZv0ctyXPt65N/yw9YlP0HkZK5CpfRkJAShkRQRl5EYLow6RT\nNkZhWlFQWRWm6d9m4yUKIcBU9gv2+hTvdpqDENI1LjQlXxBDQEFhmhpD31SgNinze3rAbhGCJYSY\n8RNawe5kE7aySXiE8XfFt7YBUwYhMdqzAFArrt1xwGewYtKh6vuQEg0hdNIYgjkkjgi08+oxGf35\n3Jo8Hye0v4Exja5kTDpRJdZU9gV3ZQxh25NYRN6C3UYf5D9IgyDTqC0HVK4p4aOGinsZdSfPPoOg\np5qWQgi695h6hlmDQEAJcB/bD5cNvtMMgh3wWuC4zwGv+or2xlgGwRlUzham7TbdnMzSu2lpi4C0\nWYjzojfinsS9EInRy0iLi8jMqIZmEOTkPuj0rwN7vNi6Zve426iklJFHMduFQ/xc6bXe/dnjsc9O\njoWDXHSHlXaaWPD+FfvxtumvO3gXVOSKaAZCSJ9VpnWFHVRWaaeOIIgQpgUNWWJ6oFSkQEZJotZ6\nlvtISSkjM4bga8sAmPfTFUNwBf+d0m3rCsf+upKSTd1+H78UoyL7zbf6mVGYZiECWZ0/NGlf3Dbn\nnXyIIKCEWAhBvAd6LyPZTkNHCAZlpD8HiiVsdwtNiCU5SwTcfYVpCePIm66+GwDwwg1X8h9kLEwh\nBDPzsakMgueEhQihRxnlS9cGIbXYBOkDz6adpkFlSoiJEAgFjv608G5dlJE77ZRkEAIx2joDboQQ\ng+K86DTc6zUIGseuZUfJ1Zx0hGB4IIEF+XWkoUkCqikgX0GVPh5xLm27nSZ74iCyZfkx2jLcWqVy\nQHilsm6I9t2ZLyJz+J4zVFA58gSVbYOQqUNgcsz+oLLeZoKBWQYBqtup3mMJDoOQpYyyIsehD8dd\nmFY2qDxRyshuscH/2MCmqXfChxDS9tdJxiBUZbsWStL1DRh3ZiaJVeNGW1GahqkPCRIhmONyUUbq\nV/HjWVc+iMfXjwqEUHwP89JO267Kf4UQTMpISlOsSz3+GEJ38iw0CJoFLkUZmS+m5Agzk45Q5ZVQ\nSkDtgp30AOLAxQgBgBlDcChgBv7C6wuJS8/S2YtIDSElW6V3SUV9ho4QjMkd2lSU32syOXHXGEjm\nc+m+K/OHgGPOMschzrLb1BrCIMwJKvNz6dlDutK3K5XtoHLkTDu16xD0GIJ5LyghqpeR3oX1qH1n\nGfvzf/WgMqc4jtp3Fm799LF4zUFz0kuHqQDdlFFJhFBC6eWKFVQOxDs1ggGNEiqijNLCNCmq5TsJ\n1CJGjPB3Rxp5o1bDer+ylcpalpGjUE5uevk9q9QxyiCEvKCyPh+Vw5QxCCZCaBTEEIrmTLdA79lX\nqTwByojXIfDPeVlGlFhBZYN28hSmOTxtAsfkybTEyFJGUjK1Euo6tMAfSSkjIu5NAx7v3F4/Iic3\nO0UI7jfSbi0AjJ+l0CmjPab1YY9ZNS+ZGyrKKH0+euCY9zJy/wakjdnMwjR/DIFTPVTblP8WJQyD\nmkHQG/ilnqtOGQnjTwjmzhww0Kg9hAkZhG4rlbMHMJ5jmHBKZJgNqBbdPoTgKkyTIikjRqhONgJI\n26TH8ibBVMy8EMwMyvOgsqSMHAaBmc3wCFDq3niDyjDnozqyooxkDMG87gYkQhhnULmXZVQgOioo\nYxDsSuVcg5C+oMSHEHTKSJbTeykje/JkvRQ5idoug+Cx95QgffG0CaYQgmYQnrtr2p8pSxm5DRkA\ns3NpDvfNxyM58/HSFCllBCb6RfkMgqpD0BGCZhAK0k4Vz20c3jYIKUKw004V5ZikGU/yEqQwB0JI\nhEtLrO1TYi7d1lWYxrqljLZTDMGFEHxOQtr+OskiBBFUJiTQ5kSKyAGzVsPI+AOcCEFu7KSMYFGJ\nHnrUlrygst4BQF2DRAjS2bJjCCw/hlBIGfUQQoHoRqDLwjSi/X9m0tFAQwg5BkGnjJSX7qeMilpX\nMPFqRQ6D4OQsIZSHZhAU1aUMQg0vmjcDZ596EPbZSQtiu1aY88BopYB8BVVWoBWAcy3ZUqKlnUKs\n6+sTSdPoweIoZkqz+mIICeM1IUpp2XmNmiQg6XCYHUNwf87sDzMQHIsYgjIECiHIv9P9J0YZTTyG\noO8qkyuG2YAzA0gXaWwDRwyhBjFXKdHMCv+kpxJLQ2wDOE3/i33dlcraBuY6GQSlKKP81hWpQ1VJ\nhCGQBkFvh28OQ+3vPm6PMpqYxF0ihEzrCv7Zl3YKcBqWGtDPQxlJo+FJO+Xsvr6vO4YAwNk7yVdg\nRvUYgqaaiegPVCf9oJQHYw1xUkb5Ss2nWFyMS7e927UjZGMyHnHVIbTjJJ0J1vXI1hXqPsNUxHwf\nfwyBwVTs+n56lpF+OFe3U5l2qtCc5nzYw96hlBGhzueoI4TitFNXDEF4ziR1vKDmm4YQnEHl8ush\npGMxs7pIWYTg+d6OIfTJpXdlPY3UBXbdk6S1vGmn2U/m7z3KyCtXP7AWr/3uTekXJZrb2SZWBZXz\nKCNaljLSyl4dE5GQ1MNSYsUQVMtgh2L2U0Z6DCHNMlIIgdUMZaXEpoysbB5dSlMU0EBT6T0cB1Cz\nO85VammlsoYQjGr0wFDa8re04ZzrfufHEHQFbQbT4fzeVZgm8+hTZGBxRjqacBamPU1TnbizyobZ\nYOksI5ITQyCEZvbXYwhOysiJEPQYgsMgMPu9QCljmde6QnfQ+mJhEGTFvUIIpk6SSMnrLBXFELqc\nVM8qg3DlvauNtQRKLaFpIQT5uH11CICgjPTzGEFlnTJKaZtyC+SQzHbpBMju6+KS5WFMg2BSRnXU\n3NA3gxD8405bV7gLqvTvfJXKpcWgjFgutFdpp2KyJwmzehkRQzm3I4kQpNfuQgi2QbBjCGaWkRSn\n0dXOpRtVmWVkOcfutFMXQvC0nPbKhGIIWRlGGcpIIoQEgV2YJuYTI1TrkpU6YABHUYkj71S97kYM\ngaVG21m3YRUs5jg/uvhiCDZC6I959baq/JdUkcWbUiINgvu4vcK0CYppEIopI2IFldUCOY4Jprh4\nQtytKwCLMtIMgguOZuoQ4M0ycr2HMXVw/pAIQZ5bo4zEuepehOCgjDwwuqgQSn+RU8ood5d8sVuB\neMTOMrKbzIEEhmGKFUIQY1RB5XyEoC+Qo3ugrmC6LVJ5Z9pfM4chcIzApeBKU0alt/Pt7kYIY+gr\nRAh5hWlVuQYIoSCyvxexDIKePmyNCcgqT7uXkX6/GbODyihJGfljCLqDphCCzDKSjIIHIXgrlQsR\nQo8y8goh2osFlKSMSi6hCX2ywl+HoCgjGyE4DAz4MvbGCaxYgzPIKX8jJbOMrH39CMFFGVkrjEll\nZgSV81/KFCHkbuYXLe1UZRl5pGJlGdmFZ5wDT//sWJRRGkMgxj66JJoZ51RPuq1rHQhbnIVpKjhv\nIpSyhWnlDYKUiSAEdxpBnvPCzyivzV+pTGjgTzvVYwjE3sp0vvVCNdf9ZjCRI6eMxt/WgzFmtKjv\nT0Z5TZQKKmvvryYpZSS+GH4KGF5rXEf+ebsb57PLIMBGCCVWTPN4+i6DQC2PxT2IbiijYoTgDHIK\niam7mRgfQjaoLGWM9bkNwsx9sgfKNHYLjHEBxS/lhIPKRtppfpZRaGUZ2U3m7MtO10MwlYYJELII\nQdINeZSRnRopxUYjgAyYMi9VpD9FVwyhfOsK7f0cj1iVyvfv9EZsZDx1eZtYy2LUU+ciPXSKxJyn\nAKoyy4hQtMRStWsIb3RIKU8H1xsAGnUujvdLb2XhyupizFw4qWzaqU8SBuyODQCARWRf/uW39wfW\niZ5GTywAfvfvwNYnAQD/3fk3PnZ5Pa0hYP5U4DsHAN85UDtuERHXnTzrsoxUtgKtdL2mMiXpRHQb\nBP5v1lvXOWoPZeQJKptHynrkeouIzNhpRdVYbWWTcFH0WjFOYpzbVv4RAgQuV+Gg04Dff9i8Fruy\n19HLv9AgyAk5kSwjPe20RJaRRAbtOEsZ6SKzkeTIVODT5CSMfXgMQQwH/qCy4ThoH+W9y7S/1jaz\n00/NLCMHWizt+01UsZj7X7/XmfjflacBAH4SvxYtVHBZfLxzzziHMkrnLcW6qQfj/e1PYFHfEXir\n+D0gBFGSxgXMrGBhzI0Ygh5Ulvdb+x3YzggBOBDLAQC30cNxUPwEUN+UbrD6bv5ffQsA4O5kf4yx\nmroP4ehT7uM6Ptnn7UaeVQgB0F6s6oCXMlq1JW3xTK1t8igjvTBNSlybCrzkw5lt08I0wLumMnGs\nFJXJMkoVwspktvmbttDNoa0L8cP4FLGtbhDSYrtVe/PpFSfMjXIqfcBHFuoDzChfGdBkuuErkHQJ\nTeD09mdw7f5fK9zHEJ0yKpllJBW93c2U2NeToYwcMQQ7VTXTusJHGXn4ZkfaqUyp9BkC/Ui57a+F\n3LlsM+5fudV5/gmJlXaqB2Y7CHFx/Brn+AAN7QqEEGucvgwqE9EzbEFyBDpajIxSIigjSaxp6EuB\nHk19sqyyZIYTY1NGWbrWfQ2e+Ahj+F/CDeODNPXwcdrF5pwWiR0JKGJQvD+8Dh8MrnZr9jt/hOrY\nunGNxyfPQoMgshUqg2ASIfzlXODvlwFLbwYAHHUu//cAshL7rvmt2pcgfblcud5ykgaU4NaXX4Kf\nRq/GyvcvBnY9VN+I/1sihkAAc4EcZx1Cysu+qv1NHNj8afpbhvNPj4sXfQDY7QjgsHep7x85/MvA\n/CHEjHmVFSbPyR2PnOxKAZVwUVSgPmG4LXk+lsw+oXAfQ2zKSPfk5hxsbBoIeiGljCyDYBnCKE4w\npS9EJTSvKy+GwLudis+MWTEE7bO2m9mtNMtpSzpE9Z1S1JG8Bem2LmfFfl/fduGdOPWHt2e2U80D\nJ+2c/a2MWJSRDcDyRHrq