DQN_Improved_version_mountaincar.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import Library"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('/Users/Jacklin/anaconda3/lib/python3.6/site-packages/')\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import gym\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot Steps in Episode Comparison\n",
    "在mountaincar遊戲中step越少代表越快破關"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhMFgMBjiYgTCYDAYDHExAmEwGAyGuPw3QZF0261YapcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x127917048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "step_result = pd.concat([pd.read_csv('ref/DQN_mountaincar.csv'),\n",
    "                         pd.read_csv('ref/prioritized_DQN_mountaincar.csv'),\n",
    "                         pd.read_csv('ref/dueling_DQN_mountaincar.csv')],\n",
    "                        axis = 1)\n",
    "step_result.plot()\n",
    "plt.xlabel('Episode')\n",
    "plt.ylabel('Steps')\n",
    "plt.yscale(\"log\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CDF Plot Steps in Episode Comparison\n",
    "在mountaincar遊戲中step越少代表越快破關"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ndZkgJEldTwADlhDkRelm9wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x127afe208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "step_result.iloc[1:,:].cumsum().plot()\n",
    "plt.xlabel('Episode')\n",
    "plt.ylabel('CDF Steps')\n",
    "plt.yscale(\"log\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prioritized DQN Memory所需要的class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class SumTree(object):\n",
    "    \"\"\"\n",
    "    This SumTree code is modified version and the original code is from: \n",
    "    https://github.com/jaara/AI-blog/blob/master/SumTree.py\n",
    "    Story the data with it priority in tree and data frameworks.\n",
    "    \"\"\"\n",
    "    data_pointer = 0\n",
    "\n",
    "    def __init__(self, capacity):\n",
    "        self.capacity = capacity  # for all priority values\n",
    "        self.tree = np.zeros(2 * capacity - 1)\n",
    "        # [--------------Parent nodes-------------][-------leaves to recode priority-------]\n",
    "        #             size: capacity - 1                       size: capacity\n",
    "        self.data = np.zeros(capacity, dtype=object)  # for all transitions\n",
    "        # [--------------data frame-------------]\n",
    "        #             size: capacity\n",
    "\n",
    "    def add(self, p, data):\n",
    "        tree_idx = self.data_pointer + self.capacity - 1\n",
    "        self.data[self.data_pointer] = data  # update data_frame\n",
    "        self.update(tree_idx, p)  # update tree_frame\n",
    "\n",
    "        self.data_pointer += 1\n",
    "        if self.data_pointer >= self.capacity:  # replace when exceed the capacity\n",
    "            self.data_pointer = 0\n",
    "\n",
    "    def update(self, tree_idx, p):\n",
    "        change = p - self.tree[tree_idx]\n",
    "        self.tree[tree_idx] = p\n",
    "        # then propagate the change through tree\n",
    "        while tree_idx != 0:    # this method is faster than the recursive loop in the reference code\n",
    "            tree_idx = (tree_idx - 1) // 2\n",
    "            self.tree[tree_idx] += change\n",
    "\n",
    "    def get_leaf(self, v):\n",
    "        \"\"\"\n",
    "        Tree structure and array storage:\n",
    "        Tree index:\n",
    "             0         -> storing priority sum\n",
    "            / \\\n",
    "          1     2\n",
    "         / \\   / \\\n",
    "        3   4 5   6    -> storing priority for transitions\n",
    "        Array type for storing:\n",
    "        [0,1,2,3,4,5,6]\n",
    "        \"\"\"\n",
    "        parent_idx = 0\n",
    "        while True:     # the while loop is faster than the method in the reference code\n",
    "            cl_idx = 2 * parent_idx + 1         # this leaf's left and right kids\n",
    "            cr_idx = cl_idx + 1\n",
    "            if cl_idx >= len(self.tree):        # reach bottom, end search\n",
    "                leaf_idx = parent_idx\n",
    "                break\n",
    "            else:       # downward search, always search for a higher priority node\n",
    "                if v <= self.tree[cl_idx]:\n",
    "                    parent_idx = cl_idx\n",
    "                else:\n",
    "                    v -= self.tree[cl_idx]\n",
    "                    parent_idx = cr_idx\n",
    "\n",
    "        data_idx = leaf_idx - self.capacity + 1\n",
    "        return leaf_idx, self.tree[leaf_idx], self.data[data_idx]\n",
    "\n",
    "    @property\n",
    "    def total_p(self):\n",
    "        return self.tree[0]  # the root\n",
    "\n",
    "\n",
    "class Memory(object):  # stored as ( s, s_, r, a ) in SumTree\n",
    "    \"\"\"\n",
    "    This SumTree code is modified version and the original code is from:\n",
    "    https://github.com/jaara/AI-blog/blob/master/Seaquest-DDQN-PER.py\n",
    "    \"\"\"\n",
    "    epsilon = 0.01  # small amount to avoid zero priority\n",
    "    alpha = 0.6  # [0~1] convert the importance of TD error to priority\n",
    "    beta = 0.4  # importance-sampling, from initial value increasing to 1\n",
    "    beta_increment_per_sampling = 0.001\n",
    "    abs_err_upper = 1.  # clipped abs error\n",
    "\n",
    "    def __init__(self, capacity):\n",
    "        self.tree = SumTree(capacity)\n",
    "\n",
    "    def store(self, transition):\n",
    "        max_p = np.max(self.tree.tree[-self.tree.capacity:])\n",
    "        if max_p == 0:\n",
    "            max_p = self.abs_err_upper\n",
    "        self.tree.add(max_p, transition)   # set the max p for new p\n",