+5I16CdtY/2ENKhsqHr1KaTEaABoIgRxfB0hIH012whxX7IPPtr5SPq7\nFVQm4tomIpeT1wLzh7AF09Mv9znePG7EaewIgbof/1W5HCSqw5Btq4Drz8Jht38wcx7b8HUjzyrK\niBCCUCCEze0QIxu2Yi/GgJvPTjeaP6Q+ziXrrf3TCW3k+H9iMQDN+yPAlplH4MtRiGPsJ+KkjLJU\nkNo8E0Nw9zwiBGjBygLyxBAoIcCUXYD385oMOjzMRySd7CTHIBiTwhVDsCkjt4/iysSVEzAvBOMf\nk9LA6Zg+vYwjQUsqlHopI3vN6k7M0VIYBECsGwRjAMY+CUvXFc5DCEEBQjCa2zFfllH2GC5nRSEc\nxuDrtwMAmD4POPmHwH4n+rfJEyuo3E1cSLZ6/nzl/wHgc6wfXEGmlFGosvmIdS+jxBNUdsUQtAp6\nBoo3tr9sjIWnpVoxhAlQRokW/9lGp6Y/1KaYx7UQgrqGjmUQRJeFSns4cy5mGb5u5FmHEKSnMZZU\nkESdTNCYLb0Fbw5uQT+amE24cfhBxKmWNVMPVdPPSDudwhfTsIthANcDkT9ohWm+5nYZyigvhpCV\nxGsQ7POIY4k3KWYeykiOVf8sxiMDgrHiY7spTCPqvPp4youOEDTKaHBm2kVSkzAgKphsr3dgK8so\nScRzlUraUZjmpIzkX3mFadn3BQCOO3CO2FOjjKzWFWlzu+z+eQvklErtPfR0YHBW8XYusdJOuzEI\ndZjddNua05UWprlfDioRgqKM9CHJGILJuOfFrBgchWkTCCozlj6zEWjtYGyULRBCzKjhEPzg+r+X\nP5dx3u5MwrMKIQBpllETNVTZcCaOQH5xMr5ZAZ5HVuBfwxsBAFfGR+H86BR8aMpBauK5qoCNVgQq\nldJGCOUpI8aYmWVEiDeG4BLm6j2ErMK1jVeS+PltY6ppL3MMCopY3RdlqKyWAS7RG77xw3ZpEfTt\nWVLI9YaUqMneiXwIjksUM6P6tAxC4GsqpwhBV+y+wjRTgUmUZSOElDLKxBK0bV2UkaKhGEOQ885M\nVNZsa+KOpZvT83ahkBrMNAgdh0EACZxxk1DGEMTfrlfIGArL956zdQjlEIKvDiFhmutmv58OhBCB\nGokA/WhlBwioDrIGKrAMXzfyrDMI8sWqo4Z+FmWaSUnZl6xRn9usghaqvK219GadMQTxL9KXNTMf\nVAWvhRA8ayqXaW5nj0Hx174YgjVb0iwf/nfCmDvLSB+/2FMhBBkItesQrPFPqVF89Mjp2HdmDYsX\nL1bnu+gNu2BKfx0XvWEXTO2vq99KybTjgBMPBxYvBl5+Pl99Kmf/806cDULaeOSRR5DE/NyLyRV8\nfBXgojfsoraVz3wxLgZYgs+z6eggxNj6J7F4k7jG2guAE69Q+/wHm4Z+cQ2dOMFXXv88dfz9akPq\n+LWQonUIT8EcrAKLd+fbvLgyCxc9hz97ud9rqwleuctUVAKKxYsX44TdYrz0DbtgUi3G4sWL8ard\nIiyexbc9lU3GYvJK45r/W4z7icceBSFEjaHoPvf19WH33XdHpZK/HrKUz1z1MJYlh6i/u4kh2Aih\nwwI1kdIFctKFqPTXmFKiivcAd1DZCHZrNQsusbOM5LnHKxzdpeN+YuAF2PegI80BAkDkpowmE20N\ncUBTLFkD1EMIJYUgDSrXUcPOiLyZRruQ1MuRwS09DTQvy0j/7KWMSqSdBpTg1c/bCXhU29dThyAl\npETx48SDEGzvX61YJkbrzTKSY1WfifJ2eDA5cqad6vLRI6fjsOfsislTp2G/OVPU+eK1Q9h5Sh/W\nDzexy9Q+zPatmOaSkaeAkXXALgcA6yO+zOa0Pbybd1ZvAwDsOmMAYUBBNo7iQHG9I1P3Q7A19cYo\nIQgowYEIABaDJLujjQr2mzMZNRFoxthGYChVmGGyK6ZOmYqdpvSh0Y5R2bAVB4rntmXq/li9lfPB\ng7UQYy3+/s0YqGL3Jn8f1tfmYX2DG4QDxf1d3b8vxlox+isB5s4cwNptDWwabWHWpBp2ndaPtdsa\n2LXO3+2VyWzMpRuNa6bJ7mihggN2nQpKiboHB+7uWJJUCGMMmzdvxurVq7HXXnvlPQEl9nNPSnFU\nXBq2QXAgBEr1ynoNbRFiLJBjZvRKdJR+V0gZWVlGxiqD4xC9hoSA4IK55+G814pkE93QxNmgMgDs\nROyMMDP7zQDJDHhP8EfMIVvA2NFdjTPXIPx/9r473o6i7P/77J5zbsstyU1PSAXSgISe0CF0BCyI\nICjFgNQXQVEQFXgR5VUUC4qivljA1/ICPxBFfEEFVARUpIceenq5N7edsvP7Y3Z2Z2Zn2znn5t6b\n7PfzSe7Z3Wm7OzvPPJ2IxkRdZ4ytT9XbEIPIFxn1OA3IsUooQZhGq73fQpZJ5Me5ifNDMO1K+EXf\ng9frm+xQMUdOX5cDXrGq2pkv9tEEIdCkN1Z/zMmUyv6xMDf1HNMk0Zj8CKZ35JFrbjV6klbthgDt\na0iDwPuJbN07iha6yO+EBYh2RFfKFXWnp54JiP0StpwGRITOzk6sWbMmvrAL/emnERnpxMRkZcQg\niYyk27QtQsXxvZFNKh5VqRwvTlGjnaImgsDFsH5bSt+ydEDmEJjlvcqJ0AhCWGgc8Bm3j/UsptOq\nuouM/gmX2wEwDcAG93cHgDcAJNs2DCOInUaPk+cWRyEioxz5D3xAmpih4a+hxpcJFxmJhdLxHePs\nfIh8ktQXb9Ah6FDk0rkQDiEwr1V9RyXSykj+0iy/LgSh1JXKaugKgho8ToYkATb3HQvPxCSy1MT2\nRqzc1A8GIE6iwcTsj/yy6iyTp8APPhZDV6aiRqsuA5FJNJSU+hx9Ua9UT+VRkr4x2TFNj+MECILg\nGOeQ/y2qZDpSZMRUk+SksYxC2wNTxq30LRMaj0NQlcoTSdt7i3XL8H4Yg5t1roJSyscfubowxmby\nPukHAO5kjP3OPT4KwHvTdTU8kHejKPYyThCciik2pAqfQ/DZT3MsIwr8Dk46SWTkEYRCyO6DBRuI\n0SHIBMHKJdMh6NKhaJGRvvVSrTqS+iGoOyRzV4OFSR0t2GHufMCpIJ/P4fDjT8RXP34ULMvCww89\njA+deAKmTJuB/r4+dI4dh7Mu+ATmHjoXAPDtr38Fv/75rZg0YTwqlTK+9KUv4bil+wAAbr39t/jK\nTT9BX8VGoVDAkr33wn9+6TqYFqckSPQtGxuM6oXFXK8NcgRSWaYvULCtoHd4CGSRkdAh8JzXwSdj\nW4SK9Lko+gXPz0UaJwOinLYYmELMkuZUDoPD/KdOpBmTW2YOQSauE2SRUWO7J10I4z5tOChQGcWU\nW4CkOoTFjLGzxAFj7F4i+kqqnoYJ8ijDoTz6yjbyVgWlchF66ngdYrfLd7f8nEmpLLOE8SIj+LGU\n7EKIf4HGIRh0CHKERkDjEBKLjISM1beMCbcy0lryzPz4X0FenRTstc5NpTc7FQ14/0WiqakJv7rv\nYUzpaMKGtWtw6kdOxdU9b+DqT3Enn133WoIbf/xLAMDyZ5/GxctOxQ4Nn8PS/fYAAHxk2bm47qor\n8PJLL2D//ffH6teewx/+9Ffc8IPbcO/PbkTPhN3Q0tyM+/7fL7F61SrkR/tKatVIK+xGgyyJrywN\nNGN+DFVcqwe8eEQEVBBUzLY25rCuJ0EMMajfmGd2CssLCS0/P4tUfYXJBDxodhreN2NqMLx6mJ2S\nwiHI1oNSu65/QRm2cv8TaT1KoyYjv9N7gSdulUTdwY0n5xAc5FEetNAV7xDR54hohvvvCgDvxNYa\ndiDkUIFj5THALM5SlRJkTZNbCGHngfBgZcYGFJFRLlw+qcsKQ/wQxIuXnZ1CRUYBKyMobXAOIewG\n5Ir+wiVirvjB7SSRUa0LffxAvL7U42gwBowdPx5fuO4buPGWX/GPVKs6d8HOOP+Sz+DGW34R6HLe\nvHnI5XJYu3Ydrv3Wj3D95y/GlEnjAQC2bePMM8/EjnPmGEcaGCWFnPfuTBd6h9+jWWQUcbGO0KPB\n6hui1sbkdizC28RXEgAAIABJREFUAIT/9kVGpnmUsyyUHcdPoWkam7ZqxhFO2UclaT6EsFZlpTIC\nOoTgaCualVEH9cDJNXOi5PiibvO7ZoNOEE4GMA7AnQDucH+fnK6r4YE8ynCsHMrIIU8VlMtpCUL4\nh+hxCJCsjKJERo4sMjLnQwjqEKJFRrLNfD5EZBQgCJp0x2EsNPOTWtGPISRSHgpR2kFzJ6iNJkA9\ndMowLOphcMCDoU2dPgMVp4LVa802Egt2WYjlL78WOP/oo4/CsiyMGz8Oz774Knbbea40HvduBmsB\nJuWP+WLsufpDD+0RJAjJzFcBiQgAyJOwMvK9KIJmp2Yu0+eA/XPxVkbq2JPmVA5tD6qOUbUEUdvl\noiIKOHc6uSYuIXDK8MNkm3UINhjyqESKxUyIJddEZAP4LGPsolQtD1PkUUaFcigyfuvl/t6YGiqi\nPqs4d3nlAmOxIiNeTg9up08ertUQzcocQt62sI614rvl44xD8JrUTGQdxkLDKmg349Xyc79aWHHd\nMcBL9wOPQGlXb/Hq3zyL597hrvc9A2XkbAvlioOGnOWFqU6ESpH/KzyK+e1FXHn0+ETVdMsd0xh5\nObXMz354E+6/53a0tbbil7/8ZWCTsPz553H5xRegv7cHX7j6Gux+0FHJ7wVQOK/QIpEjDmKwRUV+\nP3w8ptzVANDSkHxRlQmC0CEwyzKHrrAg5YzQCIL7VxZfxe1TeD4EqQ1CouB2kY5pnkiZ1IVaG4wI\nDKlHra3kWnyCIHQIRBiNLljMj4/EiY+DwmBwCIyxCoD90jWbDET0SSJiRFSln3za/jgb6lDO28lW\nin0xtVREydZ9hxlf1xBkEAwiIytvZBvJ81WQeM2QjGmeyEja2RdyFnYf+D5+VDnGOAQdgnhFWhnp\nDbn9vsq4nPx/K67dsyWLjLbUcpQOzFVCvvX6CtiWjfFjx8C0wD7/9FOYt8Ms7/gjy87FP//5Lzz8\n8MPYf//9AQALdpyFfz3NHUbmzJuP+//yKI466ij096nzSxEZpdi0x5lJJm1q0HUIIhqsl/hevZ5P\nQejlnAgNKMFhBJ5TOQjbstxYRq7IKE6HEOeYxuRcGEl1atHthVoZ6foiN7GVrqesCA6BOd5mMlfu\nxRON5+AS9hOpL19klJYiJBXoPUFEdwP4NYAeqeM7wqtEg4i2A3A4uPnqFgEByFEZFfgEoTRgJgib\nWSNGUX+wjYh5IYuMQkNXGK2MIthokfBF9lmQL2vFZYIQ9vGFiYy4URMXoyQSGUkcwko2Bkf0fwUW\nAV/zrkGR1eo76SuPXeD9fvqtTehozmNDbxFTOprQOSpO1S9h82qg621g4s7AyqeRZHkUu7Q1a9bi\nmssvwQVnnOhakqjlXnz+GXz3hq/gx1/7XGR7l19wJj51zQ24679vACZNAxjQ5xKDOHmxYXQJyggE\nP/ihJL+ib8uwCAPRIlcdMofQgn4MIK/4AqmOaVypbEipbDQBZ4i2MgJUq6RaBW5y6AqiGIW2F7pE\nC6MidAiAF/Mo50ZBPZQ94tcH59gtYmAhZvVhSEoQGgGsA3CIPG5wfUK1uAHApwHcVUMbqVFAGWXK\neaakf3rmdXxEul5kNorI4xU2GQvp1UB9eWK85ExBYafjMN09NimVA46asmOaJzIKIwjugkvcC5gr\nJ8yeyp7ISNEhJJMwe8QL/geVSGQk6RD0hOcKlQnpN7TZar++hCthX18fPnj4fmBOBblcDkce/0Fc\nec7R3vUnHnsEJx55APr7+jCmcyw+d+1XsHTpnkD/Rl9no4yRcPTS/bBm3QYc9ZEL0Vex0NExGrst\n2gWHHnY4eqVyaq0QGC7oYjcTPxkFP7ZUZLGaIRYx2+MQeIc3fnhXzOhswVfveyFxW2vRjmngTnEW\nMfQzET6GXw86pvFlnl+TOYSgPkNwh2FgjKlK5Ro5BEVkhGiRkUi1qesHK3aT//27uZiZRxzVe/MS\nDDnJLLoEEhEExtgZqVqNAREdD+BtxtiTtT7otMihgjJsj0P4y/Nv4iOuMc67rTvjo2s/gmb0Yyxt\nwo8KXwvUl8d7WPGr+MmivTyCIFsRkGFBVAopVkZmayCPt7RybtArCsgxvfR/JpFRGIeg7f690BXM\nl7MmtjJy70c4EVkUvmRFtyV/FzVShJg5ValU8Nw7XWhvyiNvE1Z29cOyuNJ4/wMOwl+fU5nW5kIO\n6GjG+q4xOOeSy90RBvs47cRjcdqJx+IFNg2tLc2Y3NGEzQNlvLpmMxjZoLbJGh1Jrgno7k9n/BCG\nwRcZ+WJTwJ9PE9sasdOU9tjQ5reUj8AZuftwbvEiOCB8v/AN79oA8tz0W/Qh1bO9BDkIXPNFRv45\nBsNmTYKuVPZe1V5nA62TgAeujr4RDbIYNsgh6CIj91vSznMOwV2yvfS/5t2DFyW5PAgEgYgaAXwM\nwALAT4jKGDszos79ACYaLl0B4LPg4qIkfZ8N4GwAmDZtWpIqkcjBUQhCA/wP7bkpJ+ClNVMBAAuw\nImQ82rE6Vu+3vMiaa8hWRhGJbPQcwTGxjHSlsgmB8Nfw5b3iI0gsMpp1MG7DkfhG6Vjevykec4J1\nmvNC1S5X6QkI/ygZBvRop2GtWxYqdiOAPq9+ZOMCbvM9o+djVGMO6PPnW3gTMQ8qFlt2kyXD4xS1\nXbl4JHGy+KvLp+Hq8mkAgMOsfyjX+lmBG9p5U0wSGVmEUsmRRFYmbl3XIURYGUEVGXntHf1VLpZM\nSRAcZpwWhgMpxLp2oWQFOQSPI5LKOkyKkhwSmicMSTU8PwNf3I8A8CCAqQC6oyowxg5ljO2k/wPw\nKnjIiyeJaIXb1r+IyEQ8wBi7mTG2B2Nsj3HjxpmKpIKNCkrMRsm1Mmokn4LKi2k/IhZp+ThkfpsC\navELBpFRSN4CXk4iCAZvSV1kJI+nkAsNWaoeedIEPylIMpERAXYO1+FMrHGzQAXqMYaT9uKB5sal\n0QtUgxQKNAIwUHGwobeIpryNV5xJWM9GgYgwsb0Rk9qb1MKG+pFD0X9VyTgBbtTPuH7CFhvvXAjH\nWiuO+zZ+23FKoB9PZOQtqirnkAT6DrkfBcXpUylLXGQknNOUMiH6vKgnoSuVEwr7QuEoHAJFcgiO\n941HcAiuRzO5VoikzThPZFRJxyEkJQjbM8Y+D6CHMfYTAMcA2DtVTy4YY08zxsYzxmYwxmYAeAvA\nboyx6OSgdQCRKzJi5LnFX5tzU042tuONCX7I4IEQghB06vKPmYHFDM+HkEBkJMpRNIfA4K+FObsK\nDsEgMkpsZcQHETglfzRn7T8Lr37p6FgbdBO7nwheBV3SHlGFCMUy/2gmtDWiB414i/ENx/jWRoxp\n8cfqreWk1pcOEg8xEd8lFXrRmYKX2JTE7ceh7iKj3T6K/zfGFxR4fgju1NOTHqXJhqfvkKNERjmL\nVLNTpR13bLrYPkpkpJmdKhxzFWJuRfzk9qAORjokSyrno2Q1Skpll0PwCILcHPOfXSWdqDEpQRCt\nbiSinQC0A0hm7D3MYMFBiVmeyEjENqqc9EtQQ4tXboCFxQHS2jPszOQQF8E5J+3UElsZCSJABsc0\nFarIyDxxw4gaY8zb0SW3MtL610VGjKdsjGuPQg/SIDlFIZL1JaYKtYhdfCFx1AIcSmCkvvtRMMbN\nSgsW+DE48JTK2q7cRFTjoBOEAIcgtWVbhHIlRQpNqEEXdTAWEZiviqinFYbEOgTHNTsNiIzscA5B\nb01wCOQMBK5HIamV0c1ENBrA5wHcDWCU+7tmuFzCFkOOHC4y0j6yAUddtMI4hADLGysy0jmE5FZG\nBBh0CNEiI/keciGONIFbkIhXxRMZGatGNwR5cU2/oCbf38c0kIRDgKQvSSgSSr4LD2TCjsX4tgYg\ngY9k9WSqFgKXHLqnsp4FL409vy4yGmB5NCpmp1JZcjkEjSPh5YLi21grIwCVSpgfQhVzW3FMi6bL\nYTqEoqJDcAmCF69XtTISSmVKySEktTL6ofvzQQCzosoOZxDI1SHkA7uuAUddPAMJ60UbAXoQZBGU\nXUxAhyDJZ4TCJ9LKSBIZGXQIupVRTlrdwiyFomIZRe+aTQOMbjvxlpQM4rXECHvYETUUA4AY7sW3\nF0w1mrhrqshDSzxUAyJl4zW1HNKmoVFxC8JRWdxROoKg7n4HUEATmZXKOVskyFH7k8eip5aMfBaM\nQXayVulBFRyCrEMABeIqyRA6BLPISFUqm3UIzH92g0EQiOgVAH8H8DCAhxljz6bqZZiAiLNSRccK\n5EQeqJCygyiGPBr9JZnmNyEYDiLYgsshkBURRdHlLeXFwqBD0HdKAnZCDsGPu+TvsBLHMoJ6jyaR\n0ZZDGh2C/9t4rxT8Gd6qeoWRP5SgXkSRE8WOM7SvyKrBi4P7FiQLF7G71UVG2rRIAr0o1yGY2xBK\nZaPISIxNmotJUmiq0VPlgaV/b04kh6ARhBAOYcBu8gfiiYwqblm1ObtKDiEpqZsP4PsAOgF8lYhe\nIaI7U/U0TGDDQYlRwHKjrwJl8dIDSwmE7a55naBSOVZkFGVhBCTQIajjkZXKYRyCbkMv22n7IqMk\nBCGKQ0j30UhraO0io5gGbNvGsQfvg/ctXYIPHr4fvnHD1+G4ypOHH3oQ7e3t2H233XDcgXvijA8c\njQf+cK/X8E1fvw6H7jEfixYtwk477YS7777ba/fW23+LXQ49Ee85aAkOP2BvLFu2DBs3blTHJI3t\ng+85AscduCdOOGxfzJ07FxdccR02boo03osQUQJXfe17uP57PwUAfOGrN+H+hx6NbqtOkDdSjjZA\nT2RUhZWRvlByHUKMHwKCYkB5fnstM/V7DfTMgLJUQf1mqiAIjn/vBG2flJRDIElkVBEiI6b8BfhT\ns0golQfBDwE8tHnJ/esAWO3+G3Gw4WDAIVRIvfV+jUMAgH87s/HryoHKuaD8PTg5iFQxjHpRXHCA\nSjnawkg0EGNlJCMJh6Cf9pTKSGllFKVUlkaYDFQDRUgnMmpqasLv/vwIuvpLWLd2Da791Ll45e01\nOO+T3Ols//33x29+8xs8/fYmLH/2aVxy1qmYMWE0dlvM4xZ9ZNm5+NoXP4/nn3+e50N4/UUlH0L3\n5MVozFn4429+jTWrVyHfOTUwUoEvf+tmLFi4K+aOb8blF34Mx595Mf73vr/E3mkc/vPSc43nB4NT\nUHbe2gi9UBLu6TRWRgGzU5ZXxLHyFBWeynp/gKxD0PblMVZGRse0wEEyyKErQPrePyGHYDUBlhtq\nJ8rsVNYhpPRUTsohdAH4BoDXAJzGGFvCGPt4qp6GCWxUUGY2CgU1iXt/mQIT5r3Fa3Bb5VDlXNju\nGlAnmB8OQocmMrLDaTIBKodAFJBfRmVMC9vlB7gcj2lJaWVksO/2dcrpRUbyfiwVAtQ3ucho7Ljx\nuPnmm/GLH/9AMxvmBeYu2BkXfeoy3HjjjYFmw/MhkJcPYYcd5yQaUaFQwFc+dxHeeHslnn36Kbz9\n5ht4/9Il3vWffO/buOnr1wEAXnnlFZz0vuNw0tEH4b1HHYrly5cH2jv9E1fif++5HwAwY+9j8N2v\nfRkfOupA7LnrIq/8+nVrcdhhh2HBggVYtmwZpk+fjrVr18Y+Ox2qB7A/J/SkNUA6HULQ7LTgiox8\nWbyA7YmMDHPAMBUZi3aDZEyN1JrWzFhHxX0eYjhROoQKzH4IJeSCOgTDXTAwKXRFOse0pBzCyeAR\nT88