    "\n",
    "    def sample(self, n):\n",
    "        b_idx, b_memory, ISWeights = np.empty((n,), dtype=np.int32), np.empty((n, self.tree.data[0].size)), np.empty((n, 1))\n",
    "        pri_seg = self.tree.total_p / n       # priority segment\n",
    "        self.beta = np.min([1., self.beta + self.beta_increment_per_sampling])  # max = 1\n",
    "\n",
    "        min_prob = np.min(self.tree.tree[-self.tree.capacity:]) / self.tree.total_p     # for later calculate ISweight\n",
    "        for i in range(n):\n",
    "            a, b = pri_seg * i, pri_seg * (i + 1)\n",
    "            v = np.random.uniform(a, b)\n",
    "            idx, p, data = self.tree.get_leaf(v)\n",
    "            prob = p / self.tree.total_p\n",
    "            ISWeights[i, 0] = np.power(prob/min_prob, -self.beta)\n",
    "            b_idx[i], b_memory[i, :] = idx, data\n",
    "        return b_idx, b_memory, ISWeights\n",
    "\n",
    "    def batch_update(self, tree_idx, abs_errors):\n",
    "        abs_errors += self.epsilon  # convert to abs and avoid 0\n",
    "        clipped_errors = np.minimum(abs_errors, self.abs_err_upper)\n",
    "        ps = np.power(clipped_errors, self.alpha)\n",
    "        for ti, p in zip(tree_idx, ps):\n",
    "            self.tree.update(ti, p)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 定義DeepQLearning\n",
    "請大家將class中function的內容完成。投影片上應該也有答案～"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class DeepQLearning:\n",
    "    \n",
    "    def __init__(\n",
    "        self, \n",
    "        n_actions, \n",
    "        n_states, \n",
    "        dueling,\n",
    "        prioritized,\n",
    "        gamma = 0.9, #遠見程度\n",
    "        epsilon = 0.9,  #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",
    "        learning_rate = 0.001, #神經網路的更新率\n",
    "        memory_size = 50, #####\n",
    "        batch_size = 32, #####\n",
    "        nueron_num = 10,\n",
    "        \n",
    "    ):\n",
    "        \n",
    "        self.dueling = dueling\n",
    "        self.prioritized = prioritized\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.lr = 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",
    "        if self.prioritized:\n",
    "            self.memory = Memory(capacity=self.memory_size)\n",
    "        else:\n",
    "            self.memory = np.zeros((self.memory_size, n_states * 2 + 2))\n",
    "        \"\"\"\n",
    "        memory用來記錄探索環境所蒐集到的資料用來更新神經網路，從公式Q_target = R(s,a) + max(Q(s_,a))*Gamma\n",
    "        得知我們需要存取的資訊有s(current state), s_(next state), r(reward), a(action) 因此總共需要 2*n_states\n",
    "        +2的dimension\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current state\n",
    "        self.state_input = tf.placeholder(shape = [None, self.n_states], \n",
    "                                          name = 'input',\n",
    "                                          dtype = tf.float32)\n",
    "        #q_target = R(s, action) + Q(s_)*Gamma \n",
    "        self.q_target = tf.placeholder(shape = [None, self.n_actions], \n",
    "                                       name = 'q_target',\n",
    "                                       dtype = tf.float32)\n",
    "        if self.prioritized:\n",
    "            self.ISWeights = tf.placeholder(shape=[None, 1], name='IS_weights', dtype=tf.float32)\n",
    "        #搭建神經網路\n",
    "        if self.dueling:\n",
    "            with tf.variable_scope('Q_table'):\n",
    "                self.q_eval = self.build_dueling_network(self.nueron_num, Trainable = True, scope = 'net_eval') \n",
    "                self.q_next = self.build_dueling_network(self.nueron_num, Trainable = False, scope = 'net_target') ##### 利用另外一個參數較舊的神經網路來給出下一步的q值並計算q_target\n",
    "        else:\n",
    "            with tf.variable_scope('Q_table'):\n",
    "                self.q_eval = self.build_network(self.nueron_num, Trainable = True, scope = 'net_eval') \n",
    "                self.q_next = self.build_network(self.nueron_num, Trainable = False, scope = 'net_target') ##### 利用另外一個參數較舊的神經網路來給出下一步的q值並計算q_target\n",
    "\n",
    "        #管理神經網路的parameters\n",
    "        self.Qnet_eval_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Q_table/net_eval')\n",
    "        self.Qnet_target_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Q_table/net_target') #####\n",
    "        #更新舊神經網路的參數\n",
    "        self.replace_net_params = [tf.assign(old, new) for old, new in zip(self.Qnet_target_params, self.Qnet_eval_params)]\n",
    "        \n",
    "        #計算q_target和q_eval的mse來更新神經網路的參數\n",
    "        if self.prioritized:\n",
    "            self.abs_errors = tf.reduce_sum(tf.abs(self.q_target - self.q_eval), axis=1)    # for updating Sumtree\n",
    "            self.loss = tf.reduce_mean(self.ISWeights * tf.squared_difference(self.q_target, self.q_eval))\n",
    "        else:\n",
    "            self.loss = tf.reduce_mean(tf.squared_difference(self.q_target, self.q_eval))\n",
    "        \n",
    "        self.train = tf.train.RMSPropOptimizer(self.lr).minimize(self.loss, var_list=self.Qnet_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",
    "        if self.prioritized:    # prioritized replay\n",
    "            env_info = np.concatenate([current_state, next_state, [reward], [action]]) \n",
    "            self.memory.store(env_info)\n",
    "        else:\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 build_network(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",
    "            x_h1 = tf.layers.dense(inputs = self.state_input, units = neuron_num, activation = tf.nn.tanh, kernel_initializer=init_w, bias_initializer=init_b, trainable=Trainable) \n",
    "            \n",
    "            x_h2 = tf.layers.dense(inputs = x_h1, units = neuron_num, activation = tf.nn.tanh, kernel_initializer=init_w, bias_initializer=init_b, trainable=Trainable)  \n",
    "            \n",
    "            output = tf.layers.dense(inputs = x_h2, units = self.n_actions, kernel_initializer=init_w, bias_initializer=init_b, trainable=Trainable)\n",
    "        \n",
    "        return output #輸出‘不同動作’對應的Q值 \n",
    "    \n",
    "    def build_dueling_network(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",
    "            h1 = tf.layers.dense(inputs = self.state_input, \n",
    "                                       units = neuron_num, \n",
    "                                       activation = tf.nn.tanh, \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.tanh, \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.tanh, \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 choose_action(self, current_state):\n",
    "        if np.random.uniform() < self.epsilon: \n",
    "            #選擇產生估計Q值較大的行動\n",
    "            q_eval = self.sess.run(self.q_eval, feed_dict={self.state_input: current_state[np.newaxis, :]})\n",
    "            self.action = np.argmax(q_eval)\n",
    "        else:\n",
    "            #採取隨機行動\n",
    "            self.action = np.random.randint(0, self.n_actions)\n",
    "          \n",
    "   \n",
    "        return self.action\n",
    "    \n",
    "    def learn(self, current_state, reward, next_state, update_params): #####\n",
    "        \n",
    "        if update_params: #####\n",
    "            self.sess.run(self.replace_net_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.prioritized:\n",
    "            tree_idx, past_experience, ISWeights = self.memory.sample(self.batch_size)\n",
    "        else:    \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].astype(int)\n",
    "        index = np.arange(self.batch_size)\n",
    "\n",
    "        #算出實際q值並用此更新神經網路參數\n",
    "        q_eval = self.sess.run(self.q_eval, feed_dict={self.state_input: current_state})\n",
    "        q_next = self.sess.run(self.q_next, feed_dict={self.state_input: next_state})\n",
    "        q_target = q_eval.copy()\n",
    "        q_target[index, action] = reward + self.gamma*np.max(q_next, axis=1)\n",
    "        if self.prioritized:\n",
    "            _, abs_errors, self.cost = self.sess.run([self.train, self.abs_errors, self.loss],\\\n",
    "                                         feed_dict={self.state_input: current_state,\\\n",
    "                                                    self.q_target: q_target,\\\n",
    "                                                    self.ISWeights: ISWeights})\n",
    "            self.memory.batch_update(tree_idx, abs_errors)   # update priority\n",
    "        else:    \n",
    "            _, self.cost = self.sess.run([self.train, self.loss], feed_dict={self.state_input: current_state,\\\n",
    "                                                                                self.q_target: q_target})        # increasing epsilon\n",
    "        self.epsilon = self.epsilon + self.epsilon_increase if self.epsilon < self.epsilon_max else self.epsilon_max  \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))\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define Training的環境和步驟"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def training(save_model, model_name):\n",
    "    step_record = []\n",
    "    reward_record = []\n",
    "    learn_count = 0\n",
    "    update_count = 0\n",
    "    for episode in range(100):\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",
    "            env.render()\n",
    "\n",
    "            # 根據現在的狀態選擇動作\n",
    "            action = RL.choose_action(current_state)\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",
    "            # 調整reward\n",
    "            \"\"\"\n",
    "            表現較差的方法，但是在一般問題中我們常常是不知道怎麼引導agent\n",
    "            達到目標的，因此這樣的方法相對直覺。\n",
    "            \"\"\"\n",
    "            if done == True:\n",
    "                reward = 100\n",
    "            else:\n",
    "                reward = -1\n",
    "            \n",
    "            total_reward+=reward\n",
    "            # 將資訊存至記憶體中以便進行experience replay\n",
    "            RL.write_memory(current_state, next_state, reward, action)\n",
    "\n",
    "            # 進行學習\n",
    "            if learn_count <= 1000:\n",
    "                learn_count+=1\n",
    "            else:\n",
    "                #進行參數更新\n",
    "                if update_count==300:\n",
    "                    update = True\n",
    "                    update_count =0\n",
    "                else:\n",
    "                    update = False\n",
    "                    update_count+=1\n",
    "                \n",
    "                RL.learn(current_state, reward, next_state, update)\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:{} epsilon:{}'.format(episode, step, \\\n",
    "                       total_reward, RL.epsilon))\n",
    "                step_record.append(step)\n",
    "                reward_record.append(total_reward)\n",
    "                break\n",
    "            current_state = next_state\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": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/Jacklin/anaconda3/envs/tensor_flow_py3p5/lib/python3.5/site-packages/ipykernel_launcher.py:100: RuntimeWarning: divide by zero encountered in double_scalars\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "episode:0 steps:9200 total reward:-9100 epsilon:0.95\n",
      "episode:1 steps:2709 total reward:-2609 epsilon:0.95\n",
      "episode:2 steps:306 total reward:-206 epsilon:0.95\n",
      "episode:3 steps:232 total reward:-132 epsilon:0.95\n",