DsIyI/gbgIcbYA6l6G2JwPwQHFRAaGguAFLuur8KCTmQhbSjH0m9/KlJykVEsh+CoRCPAIaiw\nFQ7BPHH1s/J66omMEoWuCBYKWhlFPNR7L3OD0QHTi2UQcda6MW8lCjXswSnxDyTfDLRNBo6KT+Yn\nFhOLgFmzZsFxKli/do1Who9+wS6L8N83fSugBwjNhyAhzY7ctm0snL8jXnrpRUydszC03Nlnn40v\nXf9NNI+bijdfeBLnnXcefnHX7yL77BjTiV/e+yD+eMfPcP311+M/rroe37vhv3DIIYfg8ssvx+9/\n/3v86Ec/SjFaH4qylvkbISIK+CGkWUtXYoxy3I88LPLJgIlDYIZrng5BC38d66ls2ugoLSaH4zDP\nLyj4DHSRkdkPoUh5QDjSlsPNSTmH4D73wXBMY4zdxRi7FMDHAfwOwOkA7knV0zCBBQcVBDmEvrKV\nyMolzGQzUE78CIiMxHbF4dQ7iiAQqVZGCXQIlpWAIOgcgsfNsPqJjOLqD4qW0yAvCIEvwtCehelI\nmxe3/vC7WLRoET71qU8Z8yG8uPxZHHngYsyePRt3/O+vlLZiH0vMHOzZvBl/+9vfsOy0U3DiEfvj\n05+4EO+++26gnAOgj/lza+mR7wEA7LrbblixYgUA4N+P/x0nnXQSAODII4/E6NGj9WYSQR6yF6GT\nQeEQfAKcfDF9xFmADw58AejcHgAw4IWuMM87hysGAv14JrC6lVGUyIipBESlB+mtjOTQFQEdgjdQ\nvtlzIHQIvFDf0d/CRcXz0FWYIDmmqQQhoEPwREaDY2V0O4CFAF4B8BCAjwLYMhqrOiOHCiqwUMir\nC3FfRXcnNyMY1Mw/ZtJ6ZEmLrFrczxPAlcoRIqNKEWD9QEOr1LCWU1lbaHKKH0KYyMh8B3JAr0Rm\np5FK5QQ46jrv5xvvdsG2CP2lCmZ0tqCtKXlmLfSuBza+DoyZDax/JVEVmSC8+uqrsCwbY8aOw8q1\nb0qFADDg2WeexLx587w6p551Hr52jRQOu3+Tlw/h4H33xJx5O+HeP/8dX7ny0+jrTZ5vo1Kp4Onl\nL+OSOXNRzNkKcRgYcJOvOw46Ojrwx78+ipWb+jGutQGT2puwulsO1U7YxEZhNTq8M4UGHjbEtm2U\ny+nECHEw6RBCOYSUbT/O5gKjJgLrXka/8FQ2NGJbhLLjy/1NIv90HAJDuY6eyhWH+aIuCol2mmsE\nipu9PAiewrh1Iu5yxmK+wwJ+CN6QFB0CU62M7rsi8TiTkrovA5jDGDuCMXYtY+xBxlgwWcAwB4Fg\nE+cQbJ0glFlQvGNAWNgH/Zw/CYOjAAAvyUUEh9Cw4UVg4xsqVxCTU3lMi99emB5A39HKJrK+s1Z1\nHELAyiiNDqFuXEMSDoGX2bhuLc455xycdPpZQc4JPB/Ct6//L5x//vleu6bWRT6Et95Z5Z5hXj4E\n3l/0eEqlEi7/8rex3eQJWLDzLhgzdjzWr12DjRvWozgwgIceuA8A0NrWhpkzZ+LuO2/nvTCGJ598\nsirLrEV77I1f/YpzMH/4wx+wYcOGKlrRPYB9HQIhOJ+qSjTTysOcDSA8uF3Q7NSvbrQyirQxUjdH\nQD04BCkfAvS57h64a4GXZEhYTLki47JCENTlVxFdKyKjAeCRGxOPM6kO4TkAlxPRNMbY2US0AziB\nGHFiIxsOysxCvqAGWustA06CCAGmRUNAMTuVdkpaA25hlszKCEglMprc4QdlS8whSPqOqmIZSfA4\nC+F9LbibuKaim02IZBSlr68Ph++/N/qLJRRyOXzsjNNw6Il+PJ6HH34Yu+66KzZ2bcbozrG46svX\nY+nSpejqC2O/9XwINto7OrD7ol1wyKGHoaiUVA8u/4+zUSgUQE4Zhy5ZiLv++wZUAOTzeZz9iU/j\nlPcsxfiJkzBz9g5e/dtuuw0fO+vj+Op1XwZzKjj1wyfj3Is/negpyAvRxy/+DL74yXPws5/9DEuW\nLMHEiRPR2prsfckwKZUd5i7SFZVDSKMa8uAShGYMuO0ECbNtERzFMc2/arIyivVUZqpOMSy0SFI4\nLCL8tcch8DWpom36bJcgVBJyCIAvMrIq6fbtSQnCLQD+CWAf9/ht8OxpI4sgEBcZObCQ0ziEnhLA\nrAQ6BP04ZPWKVSqve5lT+QgrIw8Kh0C8DdfcTA9dMaHN142ELeph57nclP9OJjIKOqZ5XMmkhcBh\n1wALTzb3FWgr2i48ERKyGJVKBau7+rGyqx+NORs7TmzFU29xf4EDDjoImzZtAgA88/YmOIxhTLM/\nV8695LJQ6y2RD+FFa3s05C1M72zBxt4i3lgvx6Tw69752z9gVRf/YHeZ2gG88wQAnokKAE458+M4\n5UzNmI+AmTNn4pd33o13JZHRmu4BXPXJcwAALzrANTd8163wGlY8+ls85XQCAHbbfQ/8+c9/xlNv\nbURraxvuu+8+5HI5PPLII3j88cfR0FBFRFpFh+Dfn/ANkO+6qvwnbZMBABNovSIykpvKuSIjY+gK\nI4cQTzjlHMy1m53KYjNNZKRzCEwVGakcghbcTm8DGodQTpciOClBmM0Y+xARncw7ZL20pTPb1AmW\nmw8hn1cn/u+fW4u3etaF1PIRZWVkKhdco9wLf7yG/90uQdBYQUS8bVbOczhhjDBvUpvXT0vBJx65\nRAGJVAKRysrIpFT2v3xg3/9I1L+HaumBN/7kSuXRzQWs7OpXsmIBBk6F+Sd1aVj4ePyfPcVK6LW6\nfEBVPDOxaL77zls464Qj4TgOCoUCfvCDH1Q1BLMOgXkB5wBZZ1NFB7OXAgAecnbBYksNEidguUpl\nkx+CgG7qGZ8PQbYykl9c+MexcLt243k5/DViOIQSqZtVsvMeB6Q7pnlllLH7OgQrQDiikZQgFImo\nCe70I6LZANKF0RsmEEplXWS0fHUviiEB7WQELXT83/5L9r0pA1+sPlMTiYx0KuQv+lcetwBTFu6J\nS//3yUDRsF2+rluQlW7pREaGtqvcJxDgRa6pXmLkCwzikM9ZmN7ZnCrpe9qhOA7Dus0D+MSyU7Hu\n3bc8UcFAuYKLLr8KHzj+GHPlmilFeAPybJw+czaeeOKJWjsL0SH4il55TFXNj/FzMav/Vjiw8BmD\nWQfg+yGIO1RERmQQGcV0KZtg+6M3H8nYZ/ZY/P3ypVj8ZdUivyKHrgj0rxKELtaIca0NsIvuecv2\nn2UipbIkMqo3h+ByAt8D8HsA2xHRbQD2BTc9HXGwwLiVkcYaVxti2Dw9fVlpaIIcr2ACmuzVkTgE\nF/MntwPNZkIW7pimNe/+/dr/veglt2/IJXkewfaThc02t1WzyCgl2puiibHPEJDx2C9oXi4E9/GN\nH96KeZPakLct9JcqeHFVt1IurF+BmWNb8Nransixyoh8iqz+Fr+m0BXM1SEU3WcQZuabuA/4Nvym\ntrjZadAzGvCfZ7pop8zTfwTGHaNUHt0S/B5VKyNP4is6c2+Cf3svbyJ0jMljxqgmYAMAK4ecRTxh\nT4hSWR27rEOoc/hrxhgjoksBHARgMfjzvYgxlt6lcYhB8HMqNxbywEfvBn56HICgy33iNmUOQTpH\n0ocRHIUEmUO49FX+96t6QFmtTqRmjnDlsfPxzsa+CMc0nUPwjz97J3cUG9WQnFDJ9xgXA4khglWX\nH2AqhHBjNcPdy8lisBQQu+Opo8M4kWQkIW9baCnk0FMMNxdNPrJ0zyiJb44qDffv07Z8okjeudrY\nH4skkZF0XrRbNnC4Jn1e6nwIKXQIJqLH5GkEAoOBIrhrwWY0ufcjOIScxCEk8ENwHNhuLCOrMjg6\nhH8BmMUY+22q1ocZGHjoCgcWGnIWMOtA6apPvfeaMQaPvrY+dft7zuCOPUcsmOjrEGJFRtJuoqXT\nvG0x6RAEAuUZzth3JgDg+Xe7jOPUv0nTN9rSkMjkKnBKpMsMw+sbS+js7AJrCzpBiWdVs8iozuot\nXXWQpHlZBt0gpTKtZmQkVdTrM6VQQiSkCYwxrFu3Do2NjbHlBHwOgcEmWYcgREbx/QascNSrxnsV\nBKGicSS8RnBzFsch8LZCOISYh226ysNfwxubsW+XbehhjRpHYnue2IlERtLvwdIh7A3gFCJ6HUAP\n3K0TY2yXVL0NMYTDRhkWGkJNMtN9snL5uRPbsOI6Lhd+eTUXCwRFRhEEwXTddI6SibfCzE7DPJVl\nJOIQtKn/7NVHoCWm3rcf3YALAWw/Zj3YRn+hWdXV7wUnczY0KItoLEp9QM8aoKkM9K0H1jIg3xRf\nT8KqDXwnRRsbkHN38ys38TH1NuawqSmPYtnB6u4BvvhIY0e5H9jsx3pcbfN5sb5gY31PSWmzXHGw\nqot/zAPNeWzs5aasz3c3ARt5G73NOazv8U1caWMDNvSWMFB20NuYw4amPLr7y9jUV/LGtnmgjDV9\nvP4qVvLCuz9Pa9xzfCJW1hfQkLOwamO/328EGhsbMXXq1MgyRh0CuFGD2LGLmZJEpBiUsUvXJA5B\n/i4EQShVgpsKM4cQZ2XEVCsjfRARMH1PFd3sVO3MLcTnhcchCJGVJzKSCEIUd1PxDRnqrkNwcUSq\nVocpWIWzUg6scIUrxb5vBeFl/Z2S6byHJEplfY+q6B3C2w93TIvvMW5hlxs6a/+Z+NYfX060iHcN\nOLj2oXWYPa4FD3xyV+/8J254CG9u6EVvsYLbz12CedPHRLSiYfnvgDtOBo6+HrjvU8AptwM7HBpf\nT8JRl3Hm90+fOggzx/JUqmd86QGs7OrH2QfMwmePnocn3tiAs277Gya3N+Jvly/1K6/4K3D7id7h\n5eN+h7amPA7ccRyuuec5/PsLh6HDNV19c30vjr31TwCAq49bgCvv5qlFVlx3DHDVYgDAncc/