      "episode:4 steps:661 total reward:-561 epsilon:0.95\n",
      "episode:5 steps:302 total reward:-202 epsilon:0.95\n",
      "episode:6 steps:505 total reward:-405 epsilon:0.95\n",
      "episode:7 steps:698 total reward:-598 epsilon:0.95\n",
      "episode:8 steps:489 total reward:-389 epsilon:0.95\n",
      "episode:9 steps:703 total reward:-603 epsilon:0.95\n",
      "episode:10 steps:935 total reward:-835 epsilon:0.95\n",
      "episode:11 steps:701 total reward:-601 epsilon:0.95\n",
      "episode:12 steps:539 total reward:-439 epsilon:0.95\n",
      "episode:13 steps:350 total reward:-250 epsilon:0.95\n",
      "episode:14 steps:383 total reward:-283 epsilon:0.95\n",
      "episode:15 steps:377 total reward:-277 epsilon:0.95\n",
      "episode:16 steps:455 total reward:-355 epsilon:0.95\n",
      "episode:17 steps:1224 total reward:-1124 epsilon:0.95\n",
      "episode:18 steps:2176 total reward:-2076 epsilon:0.95\n",
      "episode:19 steps:2254 total reward:-2154 epsilon:0.95\n",
      "episode:20 steps:3268 total reward:-3168 epsilon:0.95\n",
      "episode:21 steps:476 total reward:-376 epsilon:0.95\n",
      "episode:22 steps:2908 total reward:-2808 epsilon:0.95\n",
      "episode:23 steps:550 total reward:-450 epsilon:0.95\n",
      "episode:24 steps:2573 total reward:-2473 epsilon:0.95\n",
      "episode:25 steps:789 total reward:-689 epsilon:0.95\n",
      "episode:26 steps:334 total reward:-234 epsilon:0.95\n",
      "episode:27 steps:133 total reward:-33 epsilon:0.95\n",
      "episode:28 steps:199 total reward:-99 epsilon:0.95\n",
      "episode:29 steps:376 total reward:-276 epsilon:0.95\n",
      "episode:30 steps:707 total reward:-607 epsilon:0.95\n",
      "episode:31 steps:2321 total reward:-2221 epsilon:0.95\n",
      "episode:32 steps:203 total reward:-103 epsilon:0.95\n",
      "episode:33 steps:177 total reward:-77 epsilon:0.95\n",
      "episode:34 steps:289 total reward:-189 epsilon:0.95\n",
      "episode:35 steps:106 total reward:-6 epsilon:0.95\n",
      "episode:36 steps:239 total reward:-139 epsilon:0.95\n",
      "episode:37 steps:212 total reward:-112 epsilon:0.95\n",
      "episode:38 steps:186 total reward:-86 epsilon:0.95\n",
      "episode:39 steps:182 total reward:-82 epsilon:0.95\n",
      "episode:40 steps:180 total reward:-80 epsilon:0.95\n",
      "episode:41 steps:168 total reward:-68 epsilon:0.95\n",
      "episode:42 steps:179 total reward:-79 epsilon:0.95\n",
      "episode:43 steps:252 total reward:-152 epsilon:0.95\n",
      "episode:44 steps:179 total reward:-79 epsilon:0.95\n",
      "episode:45 steps:398 total reward:-298 epsilon:0.95\n",
      "episode:46 steps:469 total reward:-369 epsilon:0.95\n",
      "episode:47 steps:2223 total reward:-2123 epsilon:0.95\n",
      "episode:48 steps:781 total reward:-681 epsilon:0.95\n",
      "episode:49 steps:168 total reward:-68 epsilon:0.95\n",
      "episode:50 steps:219 total reward:-119 epsilon:0.95\n",
      "episode:51 steps:587 total reward:-487 epsilon:0.95\n",
      "episode:52 steps:163 total reward:-63 epsilon:0.95\n",
      "episode:53 steps:238 total reward:-138 epsilon:0.95\n",
      "episode:54 steps:2826 total reward:-2726 epsilon:0.95\n",
      "episode:55 steps:706 total reward:-606 epsilon:0.95\n",
      "episode:56 steps:313 total reward:-213 epsilon:0.95\n",
      "episode:57 steps:153 total reward:-53 epsilon:0.95\n",
      "episode:58 steps:146 total reward:-46 epsilon:0.95\n",
      "episode:59 steps:88 total reward:12 epsilon:0.95\n",
      "episode:60 steps:170 total reward:-70 epsilon:0.95\n",
      "episode:61 steps:89 total reward:11 epsilon:0.95\n",
      "episode:62 steps:94 total reward:6 epsilon:0.95\n",
      "episode:63 steps:204 total reward:-104 epsilon:0.95\n",
      "episode:64 steps:178 total reward:-78 epsilon:0.95\n",
      "episode:65 steps:146 total reward:-46 epsilon:0.95\n",
      "episode:66 steps:151 total reward:-51 epsilon:0.95\n",
      "episode:67 steps:234 total reward:-134 epsilon:0.95\n",
      "episode:68 steps:212 total reward:-112 epsilon:0.95\n",
      "episode:69 steps:135 total reward:-35 epsilon:0.95\n",
      "episode:70 steps:208 total reward:-108 epsilon:0.95\n",
      "episode:71 steps:1166 total reward:-1066 epsilon:0.95\n",
      "episode:72 steps:207 total reward:-107 epsilon:0.95\n",
      "episode:73 steps:123 total reward:-23 epsilon:0.95\n",
      "episode:74 steps:165 total reward:-65 epsilon:0.95\n",
      "episode:75 steps:126 total reward:-26 epsilon:0.95\n",
      "episode:76 steps:197 total reward:-97 epsilon:0.95\n",
      "episode:77 steps:216 total reward:-116 epsilon:0.95\n",
      "episode:78 steps:230 total reward:-130 epsilon:0.95\n",
      "episode:79 steps:308 total reward:-208 epsilon:0.95\n",
      "episode:80 steps:205 total reward:-105 epsilon:0.95\n",
      "episode:81 steps:158 total reward:-58 epsilon:0.95\n",
      "episode:82 steps:393 total reward:-293 epsilon:0.95\n",
      "episode:83 steps:337 total reward:-237 epsilon:0.95\n",
      "episode:84 steps:631 total reward:-531 epsilon:0.95\n",
      "episode:85 steps:257 total reward:-157 epsilon:0.95\n",
      "episode:86 steps:211 total reward:-111 epsilon:0.95\n",
      "episode:87 steps:141 total reward:-41 epsilon:0.95\n",
      "episode:88 steps:312 total reward:-212 epsilon:0.95\n",
      "episode:89 steps:273 total reward:-173 epsilon:0.95\n",
      "episode:90 steps:228 total reward:-128 epsilon:0.95\n",
      "episode:91 steps:339 total reward:-239 epsilon:0.95\n",
      "episode:92 steps:414 total reward:-314 epsilon:0.95\n",
      "episode:93 steps:280 total reward:-180 epsilon:0.95\n",
      "episode:94 steps:583 total reward:-483 epsilon:0.95\n",
      "episode:95 steps:194 total reward:-94 epsilon:0.95\n",