i4vv\nftJr78FLD8J1dz6Nv768DhccvD0+dcQc/OChV3Ht757Hx/abic+/Zx5+/NfXcPr/8TGcO/A1rGCT\neLuNH+b31/9zAMBPztwLe0wfjaOvvM/vt0bIhlqK2alBbJMorHoE5FhGMnIeQTBxCO7YNA4hyhGV\nQYtlJE/tOIJgak9RKusJckQyGy4S3Mya+LdbVpXKKkHQ+5SIF/NFi3XXIQAAY+z1VK0OV7jJJMrM\nDrUuCVMShyFsboTOGf1CXD4EwN/tjp/n1oniEPyJEe6YFn+PzYUkIiP+DC85fA4uOXxOfPmopiiY\nbjFVZcCP7FiV95PblPRbPD7xvMQmIphGVT22iC8A63s4N9HWmDcWDdss6+/HIjKaWqpD8C84EQEI\n5ET09ULYkXh7AAAgAElEQVSYlZG8IakmlpEJBFmp7J9vyPP52l8SBCFIjHoHyl49rssK7ycYy0im\nMNHzy3SLAaWyqaLMIRDgm1FbyFkW57ZyZj8RuUsmjdseJB3C1gGXEldgRwR+S0cUQq2MDKZuvII2\nmULyKSsYNYErwCfvGmwjYlaHO6bFd5lsUa7+49ZHLUIPiN/pIAiCyypXEVrAa0lesC1V7q0ThkD/\nXhvcJn59Twmjm/PKwqgEXUss0vPPJdlnxIlCkoRoqRaCGDEwI4dQvRWa345JqSy40/6SIYOYe/TA\nci5Ss62wFdmHHOhR9Ku3Fz7G4HWHaYu2KjPif1zfoh40umP0dWKW5W6YGts5l6CFtVbNTiWRUUpP\n5cExwh6msFxWqgJCPsRpyyJz8KwwhO7Y3L+prIxCwbgCvLHNHWQUHffbD3NMS8sFhXdVp3bcpnwL\nkbSVhY1vWT2uZhyyyE0jAH48npD+vXr8vb+zsU+JLQXoHELY+9GOiQJ5JgJMivRbD2ciw3E0k8c6\nwGTfL/wQdAyWyChAEAzEyD8mN4VmtJVRaOiKKu5Bdkzjwe20zgAv/3E3a3LLemZ3yFkWJwhEQHNn\nXGfeTzulUnmbIghCZMQ5hLAUmemajNuxBVUIOkFIwCHoSS4UkVHQyshrOmzBqdtbrx9BUMwE07ar\ni4zqxCH4ukszYTBWcss//fYmPPTSGhw8d7xyTbGdD+XggiIojx5oZX1fPPlK+PPjylR5N1k7t6D6\nIfjhTGSCYBLzVA3tvQC+34yJQ9C7FAQ72g9BJXRqTuUqg9uJcQNa5+5v13KoRyiVJQ7B0yEARoIg\nZ5dzpPUircho6yUIA91Av2Z26YoUyrAHnUOoq8goQBCSSfrCFef14hDqN31qii4sKoiE4om4ruim\nAH8hEQt3eKqH4ALe3V9GR1Me5x20fWj7YeKToFmwWeQUNgKHRXAIWtC2ekiPjKErmE4Q3Gc4WCKj\nvOAQXIcsCvYtYLk79GjRmhbLKJLExEO3MjIX4vO3jxX4s3OD+sHKu7Ga3J1/Cg4hrVJ5ZBOEda8A\nf/yieVZfNx347hLllOVxCOFWRml1CHGTIziyKkRGegztSLNTaQeaMGNa1aiFsERwTlXrEISJXRIi\nG9aS1LcYoniMYn2I4xDW93BZ8LL9Z6Fdy+ugLu7xYwD43cUrlf3fuQhrL12HUBd9gtKEJyw13mut\njmkUIzLq80RGwRH5Y6EEfghROZWrS5AjD8QoMmrjlmG9Qodw8i+A994EtE5QdGxxBGHb1SH8z0nA\nQ1/lyVFkrHqOi4e63lLPSwQhFyIySmvhEvuBxomMklgZRYmMDI5pAoPOIdRVZCT9TjsrBcdUEgSh\nioidLhRZvCAAlsrtxRGEl1dvBgAs2q4DOvRwCyYEw5OTKm6Qx2iI3XPrssXGdgG+ML270V8k6qFe\nDiMqCofgWRnV1lcYJ6mLjGSY9C2JPJXDcurWqkOA9tku4qbBOPV2XFC8EF1o4Vxp6wTvmuzTgeYQ\nk2zRqOM/A3ub4hCEt56mJVvx7z8ai1uOTBDcl9o6WS1D6d53WNHBFRklc0yL4oLqgnoqlZXfKdsV\nz7Dohpmuk8jIa14jCHEiI/Hd7jB+VKAtdUELez9BEZToQ9drmMZdyIWLFF9dsxnH3vgXaaz11SEI\nBEVG/G+SzUjUpowkexp5nvhKZVdkZOhbaiRBPoQIglDFRkhJkENa3K5jvgZc/hbQPhX3OEsC4weg\n6hAGuNMr8i3aoF2TW2k9tFi67HgjmyDICetdrO7ux08fWu6XKUlKFcEhMNu3wLnwnzh78h1ekbS7\n57h8CPEio3rrEBKIjCJucecp7bho6Q7xY4prKCWSiFLCKwuCwHfmyNVAEAwfuxiPmGaB5xryHMa1\nBjkVeb4kDj4ocwgJdAhWhEjjhVWbleN66xDkc4rZqfu3dpGRrEj3zzfmw81Ow5JaReoQND8F5Rar\nFBl5Oii9PcsOJJLS50betlAUOT0LLiGYuofaiauDYME0jYkxsv0QDB9Hz0AFBUiZrXrWAh08vo7Q\nIZRh+Y5phWYU860A+sOajB5C6Hl+Jd7KKIkOQSMIFGVlJDWd0IpFxm8u3C9+PP5AUpRVYfJD8FpN\n+xJEkqFS7RyCab3S0z92BKLLqpUOnjMOL67abLwPRTSmXz7hFmDlU5FWRmFQZOYRi253v5r1rS4q\nBBOHAG3xT8EhRCFMNxCwMorQIRC5nsIxZqeVsOtV3oPY9StELQQ64WxtyOHdTa6o79CrgdmHAO8+\nBbz2oF9IrBMsKDZLihFFEDb0FLGpryQp6rStm4sGhSCs8QgCuQ9MD12he/TGLUhNeduovJLhKyLr\nIDLSDccjdQhS04NuZVQ/DgFRC2UcPA7BDRFdA0GQxyHCd4jnNW9SK644eh7eu+sUrY76Tm85Y6/w\nocqckH6jO70f2On9oBdWK6dVDkGt4lkmKp604Q9QTwNaD5GRyXQ11FO5Vg4B5jDpnsioXPHKeXUM\n4jVuVhreD4MqMlL6rNK6zg8prye8DEJ/Tm1NebzoxkdDUwcw/3hg1bNqJXd9YzU4mowokdFbG/vw\n3T+/7J/w5DL+A2CMoUDSpO/1s6AJ2ZpudiovnHKY2jA8f82RfpthpUNFRhpqVSpHIIl4oTbUU2Qk\ntZqaQ6gfQZDf59hRvB05ZMVZB8wKioJSjDeRCalhAfNzMvC/h87j/g0n7D5VDFwqH/5Zd/Wrc2nw\ndAgMtoFG1eqYRgQ/ooPUlM8hCLNTqY7eBlwdQoxSWc/BrA4iPYRYi3MI0c9dt4pva8yhq0+XFGjv\nWawTrq60xNLneBlRHAIAtBTkIWvSwPWvwimPDnIIojSTOQT/YeakuEZpP49wDsHMvVQnMtJYwCRE\nRMLYUQ1Yu9m3NqifUrn6/YT+QciLY1gMplBYmsgoJN5LEsjPZqybLCh+OMnHK7cflqzN5IegcwjT\nO1uUwHRRMnMZmzQOYbCsjLjIKBj2u3YroxClsrvY9omUpREMk1DqRjumMY1DqB2NbrylsEcgx04y\ncQjd/SVureQpI7QFXyMIA8gjj3TioxHFIQC6XFLiEDavBr61Kzr/ciUKKHvhf9Er5TWQdAi5EJGR\nHJUw2HfycYomYjOmVaNU3udC6SB6qj7wyQPxq4+rZoh1S4ddTysjqSnx4SSGIAjFXgCU2HHPOA7p\nd6cbdiKQFzlQya3V3AlcsTKyaBJdiVGHINaA0CEk08Fs6tUIQh3CWISJXuTvJY1jWlQJsbvnbfrn\nC25nAyaRkdaizyFEINLKqDoIsVaYDiFUZwGgrTEPh0FNkqTbZwulspt8px/pOeURRxB65QfixbCp\nAH0bAADNb/0FDSihD268esf/ACz4oSvkOD86G2uakO/bdQo+snhG4Hy4HwK/EB/+OglB0BakWQcC\nsw6C24FWWD2ePW5UwtwG1aCeIiO/LcFaJ4YtWRnZhZoIlbyYjnVFQ+s2x9hyi3mYa4zNw6BwCDEW\nat4x/BzdofNN+m1FiBSLFZUC1CfQnVmrrPohcNRFZGT4bVmEgm15HEJYOXHcX3Lwp+WqrkYGAzQr\no9qfk88hmPUgCkeiXW5r4t+wIvILcAjuWidxCGkxogiCRYReZbfmvmlpB83A0EBF9JMgCG55xnBO\n140ADI5p0oQJe+1XHbcATYaQ0HERRQPtBTiEKqyMAOC4G4HdTgNmHhBbvW4cQbDhqqtOaFOzcJFC\nENJyCJIOoQZxEWDmENZtLqaoFY1kYSiCOgRxKkxnpS6AycdTH5GRuV1TGssaIpO77YTn3m7IWegv\nG0JXBFvBvc+8i+Uru0P7YYxF7tirQRyHoOos1AKtbgh1xUpMJ/yaX1Y/S88hjCgdgkVA74BEEDyR\nUQXitTPGUEAZ/dTgknl3R1TqxawKz1lchq3KqQNifsMUCuUEQs6740mdMc0EE0Ho2A447lvGnpOO\nUUdnS9oJVB1BOPuAWTjnwNnGlmyLQnNVhEKYnTolwG6rakwC8kIybUwzAKA5Lp1oKqWy9DuNDiGm\nK+V8Ct1OfZTKJisjpnxjdcuHIC2melsNeQvdbs4D5YqBQ+iNEQMyqDL9esDjEMhMiKN0FiKnhqJY\n1jkEQRDcTXCxCg5hZBEEi9ArXNP/8Dlg9XP8tyxSYdzsdAAN6jWpjMOs0NDQYVZGobLbOCujOJFR\nEgVxKrviWFc4I+6/5MBAqOZYVKlUft+uUwJ9ibWjMU3qTK+y9AxrCFsBQHlYe80cg2+etAiHzpuQ\nsFIC+XgCDkGXsxPF7/oVmXkKzW1dCIJh4WTQ7sP9WbNjGsK5Gt+sU+OYDDqEOOh+CPVgFjwrI5BR\nBCX7k4WKjPqScAjVi4xGFkEgQp/QIfzt2/4Fp+LNAAaGAkooUQEA+YuptMsuR4mMQt58nEdycKwh\nN1EvkVEKJBEhbG8Is5Cg4SpGY67W7FqPpRYXASqXVUNgOyAoejl+0ZTwwl7BdETMIs45Jo1lpHII\n9RUZ1UNmxBgCqyxj5hSayUJXhF+zyLyYAlBSuKp+CMnbF9D9EOoBj2CFcQgRQQc9DkEWGenzTsQt\ncse91SuVbeKeyAE4ZV80xIAGKqFIBSiZhaRFVVcqy2+HwTxhwjmEkPMkREZxZqdVKJUjYZKYDhbq\n17JI2VkVQZCtiuqoQ0heKV0t3fM52Jx6PomVkVo/jcgocdFQhDVhUirX1VNZa+vVtdwPZf6kNrRI\nYj69xySxshhjCudTj8SjPocA40OLFBk1CZFROEFgJe7JzNy1boBt5UplIklkJINVfC89l0MoIs9Z\nKifIIVQ0s1NDT4a+ExdVTsdnTKs3QTCMZbAoQh0bFl7BDWktjACXdRYyiRq8lFGjAj5h3dDMa6IZ\nw5jidQiSKGpLi4yMfghMCzDHfycRGUUNSQ5up0MsuHeev4/iWxQVCDB0DKgPsZQhOIRAxjQX0Upl\n/n10y1ZGTWo0XSZERqieQxgykRERXQjgfAAVAL9ljH06ro5tkZcoW4FT9hZNxhgaUEIv5QHYPueg\nE4QwxWXIJAi17gg5Lz76ulgZ7XFGfJkI1C1lZqDh+u0nhGlsY64KDgHghLVSrJ0gVFUpXS1RPE1o\nkbicymrZ5M9w0BzTmOrf4+elrm1MJCkR9Lbu+8QBsC1SdAleHfk4fgjcD2HQdAhm0XSU2WnettCU\nt1WR0bi5aiFNqTxidAhEdDCA4wEsZIwNENH4uDq8Xoh1gOMoC34BZRRRqIFDMPddzflYkVGcE9Xh\nXwQWnxs/wCgMnsyoboi15ImD5RKEGkVGVYk0UusQokVGRoIg/iaSwftlNrFmPM+mh5athyVNIrNT\nYWVUq2Ma+bnCdf+a6Z0tpipGjisOjiYyCmDHo4AX741tR4bPIcRbGZmIbFuTFr6iU83GxzSl8kgy\nOz0XwHWMsQEAYIyFe4hIsIlUxzQBVvE4AebqEMqU5x8qC1oZNaGoeifLTSHdhi+uaLzIKOalpdjt\nhTYxAkRG4uOuWpFn54AS6qpUTlErVWl/t5zCUCHFDlu2Mlo48MPIsvWJdprAMU1wRTUqlQmEfWeP\nxaVHzMEpe09LNL60SmVh2hrJIZx0G/DG34F1LwOTFgba2HPGaDy+YoNyTuUQgv2Gxk5y0daYVzkE\nffNTdqOhOiPPU3lHAPsT0aNE9CAR7ZmkkkWE/mIJ+P3l6gVZZASgAUWUKK8qlSv+gywiF8quh1sZ\nmceUNPxAaEOxBCHZK5o1jlsJdY4KtjfclMqmRyxiVFXtDCRMT2s1O60JKXUIIa/WbNRA4Rehztsw\nkZGJK66HstRMD5jxG6tVqWwR5zLOP3h7dDQnXfDS6RDylhWIZRQciA3M2BfY/TRg8qLA5V+fs0/g\nnKpDiBYZmbpua9IIAoBHOt+Puyq8L51D6EX6b2HQCAIR3U9Ezxj+HQ/OmYwBsBjApQB+RSErKxGd\nTUT/IKJ/9PX1YkFlOfD37ypl/rx8JTb2uIlwGEMDytzs1CAyuqV8BJ5jM9TYL3qfpnGE6hDC7t/t\nVn+zAZFRDAeQMLLpJw/fEbd+bG/sOSOYXm/wPJXrN32EVUj1HIIgCLUqlauplW7McSaYZh2C+zdR\n+8FSZ+0/UzHLFKiH4jRMhzBYnsrp62jHMU/RtoI5l+uha2lIwSGYemxrzKlKZQB3Tb4Yl5WW8Roe\nQeB1u1lz6jEOGkFgjB3KGNvJ8O8uAG8BuINxPAbAATA2pJ2bGWN7MMb2GNXSghwFo3Hd/vgKfOcB\nniWNcwglziGQDbAKShXHM8V62Nk52If227hDS6tDMLQd1cBfXlprzAebdIXK2xb228H4CAePQ6ij\nH4KwMipXm+1J6GJqyJYGDKICXu4jxsooWocQ375lWHVP2Xu6ccc+aOGvAaOncs2xjOpQRx/CNe/d\nSTnOWRTMh1DHWEagkOB2EY5pAA9foeezYEzySPYIAl/ruhEdV8uEoRIZ/T8ABwMAEe0IoABgbVwl\nIhiNziw46Onn8WaIOVyH4JqdVioV7HDFvfjFo68B4D4I1SCUE4izMoqaR4d8HmgajZdWdePUHz2K\nK+96NlhmOOsQ6rh4CoJQqVQrMnIJwpBwCOkQp0MwEQqfQ0iilA2WKeQsI0GodaETi2aXthtlTDc7\n5X9NSuW/XXZIbD9yHKC0CEuhKbDf9upGKmdTUIeQvltceIiq9BUWdGHvUN4Mmfpra8oF8lkwMFRg\no8wsycqIt9PFzEr2KAwVQfhvALOI6BkAvwBwGks4M00EIQcHthvy1XIj/hWpAbBsFIucUPzp2bcB\ncC/lKPDQFSaWPYYVCJTnfyN3YPtdDMCPYPjCKkOwrTqIZQbP7HQQlMrVLlB1EhlVJeP205al6iON\nUjku2qlS1rDoNuTUHCACtW58RZ7f/Qa+iT37v+O3CxhzKpvueXKHupM1zde7L9gPlx01tyqRkSlH\ntdqfCv6cuJXRbtO4rf/eM4Oi2Dh88vA5+PEZvnpUJOXiSuvo0BVGK6PGPNb3FPGLx97wzoliReQ9\nT2Xm6hCq4RCGxMqIMVYEcGraekSEAoJWRhY5sEkQBE4AhMjIqfDyDTYDKkCZRd8ygzkfQpyuwDRW\n3l4E3MVe7NyMNDGhDiGym8Ha9dZRhyA8lavWIQil8lB4KqesHeekFRVcMWwBl8+b6jfkbWM+j1p1\nCCKcdhdaAPg7UsbM+RCqjWU0Z2Ir5kxsjS9oQFwsI/1x5SyfQ9h3+7G447x9q+qXt02B34T40BWm\n9yyI72V3PI0P7bmd4uA2gDzyrqeysLgcVjqEwQCBWxDpyKHiZQbKOZxKDkgiIwBotPmjS8IhGPtO\nxyD4dSLdLlVZsnF3XMdFt/6oH6VpcuWr5VrMToHYfARhWDi1HcCWFRmF9RWlVK7WKqhgW0b5fa06\nhFIlXOczGBnTqkHgtmOUzDmb4DAWUIxX1XfIeMw6hGg/hKN2nuT9fnn1ZgAyh5ADczkEcgSHsA0Q\nhEYDQbDhwCL+ZPIOp5J9DucQKoJDsPjETaJDEHPgymPnS+eS7+bk8SZZ38SkM+pT6yEykob4fxfH\n50+oquEa0d6Ux/jWBlx93ILaGmpsr6raT8/cG3ect8/gWWRJsGJ2y+bT8Tqp9w9chQfGnGS8lrfJ\neG+1ioyiCYL/O86yaksiCYcgVFn1jL3k9282O40jzrtPH42HP30wAOCvL6/F+p6iV2eA5QPRTkdU\n6IqqQEATmQlCjonUefwB9bscQttbD+Jwax4aLc5umjiEjmbfmUl+JYFk6gZE7XiiknnIELti44So\nh1JZ+gR2mFAd2x3Wcr2Qsy08dsWh1TdQdPMpN1RHENqb89ht2ugqO0+3qsYrlaM4hHD8i+2Iu8cf\nhKWGa0QhIZdrpAhCjGGCKV1oreGvq0HQMS16DDnb8szF06bmCPQdEhetGg4BALYb04xpY5rxm6fe\nxVW/ec47X0QeKPO10SpzE/xeNoz8EAYD0RyCOjH7GPdDsCt9uLlwA1Zv5CxW2UADP33EXBw23495\nTwl2Y3pZE6yQF6+j4rIGZoJQ+wc0XHQI57tWF1NHp2dlY1FyCUKVHMKWhG92mvzFeCXrnMVrcDmE\n4P0NBYdQjQ5BWPykCRRo7NvEIRCwunsAP3joVeW87LM0sS1c9Lnv9p345+uqF/QA8iDXUzlX5Gvd\nVi8yAlEIQaggp+2dellB2V0L0ZGJQ2gq2Dhjnxn8gCHVxjfOzT6JyKhUERyC4WI9lMo1txDWcLqW\nj1s4GSuuO2ZwcjwXeejjISEIwju6LUHuBPjOWaEiI8P5JBxCNajVU7lYDq9vur9ad9zVIC50hc4x\n2BZ5u/V6+U2ozfCDa3/3vFJW6BA/umQ6/vP4cNHpPrOD/kaylZFd6kaFUVWeyiNKZEQIFxnZUJ26\n+pycspiK6wumjMF171kSaEN8hPIHUvPHR8k+uHIlSmRUDx3CoLEIg9RuFSi5nuqNtaXQrArtU4D3\n/xCYfXCi4mKXHPb05HX0ltP3dMtGc63VLuy1Whml5RC2hI4m2Kd2HMMx5GzLE+PWnOHN7bwxJJub\nDEGEjl80xfPLMWGf2Z2BcwPwdQh2aTM2ownVfJ8jikMIszKyJT8EgR5H5RByLkGYNaEdexjCO5iS\ncidxjYia31aYfZkGwZ4aoyvWRYcwSBgGCkIPIlvUUImMdvkg0GL2FNdhEXEnywRWRgfP5YGAfbPT\n6AkV9UZMNWvWIUQRBMMN1rrjrgYBAhDDMVjkL861EjBRvTEftLjSId5FHBHqHNWAeZPUjU+R5Xi0\nXwC5UndV4iJghBEEULgOIcAhsLwSWjpH/LoTEm7a8j64dJMgSofARUZpOARTI/W1MqoLjv0m0DJ+\neBEEgYYh4BBSgiDSYoZYrkWcq7vIqFYOwVUqG/UFRpFR/JypR8A9GXHT1MQxiG8yJPV68r7dv035\nJBwC3D7jO/3Gh9SAegMogNxNUa60Gd2sOvPrEUUQ8sVuNJkIAlUMHEJeiaSVE9cphCBUyRpG6hAS\nKpUFe2p0yqqLY1qdF+7dTwcufam+bdYLI0Kp7C4UIa/F7JiW3NAhDMaopHXiEH59zhLsoOXlNi1s\niTyt68zTBpXIcRwDJd6tx/YtREYyQQi5v2ff2QQgWQDAORNbMUXy8B5AzhMZyRzC/+z5a+DMPyQe\n74jSIeQH1qMxVIegTuweJ6eIWwQHUWLmpy1nOPNFRvFjipouFoWn+5MhREbGj3NYO6YNQxRGxZcZ\nYlhEntjIfD28br05hNp1CLyBgm0hr2mMzUrlRBShrgjqEOKPxSatXlZGMkEwcUAvrOzGN+7nm6yk\nREhONzuAvMQhdGOzyyGsb54FTNveWN+EEbXaEHOMIqMcHNikiow4h+C