      "episode:96 steps:245 total reward:-145 epsilon:0.95\n",
      "episode:97 steps:2832 total reward:-2732 epsilon:0.95\n",
      "episode:98 steps:170 total reward:-70 epsilon:0.95\n",
      "episode:99 steps:153 total reward:-53 epsilon:0.95\n",
      "game over\n"
     ]
    }
   ],
   "source": [
    "step_result = []\n",
    "env = gym.make('MountainCar-v0')\n",
    "env = env.unwrapped\n",
    "env.seed(1)\n",
    "\n",
    "RL = DeepQLearning( n_actions = 3, \n",
    "                    n_states = 2,\n",
    "                    gamma = 0.99,\n",
    "                    epsilon = 0.95,\n",
    "                    epsilon_increase = 0.001,\n",
    "                    learning_rate = 0.01,\n",
    "                    memory_size = 10000,\n",
    "                    batch_size = 32,\n",
    "                    nueron_num = 10,\n",
    "                    dueling = True,\n",
    "                    prioritized = True\n",
    "                      )\n",
    "step_record, reward_record = training(save_model = True, model_name='dueling_DQN_moutaincar')\n",
    "step_result.append(pd.DataFrame(step_record))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "step_result = pd.concat(step_result)\n",
    "step_result.columns = ['dueling_DQN']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3XeYXGXZ+PHvvTOzs71kS3pPCCQh\nIYUSgqBEIKEIKgKKAoLCT1HUVxTUV0EFGwgCKi9R0AhIkSIQqVKkSCC9l92ElN2U7b3N7Dy/P845\ns7O9zWR3z96f68o1O2fOzJw9mZ373M/9FDHGoJRSSvVU3EAfgFJKqaFFA4dSSqle0cChlFKqVzRw\nKKWU6hUNHEoppXpFA4dSSqle0cChlFKqVzRwKKWU6hUNHEoppXrFO9AHEAvZ2dlm0qRJA30YSik1\npKxdu7bEGJPT3X6uDByTJk1izZo1A30YSik1pIjIvp7sp01VSimlekUDh1JKqV7RwKGUUqpXXFnj\nUEoNLoFAgIKCAhoaGgb6UBSQkJDAuHHj8Pl8fXq+Bg6lVMwVFBSQmprKpEmTEJGBPpxhzRhDaWkp\nBQUFTJ48uU+voU1VSqmYa2hoICsrS4PGICAiZGVl9Sv708ChlDoqNGgMHv39v3Bl4DhS1UBDoHmg\nD0MppVzJlYGjqLqR2sbgQB+GUkq5kisDB0Cg2Qz0ISilBqlbb72VO++8s9fP27t3L7NnzwZgzZo1\n3HDDDVE9rquuuorJkyczd+5cjjnmGK644goKCgrCj1dWVnLFFVcwbdo0pk6dyuWXX055eXn42ESE\n++67L7z/N77xDf76179G9RjB1YEjNNCHoJRysYULF3LvvfdG/XXvuOMONm7cyM6dO5k3bx5nnnkm\nTU1NAFxzzTVMmTKF/Px8du/ezbRp07jqqqvCz83NzeWee+4J7x8rru2O26SBQ6lB6acvbGXbwaqo\nvubMMWnccsGsLve5/fbbWbFiBbm5uYwfP54FCxbw8Y9/nDvvvJOFCxdSUlLCwoUL2bt3L83Nzdx8\n88289dZbNDY2cv3113Pddde1er233nqLO++8k5UrV3Lrrbeyf/9+9uzZw/79+/n2t78dzkZ+/vOf\n88gjj5CTkxN+3xtvvLHb30lE+M53vsOzzz7LSy+9xKxZs1i7di1PPPFEeJ+f/OQnTJ06lZ07d+L3\n+8nJyWHx4sWsWLGCr371q304kz3j2sChGYdSyrF27Voef/xxNmzYQDAYZP78+SxYsKDT/R988EHS\n09NZvXo1jY2NLF68mLPPPrvL3kg7duzgzTffpLq6mhkzZvC1r32NDRs28PTTT7Nx40YCgUC379uR\n+fPns2PHDkSEE044AY/HE37M4/Ewb948tm/fzgknnADATTfdxLJly7j66qt79T694drA0RTUwKHU\nYNRdZhAL77zzDp/+9KdJSkoC4FOf+lSX+7/66qts2rSJp556CrBqC3l5eRxzzDGdPue8887D7/fj\n9/vJzc3lyJEjvPfee1x44YUkJCSQkJDABRdc0OtjN6Z39dopU6Zw8skn8/e//73X79VTrg0cmnEo\npbrj9XoJhazvisgBccYY7rvvPs4555xW++/du7fT1/L7/eGfPR4PwWB0enauX7+eJUuWMHPmTDZs\n2EAoFCIuzipPh0IhNm7cyPz588O/B8APf/hDLr74Ys4444yoHENbri2ONwW1V5VSynL66afzz3/+\nk/r6eqqrq3nhhRcAa+2etWvXAoSzC4BzzjmH+++/n0AgAMCuXbuora3t9fsuXryYF154gYaGBmpq\nali5cmWPn2uM4d577+XQoUMsXbqUadOmMW/ePG677bbwPrfddhtLlixhwoQJrZ577LHHMnPmzPDv\nGW2uDRyacSilHPPnz+fSSy9l7ty5LFu2jBNPPBGAG2+8kfvvv5958+ZRUlIS3v8rX/kKM2fOZP78\n+cyePZvrrruuTxnEiSeeyKc+9SnmzJnDsmXLOP7440lPT+/yOd/73vfC3XFXr17Nm2++SXx8PAAP\nPfQQeXl5TJ06lZycHFatWsX//d//dfg6P/rRj1p15Y0m6W372VDgHz3dvPjGuyw5buRAH4pSCti+\nfTvHHXfcQB/GgKipqSElJYW6ujpOP/10li9fzvz58/v9ujt37uS8887j3nvv5dxzz+318zv6PxGR\ntcaYhd09V2scSikVQ9deey3btm2joaGBK6+8MipBA2DGjBnk5+dH5bV6y7WBo0lHjiulBoGOejdd\nf/31vPfee622fetb3+LLX/7y0TqsfnFv4NDuuEoNKsYYnSHX9oc//GFA37+/JQotjiulYi4hIYHS\n0tJ+f2Gp/nMWckpISOjza7g249DAodTgMW7cOAoKCiguLh7oQ1G0LB3bV64NHNpUpdTg4fP5+rxM\nqRp8XNxUpSmxUkrFgosDh2YcSikVCxo4lFJK9YorA4egNQ6llIoVdwYOEV3ISSmlYsSlgUObqpRS\nKlbcGTiAgE6rrpRSMRHTwCEi3xGRrSKyRUQeE5EEEZksIh+ISL6IPCEi8fa+fvt+vv34pIjX+YG9\nfaeInNPZ+zniRDTjUEqpGIlZ4BCRscANwEJjzGzAA1wG/Bq42xgzDSgHrrGfcg1Qbm+/294PEZlp\nP28WsBT4o4h46IIIWuNQSqkYiXVTlRdIFBEvkAQcAs4EnKW2VgAX2T9faN/HfnyJWDOiXQg8boxp\nNMZ8BOQDJ3X1poJoryqllIqRmAUOY0whcCewHytgVAJrgQpjjLOUVgEw1v55LHDAfm7Q3j8rcnsH\nzwkTkWtFZI2IrGluDmpTlVJKxUgsm6oysbKFycAYIBmrqSkmjDHLjTELjTELfT6vTjmilFIxEsum\nqk8CHxljio0xAeAZYDGQYTddAYwDCu2fC4HxAPbj6UBp5PYOntMhQcdxKKVUrMQycOwHThGRJLtW\nsQTYBrwJXGzvcyXwnP3z8/Z97MffMNbk/c8Dl9m9riYD04EPu3pjHcehlFKxE7Np1Y0xH4jIU8A6\nIAisB5YD/wIeF5Hb7G0P2k95EHhYRPKBMqyeVBhjtorIk1hBJwhcb4xp7uq9NXAopVTsxHQ9DmPM\nLcAtbTbvoYNeUcaYBuBznbzO7cDtPX1fQXQAoFJKxYg7R47rOA6llIoZ9wYOHcehlFIx4crAEYdO\nOaKUUrHiysChxXGllIodlwYO0QGASikVI+4MHGhxXCmlYsWdgcNuqrLGDyqllIomlwYOwRgIhjRw\nKKVUtLkzcNi3WiBXSqnoc2fgsCOHjh5XSqnoc2ngsCKHFsiVUir63Bk47FttqlJKqehzZeCIc5qq\nNHAopVTUuTJwOE1VGjiUUir63Bk47NtGnehQKaWizp2BI9xUpb2qlFIq2lwaOLSpSimlYsWdgcO+\nDWhTlVJKRZ07A4eO41BKqZhxaeCwbrXGoZRS0efOwIHWOJRSKlbcGTjsjEPXHVdKqehzZ+Cwb7XG\noZRS0efOwKHdcZVSKmZcGjisW+2Oq5RS0efKwBGnvaqUUipmXBk4nF5VWuNQSqnoc2fg0GnVlVIq\nZlwZOAC8caLdcZVSKgZcGzh8njjNOJRSKgZcHDhEi+NKKRUDrg0c8d44LY4rpVQMuDdweOJ0HIdS\nSsWAawOHz6s1DqWUioWYBg4RyRCRp0Rkh4hsF5FFIjJCRF4TkTz7NtPeV0TkXhHJF5FNIjI/4nWu\ntPfPE5Ere/LePo82VSmlVCzEOuO4B3jZGHMsMBfYDtwMvG6MmQ68bt8HWAZMt/9dC9wPICIjgFuA\nk4GTgFucYNMVnyeOpqAWx5VSKtpiFjhEJB04HXgQwBjTZIypAC4EVti7rQAusn++EPibsawCMkRk\nNHAO8JoxpswYUw68Bizt7v3jPaJNVUopFQOxzDgmA8XAX0RkvYj8WUSSgZHGmEP2PoeBkfbPY4ED\nEc8vsLd1tr1LOo5DKaViI5aBwwvMB+43xswDamlplgLAGGOAqLQnici1IrJGRNYUFxdr4FBKqRiJ\nZeAoAAqMMR/Y95/CCiRH7CYo7Nsi+/FCYHzE88fZ2zrb3ooxZrkxZqExZmFOTo49jkNrHEopFW0x\nCxzGmMPAARGZYW9aAmwDngecnlFXAs/ZPz8PXGH3rjoFqLSbtF4BzhaRTLsofra9rUs+HcehlFIx\n4Y3x638TeFRE4oE9wJexgtWTInINsA+4xN73ReBcIB+os/fFGFMmIj8HVtv7/cwYU9bdG8d7Rbvj\nKqVUDMQ0cBhjNgALO3hoSQf7GuD6Tl7nIeCh3ry31jiUUio23DtyXJuqlFIqJlwdOLQ4rpRS0efa\nwOHXuaqUUiomXBs4fDpyXCmlYsLFgUMzDqWUigWXBw5DKKR1DqWUiibXBo54r/WrBUKadSilVDS5\nNnD4PAKg644rpVSUuThw2BmHjuVQSqmocm3gCDdVaYFcKaWiyrWBw8k4dL4qpZSKLtcGjninqUpr\nHEopFVW9Dhz29OZzYnEw0RTOOLTGoZRSUdWjwCEib4lImoiMANYBfxKRu2J7aP3T0qtKA4dSSkVT\nTzOOdGNMFfAZ4G/GmJOBT8busPrP59Uah1JKxUJPA4fXXub1EmBlDI8nauK1O65SSsVETwPHz7CW\na91tjFktIlOAvNgdVv+1dMfV4rhSSkVTj1YANMb8A/hHxP09wGdjdVDREB4AqE1VSikVVT0tjk8R\nkRdEpFhEikTkOTvrGLSc4rjWOJRSKrp62lT1d+BJYDQwBiv7eCxWBxUN8dodVymlYqKngSPJGPOw\nMSZo/3sESIjlgfWXNlUppVRs9KjGAbwkIjcDjwMGuBR40R7XgTGmLEbH12c+natKKaVioqeB4xL7\n9ro22y/DCiSDrt7RUuPQXlVKKRVNPe1VNTnWBxJtfo8H0HEcSikVbT3tVZUkIv8rIsvt+9NF5PzY\nHlr/+Lw65YhSSsVCT4vjfwGagFPt+4XAbTE5oijRSQ6VUio2eho4phpjfgMEAIwxdYDE7KiiwBun\nGYdSSsVCTwNHk4gkYhXCEZGpQGPMjioKRIR4T5wWx5VSKsp62qvqVuBlYLyIPAosBr4cq4OKFp9H\nNONQSqko62mvqldFZC1wClYT1beMMSUxPbIo8HnjNHAopVSU9bRX1evGmFJjzL+MMSuNMSUi8nqs\nD66/4j0aOJRSKtq6zDhEJAFIArJFJJOWgngaMDbGx9ZvPk8cTUGtcSilVDR111R1HfBtrIkN10Zs\nrwZ+H6uDipZ4b5zOjquUUlHWXVPVf7HGbtxojJkC/BTYAvwHa8bcQc3nER05rpRSUdZd4HgAaDTG\n3CcipwO/BFYAlcDynryBiHhEZL2IrLTvTxaRD0QkX0SeEJF4e7vfvp9vPz4p4jV+YG/fKSLn9PSX\n82mNQymloq67wOGJmPn2UmC5MeZpY8yPgWk9fI9vAdsj7v8auNsYMw0oB66xt18DlNvb77b3Q0Rm\nYk2mOAtYCvxRRDw9eWOfR5uqlFIq2roNHCLi1EGWAG9EPNZtV14RGQecB/zZvi/AmcBT9i4rgIvs\nny+072M/vsTe/0LgcWNMozHmIyAfOKm79warxqEZh1JKRVd3X/6PAf8RkRKgHngHQESmYTVXded3\nwPeBVPt+FlBhjAna9wto6Z01FjgAYIwJikilvf9YYFXEa0Y+p0vxnjjqA8092VUppVQPdRk4jDG3\n2+M1RgOvGmOcvq1xwDe7eq49e26RMWatiHw8GgfbzftdC1wLMGHCBMAqjlc1aMahlFLR1G1zkzFm\nVQfbdvXgtRcDnxKRc7GWmU0D7gEyRMRrZx3jsGbaxb4dDxTYzWPpQGnEdkfkcyKPaTl2wX7hwoUG\nnHEcGjiUUiqaejrJYa8ZY35gjBlnjJmEVdx+wxhzOfAmcLG925XAc/bPz9v3sR9/w85wngcus3td\nTQamAx/25Bh8Oo5DKaWirqeTHEbTTcDjInIbsB540N7+IPCwiOQDZVjBBmPMVhF5EtgGBIHrjTE9\nKlzolCNKKRV9RyVwGGPeAt6yf95DB72ijDENwOc6ef7twO29fV9rAKBOOaKUUtEUs6aqwUC74yql\nVPS5OnDoAECllIo+VwcOrXEopVT0uTpwaHdcpZSKPtcHjpCB5pAWyJVSKlrcHTi81rpT2lyllFLR\n4+rAEe+xfj0tkCulVPS4O3B4rV9PF3NSSqnocXXg8NkZR6BZaxxKKRUtwyRwaMahlFLR4vLAYRXH\nG4dxU9X9b+3m0Q/2DfRhKKVcxNWBI14zDlb8dy8rNx4a6MNQSrmIqwPHcG+qqm0McriqgerGwEAf\nilLKRdwdOLzDO3B8VFILQE1DsJs9lVKq51wdOMLjOIbp1Oq7i2sAqGnUwKGUih53B45hPnJ8d7GV\ncVRpxqGUiiJXB47hXuPYY2ccTcEQjcEeLZqolFLdGhaBY7jOkOtkHAC1jRo4lFLRMTwCxzDMOEIh\nw0clNWQk+YD2BfLmkGHp797m5S3aVVcp1TuuDhzxw3jKkUNVDTQEQswdlwFAVUPrLrmV9QF2HK5m\nY0HlQBye6oPKuoB2dFD90hQMsW5/eb9fx92BYxh3x91dZNU35o5LB9r3rKqqtwJJZb2O8Rgqrntk\nDT/555aBPgw1hL2w8SCfvf+/FFU39Ot1vFE6nkHJmXJkOAYOpzA+d7yVcbRtqqrUwDHkFJTXExyG\n2bOKnuKaRoyBstomclMT+vw67g4c3uFbHN9TUkuq38vk7GSAdqPHnaarKg0cQ0Z1QxC/t2mgD0MN\nYdX23311P7vouzpwJPo8ANQ1Db8eRbuLa5iSk0xqQsfFcSfT0MAxNBhjqG4I4I2TgT4UNYQ5AaO6\noX9/966ucfg8cSTFe4bll+Oe4lqm5qSQmmBdG1S3q3FY97WpamiobWomZKCiPoAx2lyl+sb5Luxv\nxuHqwAGQluAbdl+OtY1BDlU2MCUnGb83Dp9H2mUcTlNVR+emsi7Ajf/Y2K4nlho4zhVic8joTACq\nz5yA0d/PkOsDR3ri8AsczuSGU3NSEBFS/N52VxjhpqqGYLsr2A/3lvHU2gLW7et/tz0VHU6GCFBR\np3UO1TdV4RqHNlV1KS3RO+yunJ3JDafkpACQkuDttDtuc8i0e6zc/mIqqdEvqMEi8g+9om54fZ5V\n9LTUODTj6JKVcQyv1H5PcS0iMDErCYAUv6/TjKPtzwDltVbAKK1pjPGRqp6KvPgp14xD9VFLjUMz\nji6lJfqGXXF8d3EN4zOTSLB7laUmeNt9UCLbONsGjjL7i6m0Vr+gBovIwK8Zh+orzTh6KC1h+AWO\nPcW1TMlJDt9P9XfcVOWMrG8bOCpqrfsl1bHJOGoagxyqrI/Ja7tVZKDXjEP1RShkqGnSwNEj6Yk+\nqhuDNIeGRxdGa3LDWqZkp4S3dVbjGJ+ZGP45kpNxlMQo4/jda7u45IH3Y/LabhX5f1SuGYfqg+rG\nIE4/GG2q6kZaojUArr8naigINod4bPV+6gPNTM2NyDgSvB12xx0/wqqBHO0ax97SWg5WNLhuPEIo\nZKhrik09rbohSLw3jrQEr/aqUn0S+R2oGUc30u3A4eYuucHmEE+tLWDJXf/hR89uYfbYNM6ZNSr8\neNviuDGGyvoA4zM7DhzhjCNGgaOoutGV4xGeWHOA0379ZkymuKlqCJCW4CUzOV5rHKpPnC7dyfEe\nDRzdcQJHlYt7Vj3w9h5u/MdGUvxe/nTFQl74xmlkp/jDj6cmeGlqblkFsCEQItBsGJ2RgCdO2tc4\n7C+m0pqmmGQFxXbtxG1XzjsPV1NW29TvmUc7Ut0QJDXBR0ZSvNY4VJ84GcfYzMR+D1GIWeAQkfEi\n8qaIbBORrSLyLXv7CBF5TUTy7NtMe7uIyL0iki8im0RkfsRrXWnvnyciV/bmONLsKTfcnHGs2VvG\njJGprPzmaZw1cyQireczSvFb58BprnLORXqij7QEb6tz0xwyVNQ1keL3EgyZqJ+3UMiEA4fb2uqd\nDO1IVfQDR1W9nXEk+TTjUH3iZPhjMxKpaQwS6kfdN5YZRxD4rjFmJnAKcL2IzARuBl43xkwHXrfv\nAywDptv/rgXuByvQALcAJwMnAbc4waYn0pPc31SVV1TDjFGp7QKGIzxfVXi6AetcpCX4SE/0tcrG\nquoDhAxMy7WK69EeBFhe10TQ/sC67cq51D5XhypjkXEErIwj0ee686aODifjGJORiDFQ2496XMwC\nhzHmkDFmnf1zNbAdGAtcCKywd1sBXGT/fCHwN2NZBWSIyGjgHOA1Y0yZMaYceA1Y2tPjCDdVubQ4\nXtcUpKC8num5KZ3uE8447J5VVREZR9spWZz6hvN60S6QF0V08a102ZWzk3EcjkngCJKW6CUjSWsc\nqm+cC8exdm/K/tQ5jkqNQ0QmAfOAD4CRxhhnoevDwEj757HAgYinFdjbOtve9j2uFZE1IrKmuLg4\nvD0twd0Zx+4ia16q6SO7CBxtMg7nXKQl+khrEzicuoPzetHOOIojAofbrpxjGTiqGgKk+n1kJsVT\n0xgclmvMqP5xLhjHZgyBwCEiKcDTwLeNMVWRjxmr8hqV6qsxZrkxZqExZmFOTk54e1K8B2+cuHYQ\nYF5RNQDTclM73ccJnuGMo6F1xhF5bsrswX/T7dcrrY1dxuGmGkegORT+fQ7FoMbhZByZye6+EFKx\nU90YJMEXR2ZSvHW/H60wMQ0cIuLDChqPGmOesTcfsZugsG+L7O2FwPiIp4+zt3W2vafH0O6q2k3y\nimrweSQ8L1VHnKYq54PiNBGlJXjbNVU5YzismXWjn3E4PY783jhX9aoqjxgsGe2MI9Acoq6pOdyr\nCtzXI03FXlW9VSdrW/Psi1j2qhLgQWC7MeauiIeeB5yeUVcCz0Vsv8LuXXUKUGk3ab0CnC0imXZR\n/Gx7W4+5eWr1vCM1TM5Oxufp/L/SaapqyTis27SIGofT7dapcWSlxDMiKT7qYzmKqxtJ8XsZlZ7g\nqoyj2D5PyfGeqAcOpzdcqt2rCtyVramjo7ohSFqCN7wqaH/qvrFcOnYx8CVgs4hssLf9EPgV8KSI\nXAPsAy6xH3sROBfIB+qALwMYY8pE5OfAanu/nxljynpzIGmJPtcNNnPkF1Uza0x6l/u0ZBwtxfGk\neA8+TxzpiT6CIUNdUzPJfi/ldU3Ee62VE7NT/DEpjuem+klN9LnqqtnJzGaNSWfd/nJCIUNclJZ5\ndf7f0hJ84WYGt9WHVOxV2T3z0qKQccQscBhj3gU6+8tZ0sH+Bri+k9d6CHior8fSdqyCWzQEmtlf\nVseFJ7TrK9BKgs9DvCeuVXHcqXtEjqxP9nspr21iRFI8IkJWSnz0i+NVjeSk+kmM94S7r7qBMyHk\n7LHpfLi3jJLaRnJTE6Ly2s6VYardtAjaVKV6r6ohSHqiL5xxDMqmqsEkPdFHtQsDx57i2lZjLrpi\nTXTorPoXCH8BtZ2Spaw2QGaydVWbFZOMo4GcVD+ZLhsB7XQimD02DYAjldE7b+FxN4m+8P+NNlWp\n3rLGAnlJ8MXhjZPBWxwfLNxaHHd6VHXVFdeR4ve2Gjmelmglm2ltAkdFXVO4HT07JT7qWUFxtXUl\nnuGyEdAlNU34vXHhIB7NaeOdAZqpCV6S4z34POKqc6eOjqr6IGkJPkTEXqNHM44utS0Au0V+UQ1x\nApOzk7vdNzViavWq+mDnGUddU/iqNjvFT3VjkIZAc1SOt7YxSG1TM7lpfteNRyipbiQ7xc+odKt5\n6nAUu+RWR4z0FxF7EKB7sjV1dFTbE2UCpCb4NOPojlMAro/SF+BgkV9Uw6SsZPxeT7f7pvi94Q4C\n1kyrHQcOp8YBVsYB0VsJ0BnDkZPiD2c1FfXu+AIsrmkkOyWe7GQ/3jiJas+qqojiOEBmkk47onqn\nMdhMYzAUbmHQjKMH3Dp6PK+opkf1DWi9JofVVGWdE+e2qj5gTXBYHwh/qWclWzPsRmslQGfUeG6a\nn/TweAR3/J+U1jSRneInLk4YmZYQ1cDhXBk63aqtGXLdcd7U0VEd0aXbudXA0Q03Tq3eFAyxt6S2\nR/UNsGsc9oyYNY3BlisPvxcRK5hU1QcwhojiuJNxRCdwOIP/clMTWjIOl3wBltQ0hqeyH5WeENWJ\nDqsbgqT4vXjs7r0ZLuvKrGLPmR0iNaKpqj/jOIZF4HAKwW7KOPaV1hIMmfDUIN1x2jSrG6zlI522\nzrg4Ca/L7gz+GxFR44DojR4vqrIzDrtXFbhjPEIoZCitbQoH2lHpCVGdWt0a8dvScz5TMw7VS9Vt\nmjs14+gBN64CmFdUA/SsKy60rDseuRaHw+k84EybkZnUNnBEqamqphGfR8hI8pGR5J7xCBV2M59z\nvkanWRlHtDpjWCN+W/6/MpJ9VNa5r7OHip2Wpiq7iVqL491Lj2jHd4u8IzWIWHNK9USK30ug2VBc\nY10Jp3UQOMraBI7EeA/JURyoV1TVSE6KHxGJyDiG/v+JM9YlO7Wlqao+0By1ptGqhvYZR5M9f5VS\nPdEyFqilxlHTGOzzxcewCBxuLI7nFVUzPjOJxPjue1RBS9NUQbk1vqCjjMOpNzgzsEJ0BwEWVTeQ\nk2Z1V02Kt0azu6GpypmnKjuiqQqi1yXXWjY2MnA481UN/XOnjo7q8OwDLU1VIQO1fbz4GBaBw/mj\nc9NiTvm96FEFLT1yDlbYGUdCBxlHmxoHENVpR4qrrYwDsMcj+FyxmJNzfsJNVXbgiNYgwKqGQKsM\nMcNlPdJU7DnZb+Q4Duj71OrDInB4PXGk+N0zX1VjsJndxTU97lEFkOK3PiiFFXVAy5K64IysD1Je\na01wmOhryWKyU/zRq3FUN5Kb5g/fz3DJeASnu7ITOEbaWVW0uuS2zzjc07GgLWNMVEfdK0t1QwAR\nSI5vaaqytvetOXVYBA5w19Tquw7XEGg2zBmb0ePnODPktmQcLV9EzmJOZRETHDqyo5RxBJpDlNY2\nkZsaGTjc0TuopKYRT5yQYWcFuakJiESnqcoYY4/4jcw43Du1+jt5JZz6qzfYXVwz0IfiKlUNQVL9\n3vCMzZpx9FBaos814zg2FVYAMGdc19OpR3KuMArL64mLuPIAq2DW1BziUGVDeAyHIyvZT1ltI6FQ\n/3rwOFlLTkTgyExyx3iE0pox9NADAAAc50lEQVQmspLjw3+U8d44slP8Uck4GgIhAs0m/IcOLYGj\n0gXnrq2NByowBrYUVg70obiKM6W6o6X5XjOOLqUleF3Tq2pzQSUZST7G2YvO90Q4cFTUk5boa7VW\nhFMo/6iklhERhXGwMo6Qsbqc9kd41HjEVONuGY9QUtNIVoq/1bZRadEZBFjdpjcMQEaie3qkteV0\nM887ohlHNLVt7uzvmhzDJnCkJ/ZvpORgsqmgkuPHprdqUuqO01RV09h6TAC0BI6DlfXhwqsjK0pj\nOSIH/zmcyfqG+niEEnueqkjRGgRY1aY3DFgZTYq96Jbb5NuBw7lV0VFV37qDhTZV9dDRnlo90ByK\nSY+hhkAzu45U96qZClp6VUHrrriR940hPMGhI1qDAIsi5qlyZCb5CDSbIT8eoaSmKdxbzDE6StOO\nVLWZY8jhtmnpAZpDJlzbcJYMUNHhLBvr0OJ4Dx3N4rgxhqv/uppP/Pat8PxM0bL9UBXBkOH4XhTG\nAfxeD/Fe6787stkDWgeStjWO8Ay5/SyQO+fBmTgR3NE7yBhjZRyprQPHyLQEKusD1PczKLadKsLh\ntoWwAArK62gMhshN9bOvtM41U+4PBlVtOlgk+jx4+rGY07AKHHVNzQSaY/9hfPSD/byTV0J5XRP/\n++yWqDbFOEXD3mYcYE1oCJ1nHNAyuMwRraaq4upGRiTHh4MXEDHtyNC9cq5pDNIYDJHVJuCOjtIg\nQKcul9ZBxuG2GodT11g6exTBkGFfae0AH5F7tK1x9Hcxp2ETOJw/vFgXyA+U1fGLF7dz2rRsblp6\nLK9uO8LzGw9G7fU3FVSSnRIf/mLqDae5qrMaB7Qe/AfWTKyeOIlCxtHYqr4BLQPZhvKVc9vBf45R\nURoEGM442gT7jKR41/WqyrebqZbOGgW0FMpV/4S7dLf5DGng6AFnwFtfu5/1RChk+N5TG/GI8OuL\n5/DVj01h3oQMbnl+a9SarDYX9r4w7nCuONp/gCIzjtaBIy5OyEqO52BF/74Ai6obW3XFtd5r6I9H\nKGkzT5VjlD0I0Bk301ctxfHWGUemSzOOkWl+TphgNcNqgTw6apuaCZn2n6FUf98nOhw+gSNKM+SG\nQqbTMQ0Pr9rHqj1l/Pj8mYzNSMQTJ9xx8Vzqmpqj0mRV32QVxo8f17v6hiOlk6YqT5yEP1RtMw6A\nkyaP4K1dxf1q5ivpIHC0TJ0xdK+cS9vMU+UYPyKJ5HgPGw9U9Ov1qxsCeOKk1Wh+sM5dVYM1K69b\n5BdVMy03haR4L+MyEzXjiJLIpYcjpSZ4dRxHd7qb6PBgRX14Te7ONASauXT5+3zhz6vafYnmF9Xw\nq5d28PEZOXxu4bjw9mm5KXz3rGN4ddsRHnh7T79+h22HKgkZOH5s7+sb0DLtSNv2cmgJJhltahwA\nF8wdQ1ltE//dXdqn9y2qbuBIVQNjM1qPO3FDjaO4k6YqnyeOhZNG8P6evp0zR1W91RumbYaZk+rH\nGFxTBzDGkF9UE15fZnpuimYcUeIMfE5tFzh82lTVna6mVm8MNnPBfe9y01ObOn2+MYYfPruZ1XvL\nWbWnjDte2Rl+rCHQzDf+vo7EeA+/+eycdn/kX/nYFM6fM5pfvbSDv72/t8+/w6aCvhfGoSVgtG2q\ngpbz01HGccYxOaT6vazsY63mL+/tJWQMn50/rtV2nyeO1CE+HsGZp6qj83bq1Czyi2r61UxZ3WbE\nr+Os40YSJ/DU2oI+v/Zgcqiygdqm5vDEndNyU9hdXDPoM6ojVQ3c9erOQd0DrLqT5s60BK82VXUn\nrYumqjd3FFNa28RLWw5xoKyuw+c/9N5enllXyHc+eQyXnzyB5W/v4Y0dRwD45Yvb2XG4mt9+bi65\nae2L1p444e5LT+CsmSP5yXNbeWL1/j79DpsLKslN9Ycn0eutlG4Ch7/NBIeOBJ+Hs2eN4uWth2kM\n9q57aVVDgEfe38ey40czKTu53eMZyUN7PEJJTSOZST58nvZ/SoumZgGwak9Zl69x/1u7Wbe/vMPH\nqhuC7bpPg1V8P/PYXJ5cU3BUegrGWtuFyabnptIUDFFQ3vHf42Bx+7+2c+8b+azcFL0OMH2x83A1\nj33Y8fdKy1ocWhzvtXDG0UGEfXZ9ARlJPkSEh1fta/f4e/kl/OLF7Zw9cyTfPHMaPz5/JseNTuN/\nntzIw+/vZcX7+7jmtMl84tjcTt/f54nj91+Yx+nH5HDzM5v71NNqc2Fln7MN6LzG4WzLbDPBYaTz\n546muiHI27tKevWej6zaR3VjkK+dMbXDx4f6eITSmqZ2zVSOWWPSSU3w8n4XTXybCyr59cs7uOPl\nnR0+XtUQINXf/v8L4PMnTaCkppHXtxf1/sAHGadZarodOKbat4N56pGtByvDf8cdfW8cLcZYnXJ+\n8MzmDpv3qjsZRJqa4OvzYk7DJnAk+KwBcG0zjoq6Jt7YUcRn549j6exRPP7hfuqaWqJwQXkd1/99\nHVOyk7nr0hOIixMSfB7+ePl8AsEQP35uK7PGpPH9pTO6PQa/18MDX1zAiRNHcOM/NrK+k6vMjtQ2\nBskvrun1wL9InXXHBbjq1EnctKzz3+G0adlkJPl6dWXVEGjmoXc/4vRjcpjdSV0mPXFo9w6y5qlq\n30wFVqZ58uQRrOqizuE0Xb6/p7TDq+vOMg6wmhBHpSV0eqU5lOQXVTMiOT48bsjJPAZzgfzOV3aS\nnujj25+czvr9FQM2MeNr246Em7Ef7+Cz0DIWqH3G0Rzq28wNwyZwgHXi2tY4Vm46RKDZ8Ol5Y7l6\n8SSqGoI8va4QsGofX390Hc3NhuVXLAxfsQNMzk7mjs/NZVpuCvd9fh5+b89W4kuM9/B/X1rAyDQ/\n1z68tkf9/BuD1hewMX2vbwCMzUgk3hPXbnoMgJOnZPHpeeM6eJbF54lj2ezRvLbtSI9HQ/9jbQEl\nNU2dZhtgZRxDsVfVwYp6bv/XNjYVVjImvfPJJk+ZksVHJbUd/j+X1zbx3MaDfGJGDgD/XF/Ybp+q\n+o5rHGCtM3PJwnG8nVc86Jt0upN3pIZpEcsgpyf6GJnmH7QF8g/2lPLmzmK+9vGpfHnxZBJ9nn7V\nL/sqFDLc9douJmcns3TWKJ5aV0BDoPXfZ2fT1rTMV9X75qphFTjSE73tplZ/dn0hx4xMYdaYNOZP\nyGTOuHT++t5HhEKGn6/cxqaCSu743Fwmd9A+f+7xo/n3/5zBlB6u++0YkRzPg1eeSF1jkGv/tjb8\nRdwQaGZLYSWbCirIO1LNgbI6Hn5/L2f85i1++9ouFk3J4pQpWX3+/c+fM4Y3bjyj1SJOvXHBnNHU\nNTXzxo7um0aCzSGWv72beRMyOGXKiE73yxxicy5VNwT47pMbOf03b/LQe3tZOmsU3z2n80zNqXN0\n1Fz1+OoDNAVD3LzsOE6ePIJn1hW2azZoO+K3rUtOHA/Ak6sP9OXXGRSMMeQV1TCtzcJk03JTyB8E\nc1YZY3huQyFbD1ZijMEYw29e2Uluqp8rF00iPdHHRfPG8NyGgzFb0bK4upFfvrSds+/+D+/mtTQX\n/2vzIXYcrubbn5zOlxZNpKIuwCtbD7d6bnVDkHhvHAlt6pct81X1/pg7/0S6UNuJDveV1rJ2Xzk3\nLT023Lb/5cWT+M4TG7n5mU08uaaAa0+fwtLZo6J+LMeMTOWey+bx1YfX8Pk/rSJkDNsPVRFobt/e\neOKkTO66ZC6Lpmb1aeCfwxMnjMtM6vPzT56SRXaKnxc2HuS8OaO73Hf5O3s4UFbPj8+b2eUxR45H\n8MT1/Xc7Wn7y3Fae33iQKxZN5OrFkxk/ouvzedyoNDKSfLy/u5TPRPQqaw4ZHlm1j0VTspgxKpXP\nLhjH95/axIYDFcybkBnep7qD2YwjjctM4vTpOTy5poAblkzH20GRfrArqWmisj7QKuMAq0D+jzUH\nMMb063PfX3//cD8/enYLAMeMTGHBxEzW7ivn9k/PJjHe+jL+0imTeOzDA/xj7QG+8rEpUXvvI1UN\n3P/Wbh77cD9NzSFyUvxc9ZcPuf3Ts/ns/HHc/e9dzBiZygVzxgAwKSuJRz/Yz4UnjA2/hjVPVfuv\n+v6syTH0PmX9kJ7o42BlPbX2eI1n1hUiAhfNGxPe57zjx5CT6ufJNQWcNGkE3+/iarK/PjlzJD86\n9zj2ltaS4vfylY9N4Y+Xz+fBKxfy+y/M446L5/DkdYt48rpFnDote0D/eMAKPOfPGc0bO4rYdrCq\n0/2e33iQ37y8k/PmjOasmSO7fM3MJB/GdNzbrTlkwut4DAbPbSjk2fWF3HDmdG65YFa3QQOskfcn\nT24/nuP17UcorKjnylMnArBs9igSfHE8s66lucoZV9RVxgFWkfxwVQMvbjnc5X6DlTMTbtulkKfm\nplDb1ByVWYY7svFABXe/tov9pZ038xWU1/GLf23n1KlZ3HbRbNISfDz24QEmZydzycLx4f1mjklj\n4cRMHl61r9tFz4wxPPjuR9zy3BZe2nyIstqOm2pXbjrIWXf9h0dW7eNTc8fw+v+cwevfPYNFU7O4\n6enNXPHQh+wpruU7Z00nLk6IixMuO2kCH35U1ipTs7LW9hcf/ZlafVhlHKdOzeKtncWc+qs3uOrU\nSTy7vpBFU7IYHdFGHe+