/hAbwGEd9FfPDVuKupLIy\nihIZJZPRRouMhqFYZjij1vjKWwCWyyKEK5UjrIxC5lOSzYtpIaqZQ3BFRnnb8jKZCRgJQsr5LJuD\nV49ojiDguSzpEOql85B1CGVDAMd1PQPe76R9yorqouSYliv1bBs6BEIFTRgInLfg4I21amC4XtdT\nWaDZrdfnhMhtDR9cEllm5KtLKzIaJMe0bQIf/hVwQGxa7mEBIs4FhHu/G86ljJNkgjkGf6JmQyGs\njEzRVKu1Mjp0Hlekn7TndvjBR/eobYBIYHYa0CHUzw9hoMTbaSpIi7fBmU/OA520zyCH4CqVyz6H\nkJbgjyiRkQhtrSOHiq8jcNHj5MAs23vVzeSGtCiZaaAXYA4pRUaRVkbJXmy5komMasaOR/B/IwBC\noRw2O6KC26XFOQfOxsS2cHv0eukQ8jYFMrKZ5n+cOGTFdcfgfx57A797emVN45IRZAjiOQYRhr1W\nR7o+N8eJrFQ2WWbJjyUphyAnPOKeyq5jWmmzxyGkfb0jiiCAOWhGf+C0BQeWRBD6WR6cJ/dfQotb\nr9eQkhlQ8xeksjKKEhlRQpGRk4mMtiVYluAQQq5H6RBSfuCXHTXXr2tqL11zAXgcgp2MQxgSs9M4\nJbJeHj63XqsjnUh61ZCP5hDkoI6JdQgSV+FxCMVe2E7Ry0/Ra0q6FYERtf0kVkE7egLnc1CtjERQ\nJ0faXXscQghBML34Wk3yKGEbgkMwWhllHMJWB65QDucRTLtSsUbUI8OZjHrFMsrbFnKa/saUwznJ\ndJ7U3ggAmDUufYIXE2KVyAaCIb7FLcUhlCQikVRkJOsliiwHYhVgDc8cuYJNQM4iPP9uV6rxjqzV\nhjGMpeANWppSWRAEBplDGECJ2egPCcYVlWO5Wlj1CG43nLyBM9QFlMIxTa4DhO/oEymV66RDuPfp\nd7G6m2+wZB2CrlSWZdwCSTiEg+aMxy/OXoxl+81KPzgDgiIhXWegXQbVzVN5RicnantJCXZM3t0l\nmUNILDLi69ussS1oaXaVyCufAgC8xKZi56ntePqtTanEgiOLIABopgG86YxTztlgCodQFByCJSuV\n+1GBhYEQgjAYrCwXGfnHq7r6MeOy3wbKiclnpgcZQdjaYEVYGAFhSmWOuG877WxJq0PoK1Zw7m3/\nwkd/9BgA3+yUcwhq782FoEQ66QK7eFZnzQpdAZ0TC2ZQCx5XPMe02saw18wxeOjSg/HB3ad650wi\nI5VDSNa20CE05m10drgOmb//LABgBZuIhVM7sK6niJVdQTF7GEYcQQCAn1QOx5dKJ3vHOVKVyiWL\nK9GYNBFaqB9l2OZE9jDvympnzkn5gJ97x8y+RYqM6sghTOmoznsxQ31hEcGywv0QjOeFFVxNs9Jk\ndpquhYEy/37e3shjR/kiIwos9s2FoIXcUOxv4nUGQZFRvfwQAGBaZ7MigSgazE5lriExh5D380xb\nIlNgqQcMFkrIeaK3dZuDpvphGFlKZReb0II2SZegK5XLLkF4Z2MfZrrnBIfQXw4hCDJpjLH5Tgo+\nl/w2wiZXtNlpfb6gO87bB9PGDELY6QypYVF06AqjyChFSPY0SKtD6C+poSqKFcdLviN0CPvM7sRV\nx4xgyesAABW/SURBVC1Q5OYCQ6NU1o+jx0CguukQTDBxCLJeQdfFhEH4IRABVsHf7D24+AfAn3nS\nKQDYPBCiODVgRHIIXawZjjR0bnbqL/RNLTwJzOuSb0ILOIcg7IJ1qMHt6sSqkpoFLexjEI4qZVOg\nsDoplXebNhpjR9WWbzhDfSBIQTodQu39mkNXxNd7cVU3Tr/lMfSXKp6SVMzlUtlBwVZzg5+4x3bY\ncUKrYnsvMDQJcqJ1BkbHtDpZGZlg0iHIzmqJRUYuh2ARwc773/abrTzO0XjX3LhnqycIaEFFGvr2\n9A7y8G+60DoWx+wyCU7FJxLNNIAybIxqDAtuV3+RkZ4qL+xjKHkJckxEIdMhbG3gCuUIc+WIc7Vy\nrTqScAhfuOsZ/PmFNfjX6xs8kavgdksVB3lXli2sgka3cB2eWWQ0BAQhMAbzcc4ivPjFowDUz8rI\nBKMOQRYZpTQ7rTgMuQafQ9hc4vXHjeIiozQcwogUGXWxFsWkdJ71BuZZb3jHrLEDrQ05VCSC0IJ+\ndFtt+MkZexnblKmyNwdq/Pb0jGlh8t+KtDsoVhz1pWRK5a0OXvjrkOtGpbInxjTXqXaqJrEyEjGK\nSg4LcAhcZMSvX3Dw9li4XQcO3JEbfcihFYYS8Z7JHLZFKOQshWgNBkdjtDKSziUlQkKpXCw7yBUa\nvfN9xTKIgM5RnDD3FpP7IoxIgrAJqshIBzWNRktDju+23WJNVESlsQHjQ+TopnwItYJHTfSPTTFM\nANUpZWNvSYtCkhGErQ2+Y5r53Ro9lYUOIa7xiOkSEgUprkXPeqhccTwOwdMhlJknMsrZFg6eM96r\nVy8roVoREAHH6BSq8RpOAzOHwN/Dz8/aWwmEFwVRrru/jHzBXzV6ihW0FHKeNGTrFRm5L6eLNeMD\nu08LLWY3t2NUQw6MqQ+eheRCAMwvvh5x2eU2yo5ZfyHvDt7d1IefjvkP/2LGIWx1iAtdYUIch1At\nknAItuXvRH2REb9Wqjgo5Ib3MlJNtFO/7iAQhAgOYdF2HYnbEUYis8a1oNDocwi9xTKaCjZaXLPf\nrVap7HTOxX+VTkIXRmG3GcE0cgK5Jk4QLG1BjyIICodQp4/PsqBswEohHIJsbvrOxn48MOpYvGLP\nVgeTYasBEUWGrjDWcf/Wskkx6R+S6BDyrsNZ90DZtzJyB//G+l60hejlhiuIVFFQlBXSYIiMonQI\nSS2MAJ5N7y+fORjf+fBuKDRIHMJABS0FG7ZFaMxbqTiEkfUm8w24qXIcAIAiEsc0t3ZglJNTTFEB\nKBnUdCg6hHpZGWkZ08Lyz5YqDK2NOXT3l/Hupj4eYI/celnoiq0OQlyUavcpxdoyoVplc5JqOVck\ntLm/7BECyyK8tKob/35zI644el5VfW8p6ESPQCjYFr770d3QPVAO1SkAg29lxBgDEXm+CfmUKdqm\njuaEoEHhECqeU+Cohhw2D2y1OgTpYbmB65Y72+Gn+Q/iS5Wve5caWjrQUsxBT0+TlEPwytc6WlLb\n0HUIq7r6sfeXHgAATO9sBmOcQ2CMSWPPOIStDVHJccIQZ+fgz5bwhk11k3AIQofQ3V9WTB3/9cYG\nAMDhC2rPWfCbC/bDP19fX3M7JuhiMSJusinyVevOqvK7GQwrox0ntGL5ym5vbDZx/Yzw56gGjU0+\nh9BbLKOlga+PLQ25rVeHoDwql0PYhBZ85OgD1YINrWg1iIzKDWMQBhMLWbPIiEgRB+kcgpgUAP/o\nJrU34p2NfWAM/tgzkdFWh7TiIkC2fIuelGnbTTLHhcx780AJfUXfymhVF88xMrG9MbRuGPTAdTtP\nbcfp+84MKV0bHI0iEMFThJsxuCKjn5+12LPEEnrFUsVJJS7S0dDIzU4dyqFH4hBaClszQVC0PW4c\nj5yFWdttpxZs7kRLQ1BkVGoMJwhy2/Vag6eObsKra3yPal2HIMsSc5aFBZPb8NiK9ZrSKSMIWxu4\nDiHde01sZRQBc3A79WR/qYKlX/sz/vryWu/cgLuD7u4ve7HALIuwsqsfY1oKShjmJHj40wfjrvP3\nTTn6+oFAxsB73vVB5hDGtBSwz2yuAxUbxlKFpRYXyWhyOYQ+x8aKtT2eDwgXGW2lBEGBSxAWTu1A\nQ6ukYH7PDcCEBWhpsAMcQqkhXBE9GFZGu27XgeUruzwKrVsZKQTBJhy50yRs7C3hsdfW+yKjjEPY\n6hAX3M6EenCtQs8gTym9vZdXb8Yra3pwzT3PeeeE7wFXKvvildVd/ZjQlp472G5MM1ob86nrVYuA\nDiGGQ1B1CIPz/Yl2yx5BqM1aq9klCK+wydjUV/I5hAYbPcVtgSDISuWGNv/3HmcCRGhtyHuLaonx\nsqWmcIKg7gSiFXgAsHTu+PCLLnadNhoOA558cyOAoHWB/KJytoUDdxznyWsfbdyPXxhVj5yyGYYT\nLI1DOGbnSYEyh82fgK+fuNA7rsbKaFRDvFf+J3/9pDcvn3prI9Z0czFQW5O/YAvLos39ZU9kVKo4\nWNU1gAkR2diA+uU0qAUdTYXAOZmrEd/c8YsmA/AJZmPewsyxgzN+0adwSi1JDn7VYFT7GDy25zdw\n1airAMDTIUwb04xXVgdzyISOq+oRDDW8bGjMGPyjpcH2CIJHGCJ0CFaIGdq3Tt7VW9Bl3PzRPUKt\nhgT2nDkGhZyFO594G794/E0v2NSe/d9BCTlc0OenA81ZhKaCjfmT2/DUW5twR+sp+OD51wJNoyP7\nyDDyQARvhX/qqsPRbHBE0nMJn7J4Oh59bT1O38csZ99/h7EAgFMXTwcA/ONzhwZ2nNe+b2d86XfP\nY9n+sxQOoKu/hGLZwXE3/tXzfm1rzOGlVd345gMvYWMvj5bZ3V/ygkMWyw429ZUwf1IbovDbC/f3\nOIyhQntzHs9cfQTWdg/AYQw//tsKdDT7RCJnW/j3Fw7zCKgQzy2Z1ZnYSSwtbHfxL0siIz2fRFrs\ndcwZeE/7a/jXPc9hXQ9/ZwfOGYefPPJ64jZGLkEQHELINr6jueCJjARBKDemW1wZgOMWTsZxCycH\nrtkWwY4wfQX4Du2AHcbh1/98yzsGgDXg4+jSCALAxUxPvbUJDYV8Rgy2Uhy3cDJ2m8bDR7clFJ2M\naSng1mV7h16f1N6EFdcd4x2bAhkeu3AyjnXn8il7T8Pcz/8eAM9xILhVoSNobczj76+uwz1PvevV\nf3tjH2aOHQWAW8gVK06sQrmpYBuD3G1pjGrIed/ffx6/U+