N44Yl0/nLex/x+y/Mi/kV3Fc+NiWqVyixdv0npvHi5kN87dG1PP+N09r1\n0Fq1p5Qbn9zISZNH8NvPze022Dmz8ZbXNbUaC7G/tI5vPbGezQWV/PXLJ3Ha9OxWzzPGUB9oJik+\n+h/hQHOI9/JLmJKdwoQsKzgcKKvjf5/dwoKJmVz/ic5rNh1ZNCWLV7Ye4UBZXTjYrHh/L2PSE/jk\ncVZgTU3wcc6sUbyw6SD/e/5x+L2eiEWcum5aXHJcLseNTuP7T20kN9Xfr+bMgbA73KMqtdX26REF\n8jEZPV+0rDsbD1Rwz+t54SbX5W/v4aalM7hi0aRWC5wZY/jBM5sxwK8/O4fxI5L44ikTKSivw+/1\ntOuC/aVFE/nW4xv463/3csmJ41vVRCPd83oev/t3HvGeOFa8b/XGOnZUKmfMyOETM3I5dlQqP1+5\nnafXFXDC+AzuvvSEVk3lD111Ij95biuPfbifWWPSOGdWS4vIxQvG8dtXd/LYhwf48fkzAeuirKOM\noz+LOQ2rwHHt6VM5cdII/vjWbu55PQ+Ab545rd1+XzplIl88ecKAX+EPRjmpfv54+XwuW76KG/+x\nkQe+uCD8x7bjcBXX/m0NE7KS+NOXFrZrU+2IM+3IlsJKxmYkkuDz8NyGQn707BZEYFxmIl97dC3P\nfv1UptlfLAcr6vnKijVsO1TF2IxEZoxKZcaoVI4ZmcL03FSm5qSEmxB6o76pmSfXHGD523sorLCW\n2D1/zhiuPX0Ktz6/FQP87tITen0xsWiqFfTe31NKvDeOlZsO8V5+Kd87Z0ar1/rM/HE8t+Egb+4o\nYvyIJO75t/UZbTvzbls+TxwPX3MSn1++iqv/upq/XX0SCyd1XlfqrdrGIB98VMqYjEQmZyd32xEk\nFDI8v/Eg97yeR3ldExOzkpmUlcRxo9O47MTxrRYLM8awqaCSFL+XkWmtO204Pat2HKrijGNyeny8\nweYQmworOVBWR2FFPYXl9RRVN1JSY/07UFZPRpKP750zg3NmjeS2f23n1he28a/Nh7h52bHMG59J\nXJzwxOoDvJNXws8vmt0qu+ysuXfZ7NH8Pjefn63cxi9f2s5Jk0ewdNYoPjN/HMl2EHl41T5+9+88\nLl4wjl98+ng2F1ayak8p7+aV8OA7H/HAf/YgAgLccOY0vrlkersA5fPE8YtPz+Zj07M5bnRaq++p\n7BQ/Z88axdPrChiRHM/LWw6zubCSJR0MFehPcVyG+uprHVm4cKFZs2ZNl/vsOlLN27uK+dKiiT3u\nEaVaPPTuR/xs5Ta+v3QG8ydk8rf39/LK1iOMSI7n2a+f2uNaSn5RDZ+86z8AxAmMTk+ksKKehRMz\n+d1lJwBw0R/+S2J8HP/8+mIOVjRwzYrV1DU1c+WpEzlQVs+uI9XsLq4J14fEXlM93huHzyP4PHF4\n4gSPCF6PkJbgIzM5PrxoVUV9ExV1AXYdqaa8LsCCiZlcvXgymwoqeGTVPmrtzgt3Xzq3y55nnTHG\nsPC2fxMMGaoaAhgDs8em8fDVJ7da/yTYHOLUX71BU3OIiroAKX4vV546kf85a0aP6j9FVQ1cunwV\nxdWN3Pm5uUwYkUSy30N6oq/dyo7GGF7ecpj8ohouPWl8qyV9I63fX863n9jAPrs5xxMnTMxKYtaY\ndOaMTWfOuHQmZCXRFAzREAjxUUktv/v3LnYcrmbm6DROmJDB/tI69pbWUlBeH/6drl48mdV7y7j/\nrd1sLKjkrJkj+dMVC9sd4yfv+g8fldTyuQXj+fZZ0xmdnkhTMMT6/eVsLKhgak4KCyeNID3RR21j\nkCdWH+DBdz+iMGJSzowkHyNTE8hOjSc7xc/M0Wl84eQJ4S9OYwxPryvkZy9spaohSHZKPJ+YkctL\nWw5z/Nh0Hv3Kya0yka4Em0Os3VfOGzuKeH1HEflFNaQleLn8lImMz0ziR//czJkzcnngSwvaXYBU\nNwR4L7+UtfvKOGfWqD4H///uLuELf/oAgBPGZ3Du8aP49Lxx7eaKq20MMuuWV/jY9GyWHJvLqPRE\nlh0/eq0xZmFHrxtpyAQOEVkK3AN4gD8bY37V2b49CRyqf4wxfPOx9azcdAiw6keXnTieqxZPatX0\n1xMfldSypbCSvCPV4bEqX/3Y5PAf1oYDFVz6wPtMykpmf1kdI5LjeeiqE5kxqqVpI9AcYl9pLbuO\n1JB3pIbK+gCB5hBNwRCB5hDNxtAcMgSaQ1TVBymvawq3LWcmxZOe5GN0egJfPGUiJ0b8wVbWBXjk\nA6s54esfn9rnLPSOV3bw1s5izpk1imWzRzF9ZGqH+/3+jTz++t+9XLloElecOqnDwZpdOVRZz6UP\nrGJ/mxkQTpyUycULxnHu8aNZv7+CO17ZyWZ73EGiz8OVp07i/50xJRxggs0h/vDmbu59I49RaQn8\n+PyZNAabyTtSw84j1WwtrORgJ7WHCSOS+O7Zx3DBnDGtvnB3HK7ivjfyeXGz9ZkxxirmXnv6VD4z\nf2yHGWpZbRN/eDOfh9/fhwgsmJjJhgMVrcYexAkcNzqNA2V1VDUEOWnSCK44dSIzRqYyJiMxfLXf\nncr6AG/tLOK1bUf4z65ijIEXb/hYuLmyL9buK+fP7+zhla2HCRmsOsg1J/cpI+6Nd/NKmJKT3GUT\nnzGGz/9pFev2VdBkzz6w79fnuydwiIgH2AWcBRQAq4HPG2O2dbS/Bo6jo7YxyC9f2s6csRlcMHdM\nTP8YXtx8iK8/uo654zP40xULOr1CVlYvmvX7K6hvClLX1ExBeT3PbShkd3Et3jghGDKMzUjkO2cd\nw7wJGdz3eh7PbTxIos/KThqDIeqagjQEQlx0whh+euHsDgNYcXUjmwsrOFLViN8bh9/rISXBy6Ip\nWa0W7Gpr15Fq/rHmACeMz2Tp7FE9yqYOlNXxu3/nsaWwkpMmj+C06dnMn5DJ7uIaVu0p5YM9ZWSl\nxHPNaZPDvdL6I9Acoq6pudeBuzP7S+t4ddthPrdgfJ+7w8eKMYay2iYOVTZw/LgMVwWORcCtxphz\n7Ps/ADDG/LKj/TVwuNPu4ppwHUT1jjGGDQcq+NemQ0zISuLSE8e3aqLdebiah1ftpTEQIsHnwe+1\nZveNRVd0NXiJSI8Cx1Apjo8FIkc4FQAnD9CxqAEytZcDLVULEWHehMxOr8ZnjErltouOP8pHpYYq\n14zjEJFrRWSNiKwpLi4e6MNRSinXGiqBoxAYH3F/nL0tzBiz3Biz0BizMCen5133lFJK9c5QCRyr\ngekiMllE4oHLgOcH+JiUUmpYGhI1DmNMUES+AbyC1R33IWPM1gE+LKWUGpaGROAAMMa8CLw40Meh\nlFLD3VBpqlJKKTVIaOBQSinVKxo4lFJK9cqQGDneWyJSDewc6OMYRLKBkm73Gj70fLTQc9HacD8f\nE40x3Y5nGDLF8V7a2ZNh88OFiKzR89FCz0cLPRet6fnoGW2qUkop1SsaOJRSSvWKWwPH8oE+gEFG\nz0drej5a6LloTc9HD7iyOK6UUip23JpxKKWUihHXBQ4RWSoiO0UkX0RuHujjOZpEZLyIvCki20Rk\nq4h8y94+QkReE5E8+7b/S6QNISLiEZH1IrLSvj9ZRD6wPyNP2BNnDgsikiEiT4nIDhHZLiKLhuvn\nQ0S+Y/+dbBGRx0QkYTh/NnrDVYHDXmL2D8AyYCbweRGZObBHdVQFge8aY2YCpwDX27//zcDrxpjp\nwOv2/eHkW8D2iPu/Bu42xkwDyoFrBuSoBsY9wMvGmGOBuVjnZdh9PkRkLHADsNAYMxtr8tTLGN6f\njR5zVeAATgLyjTF7jDFNwOPAhQN8TEeNMeaQMWad/XM11pfCWKxzsMLebQVw0cAc4dEnIuOA84A/\n2/cFOBN4yt5l2JwPEUkHTgceBDDGNBljKhi+nw8vkCgiXiAJOMQw/Wz0ltsCR0dLzI4doGMZUCIy\nCZgHfACMNMYcsh86DIwcoMMaCL8Dvg+E7PtZQIUxJmjfH06fkclAMfAXu+nuzyKSzDD8fBhjCoE7\ngf1YAaMSWMvw/Wz0itsChwJEJAV4Gvi2MaYq8jFjdaMbFl3pROR8oMgYs3agj2WQ8ALzgfuNMfOA\nWto0Sw2Xz4ddx7kQK5iOAZKBpQN6UEOI2wJHt0vMup2I+LCCxqPGmGfszUdEZLT9+GigaKCO7yhb\nDHxKRPZiNVueidXGn2E3T8Dw+owUAAXGmA/s+09hBZLh+Pn4JPCRMabYGBMAnsH6vAzXz0avuC1w\nDOslZu32+weB7caYuyIeeh640v75SuC5o31sA8EY8wNjzDhjzCSsz8IbxpjLgTeBi+3dhtP5OAwc\nEJEZ9qYlwDaG5+djP3CKiCTZfzfOuRiWn43ect0AQBE5F6td21li9vYBPqSjRkROA94BNtPSpv9D\nrDrHk8AEYB9wiTGmbEAOcoCIyMeBG40x54vIFKwMZASwHviiMaZxII/vaBGRE7A6CsQDe4AvY11A\nDrvPh4j8FLgUqzfieuArWDWNYfnZ6A3XBQ6llFKx5bamKqWUUjGmgUMppVSvaOBQSinVKxo4lFJK\n9YoGDqWUUr2igUOpHhCRZhHZEPGvy4kAReT/icgVUXjfvSKS3d/XUSqatDuuUj0gIjXGmJQBeN+9\nWDO4lhzt91aqM5pxKNUPdkbwGxHZLCIfisg0e/utInKj/fMN9hopm0TkcXvbCBH5p71tlYjMsbdn\nicir9joRfwYk4r2+aL/HBhF5wF5GQKmjTgOHUj2T2Kap6tKIxyqNMccDv8eataCtm4F5xpg5wP+z\nt/0UWG9v+yHwN3v7LcC7xphZwLNYo7kRkeOwRjkvNsacADQDl0f3V1SqZ7zd76KUAurtL+yOPBZx\ne3cHj28CHhWRfwL/tLedBnwWwBjzhp1ppGGtl/EZe/u/RKTc3n8JsABYbU2tRCLDYzJCNQhp4FCq\n/0wnPzvOwwoIFwA/EpHj+/AeAqwwxvygD89VKqq0qUqp/rs04vb9yAdEJA4Yb4x5E7gJSAdSsCaj\nvNze5+NAib12ytvAF+ztywBn/e/XgYtFJNd+bISITIzh76RUpzTjUKpnEkVkQ8T9l40xTpfcTBHZ\nBDQCn2/zPA/wiL1sqwD3GmMqRORW4CH7eXW0TGv+U+AxEdkK/Bdr+m+MMdtE5H+BV+1gFACux5rN\nVqmjSrvjKtUP2l1WDUfaVKWUUqpXNONQSinVK5pxKKWU6hUNHEoppXpFA4dSSqle0cChlFKqVzRw\nKKWU6hUNHEoppXrl/wN3oxdwQdcpCQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116473908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "step_result.plot()\n",
    "plt.xlabel('Episode')\n",
    "plt.ylabel('Steps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "step_result.to_csv('dueling_DQN_mountaincar.csv', index=False)"
   ]
  }
 ],
 "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
}