C6TCD+8Tq3njpiwcRBG4/HIbgEoVgj\nhyBw4JxxwD1AryumXjJrbCpR1MglCBQ9yWyLYBOf3J781Y6Xd35gt6l4/65T8PNH38C+s8NFTElx\n1v4zcf/zqwBwFnTzgH9txbpe77eYDCfuuR1Wdw/gksN2rLnvDMMTR+4UFBFtaTTmbdz44V1xwc+f\nwAnf+5uX2UuAgMDOflXXANa6E1gQjmN2qf5efnzGnsMyJLtQyB6/aMqg9eHrEPhzLFecGMunZJg9\nbhS+csIu2G97zjE2FWzced4+2OnaZPVHMEHQ2Kv3fT9QRHAIT7fsg4U9f8WoSdtHNrn8miNRsC1Y\nFim7rVqw96xO3H7uEnzgpkewSeIIAOA5Kd+psDBYMLkdN526e136zpAhCiJfwaquAc+EVKC3WEFf\n0ReJjm9twOruATz7zibv3PxJbdhxQmvV/R80J14PNxS4bdne6O4vDypnI4xYKnUUGQmcuIdqdblg\ncnviuiOXIFiSDgEAFp4UKNJcsIAy0LH0E1jR+R3MmB5t5zxY8sJF23HRj252+vLqzd5vmWXNkGFL\nwJTARqCnWFY4hJ2mtOOPy1djrZR9a1zr1plfY7stwLWIxf+1tT1oLuRqVirXC0M/gmoRIzICgGm7\nHgoAmDpt+1hiMJgQMUUEzjlwNnafruoHOpq3nBlehgwA0BixA+4rVtAnWcHtMH5UoMyYlmwTUy2E\nyOj0Wx7Hki8/MGwIwlbAIYSj7airgMWnw+6cNfjjiUFLIYf+UhFjWgq47Ki5AIDvPfgKrrt3OQBg\ndMYhZNjCiOIQeosVhUNoa8qjIWd5ugPAT9GYIT1yklVj2WEoVVjk+9hSGBKSRESLiOjvRPRvIvoH\nEZmz1kQ24g49ylnAsoHO2VWOsr5odu2CZW/Ecw6cjcWzuCns6IxDyLCFYcpoJvDcu1341T/ewtTR\nTdhrxhgsnNqBVi2qqbB1z5AetmYq/8/XNyhhboYKQ8UhfAXA1Yyxe4noaPf4oHRNjCwPXhGbXI9X\nIpx+Mh1Chi2NJDvS9qY8fnXOEgDcFFXWIbSEOL5liEfO4AFdTUyoemOohFYMgPBoaQfwTm1NDX8I\niwU9XokIBZCJjDJsaUTpEARkoqFzBGGe0BniYQqJceWx84dgJCqG6o1+AsB9RHQ9OFHaJ3ULg5VC\napAgOARdcSTCBmRK5QxbGkk4BNn0UuduxZzOkB46h7D9+FHDQkowaG+UiO4HYHL1uwLAUgAXM8Zu\nJ6ITAfwIwKEh7ZwN4GwAmDZt2ggTFPkQ8tqcrYuMOIeQKegybGkksWqRy+iuP5mVUfXQOYThoFAG\nBpEgMMaMCzwAENFPAVzkHv4awA8j2rkZwM0AsMcee7C1YQWHOYS8VRcZHbtoMr7/4KvZx5VhWEIO\nyCiC4l157HxUHObF9M+QHroTWi2hr+uJodIhvANAZLU5BMBLaSqPHVVAfA6p4QVfh6A+8k8fMRdP\nXXV4pqDLMGzwlQ/sghZ3vg6UfdNTsamd1N6IZfvPUgJCZkgH3cpIlxwMFYZqFToLwDeJKAegH65I\nKAmWX3MkZ11XPsFPjBgdgisy0j4i26LEAc4yZNgSmD+5Dd85ZTecfsvjCocg0js2Z7qDmqGvA/WI\nY1QPDMmbZYz9BUBVAXsGK7zEYEMkBIkLmZ0hw1AjZ5On05Id0cQSNlhJY7YldPWrcc22dZHRNgcR\nquJfbwRzK2TIMNRoc53O9t9hLGaNHeWZlMrpXIUOQc9AliE9dp7SjjkTWnHbsr0xvbMZFw+T6MYj\nmPcbWTqEPWZkuQ0yDD/cfcG+sC3CDuNbQeTruNpdM2g5htHHD5yFR15dF5sUJ0M8WhvzuO/iAwAA\nD1568BCPxsfIJQgillGuaWjHkRANORvXf3Ahxm+lESIzjEzsMrXDeH58ayN+vmxv7DzVD5180Jzx\ndQsLn2F4YuQShEkLgQM/A+x++lCPJDFO2H3qUA8hQ4bE2MdNspJh28HIJQhEwMGfHepRZMiQIcNW\ng0ypnCFDhgwZAGQEIUOGDBkyuMgIQoYMGTJkAJARhAwZMmTI4CIjCBkyZMiQAUBGEDJkyJAhg4uM\nIGTIkCFDBgAZQciQIUOGDC6IjaBAVUTUDeCFoR7HMMFYACM1X1C9kT0LH9mz8JE9Cx/TGWOxGY1G\nmqfyC4yxPYZ6EMMBRPSP7FlwZM/CR/YsfGTPIj0ykVGGDBkyZACQEYQMGTJkyOBipBGEm4d6AMMI\n2bPwkT0LH9mz8JE9i5QYUUrlDBkyZMgweBhpHEKGDBkyZBgkjAiCQERHEtELRPQyEV021OPZEiCi\n/yai1UT0jHRuDBH9HxG95P4d7Z4nIvqW+3yeIqLdhm7k9QURbUdEfyKi54joWSK6yD2/zT0LACCi\nRiJ6jIiedJ/H1e75mUT0qHvfvySignu+wT1+2b0+YyjHX28QkU1ETxDRPe7xNvkc6oVhTxCIyAbw\nHQBHAZgP4GQimj+0o9oi+DGAI7VzlwF4gDG2A4AH3GOAP5sd3H9nA7hpC41xS6AM4JOMsfkAFgM4\n333/2+KzAIABAIcwxhYCWATgSCJaDOC/ANzAGNsewAYAH3PLfwzABvf8DW65rQkXAXheOt5Wn0N9\nwBgb1v8ALAFwn3R8OYDLh3pcW+jeZwB4Rjp+AcAk9/ckcL8MAPg+gJNN5ba2fwDuAnBY9iwYADQD\n+BeAvcEdsHLuee+bAXAfgCXu75xbjoZ67HW6/6ngm4FDANwDgLbF51DPf8OeQwAwBcCb0vFb7rlt\nERMYY++6v1cCmOD+3iaekcvm7wrgUWzDz8IVk/wbwGoA/wfgFQAbGWNlt4h8z97zcK9vAtC5ZUc8\naPgGgE8DcNzjTmybz6FuGAkEIYMBjG91thkTMSIaBeB2AJ9gjHXJ17a1Z8EYqzDGFoHvkPcCMHeI\nh7TFQUTvAbCaMfbPoR7L1oSRQBDeBrCddDzVPbctYhURTQIA9+9q9/xW/YyIKA9ODG5jjN3hnt4m\nn4UMxthGAH8CF410EJEIRSPfs/c83OvtANZt4aEOBvYFcBwRrQDwC3Cx0Tex7T2HumIkEITHAezg\nWg8UAJwE4O4hHtNQ4W4Ap7m/TwOXp4vzH3UtbBYD2CSJU0Y0iIgA/AjA84yxr0uXtrlnAQBENI6I\nOtzfTeD6lOfBCcMJbjH9eYjndAKAP7oc1YgGY+xyxthUxtgM8DXhj4yxU7CNPYe6Y6iVGEn+ATga\nwIvgstIrhno8W+ie/wfAuwBK4LLQj4HLPB8A8BKA+wGMccsSuCXWKwCeBrDHUI+/js9hP3Bx0FMA\n/u3+O3pbfBbu/e0C4An3eTwD4Avu+VkAHgPwMoBfA2hwzze6xy+712cN9T0MwjM5CMA92/pzqMe/\nzFM5Q4YMGTIAGBkiowwZMmTIsAWQEYQMGTJkyAAgIwgZMmTIkMFFRhAyZMiQIQOAjCBkyJAhQwYX\nGUHIkCEhiOg/iejQOrSzuR7jyZCh3sjMTjNk2MIgos2MsVFDPY4MGXRkHEKGbRpEdKqbX+DfRPR9\nN3DcZiK6wc038AARjXPL/piITnB/X+fmaHiKiK53z80goj+65x4gomnu+ZlE9AgRPU1EX9T6v5SI\nHnfrXL2l7z9DBhkZQciwzYKI5gH4EIB9GQ8WVwFwCoAWAP9gjC0A8CCAK7V6nQDeB2ABY2wXAGKR\n/zaAn7jnbgPwLff8NwHcxBjbGdz7XLRzOHjehr3AcxvsTkQHDMa9ZsiQBBlByLAtYymA3QE87oaT\nXgoe+sAB8Eu3zK3g4TNkbALQD+BHRPR+AL3u+SUAfu7+/plUb1/wUCTivMDh7r8nwPMazAUnEBky\nDAly8UUyZNhqQeA7+suVk0Sf18opijbGWJmI9gInICcAuAA82mYUTMo6AvBlxtj3U406Q4ZBQsYh\nZNiW8QCAE4hoPODlaZ4O/l2IiJkfBvAXuZKbm6GdMfY7ABcDWOhe+ht45E2Ai54edn//VTsvcB+A\nM932QERTxFgyZBgKZBxChm0WjLHniOhzAP5ARBZ4ZNnzAfQA2Mu9thpczyCjFcBdRNQIvsu/xD1/\nIYBbiOhSAGsAnOGevwjAz4noM/DDMYMx9gdXj/EIj/KNzQBOhZ/bIUOGLYrM7DRDBg2ZWWiGbRWZ\nyCjD/2+/jmkAAAAYhOFf9Z7JaE0QACqHAMA5BAAqQQDgBAGAShAAOEEAoBIEAG6jqbZkRezDvwAA\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b23e390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "reward_result.plot()\n",
    "plt.xlabel('episode')\n",
    "plt.ylabel('reward')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZcAAAEKCAYAAADenhiQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlFLK6TRclFJKOd3/A9rNDwAngQFCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x124d7fe48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "reward_result.cumsum().plot()\n",
    "plt.xlabel('episode')\n",
    "plt.ylabel('reward')\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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}