{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## 1、定义算法\n", "\n", "在阅读该教程之前，请先阅读Q learning教程。Sarsa算法跟Q learning算法基本模式相同，但是根本的区别在于，Sarsa是先做出动作然后拿这个做的动作去更新，而Q learning是假定下一步最大奖励对应的动作拿去更新，然后再使用$\\varepsilon$-greedy策略，也就是说Sarsa是on-policy的，而Q learning是off-policy的。如下方代码所示，只有在更新的地方Sarsa与Q learning有着细微的区别。" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "from collections import defaultdict\n", "import torch\n", "import math\n", "class Sarsa(object):\n", " def __init__(self,\n", " n_actions,cfg):\n", " self.n_actions = n_actions \n", " self.lr = cfg.lr \n", " self.gamma = cfg.gamma \n", " self.sample_count = 0 \n", " self.epsilon_start = cfg.epsilon_start\n", " self.epsilon_end = cfg.epsilon_end\n", " self.epsilon_decay = cfg.epsilon_decay \n", " self.Q = defaultdict(lambda: np.zeros(n_actions)) # Q table\n", " def sample(self, state):\n", " self.sample_count += 1\n", " self.epsilon = self.epsilon_end + (self.epsilon_start - self.epsilon_end) * \\\n", " math.exp(-1. * self.sample_count / self.epsilon_decay) # The probability to select a random action, is is log decayed\n", " best_action = np.argmax(self.Q[state])\n", " action_probs = np.ones(self.n_actions, dtype=float) * self.epsilon / self.n_actions\n", " action_probs[best_action] += (1.0 - self.epsilon)\n", " action = np.random.choice(np.arange(len(action_probs)), p=action_probs) \n", " return action\n", " def predict(self,state):\n", " return np.argmax(self.Q[state])\n", " def update(self, state, action, reward, next_state, next_action,done):\n", " Q_predict = self.Q[state][action]\n", " if done:\n", " Q_target = reward # 终止状态\n", " else:\n", " Q_target = reward + self.gamma * self.Q[next_state][next_action] # 与Q learning不同，Sarsa是拿下一步动作对应的Q值去更新\n", " self.Q[state][action] += self.lr * (Q_target - Q_predict) \n", " def save(self,path):\n", " '''把 Q表格的数据保存到文件中\n", " '''\n", " import dill\n", " torch.save(\n", " obj=self.Q,\n", " f=path+\"sarsa_model.pkl\",\n", " pickle_module=dill\n", " )\n", " def load(self, path):\n", " '''从文件中读取数据到 Q表格\n", " '''\n", " import dill\n", " self.Q =torch.load(f=path+'sarsa_model.pkl',pickle_module=dill)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 2、定义训练\n", "\n", "同样地，跟Q learning差别也不大" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "def train(cfg,env,agent):\n", " print('开始训练！')\n", " print(f'环境:{cfg.env_name}, 算法:{cfg.algo_name}, 设备:{cfg.device}')\n", " rewards = [] # 记录奖励\n", " for i_ep in range(cfg.train_eps):\n", " ep_reward = 0 # 记录每个回合的奖励\n", " state = env.reset() # 重置环境,即开始新的回合\n", " action = agent.sample(state)\n", " while True:\n", " action = agent.sample(state) # 根据算法采样一个动作\n", " next_state, reward, done, _ = env.step(action) # 与环境进行一次动作交互\n", " next_action = agent.sample(next_state)\n", " agent.update(state, action, reward, next_state, next_action,done) # 算法更新\n", " state = next_state # 更新状态\n", " action = next_action\n", " ep_reward += reward\n", " if done:\n", " break\n", " rewards.append(ep_reward)\n", " print(f\"回合：{i_ep+1}/{cfg.train_eps}，奖励：{ep_reward:.1f}，Epsilon：{agent.epsilon}\")\n", " print('完成训练！')\n", " return {\"rewards\":rewards}\n", " \n", "def test(cfg,env,agent):\n", " print('开始测试！')\n", " print(f'环境：{cfg.env_name}, 算法：{cfg.algo_name}, 设备：{cfg.device}')\n", " rewards = [] # 记录所有回合的奖励\n", " for i_ep in range(cfg.test_eps):\n", " ep_reward = 0 # 记录每个episode的reward\n", " state = env.reset() # 重置环境, 重新开一局（即开始新的一个回合）\n", " while True:\n", " action = agent.predict(state) # 根据算法选择一个动作\n", " next_state, reward, done, _ = env.step(action) # 与环境进行一个交互\n", " state = next_state # 更新状态\n", " ep_reward += reward\n", " if done:\n", " break\n", " rewards.append(ep_reward)\n", " print(f\"回合数：{i_ep+1}/{cfg.test_eps}, 奖励：{ep_reward:.1f}\")\n", " print('完成测试！')\n", " return {\"rewards\":rewards}" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 3、定义环境\n", "为了具体看看Q learning和Sarsa的不同，笔者决定跟Q learning使用相同的环境\n" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "import gym\n", "import turtle\n", "import numpy as np\n", "\n", "# turtle tutorial : https://docs.python.org/3.3/library/turtle.html\n", "\n", "def GridWorld(gridmap=None, is_slippery=False):\n", " if gridmap is None:\n", " gridmap = ['SFFF', 'FHFH', 'FFFH', 'HFFG']\n", " env = gym.make(\"FrozenLake-v0\", desc=gridmap, is_slippery=False)\n", " env = FrozenLakeWapper(env)\n", " return env\n", "\n", "\n", "class FrozenLakeWapper(gym.Wrapper):\n", " def __init__(self, env):\n", " gym.Wrapper.__init__(self, env)\n", " self.max_y = env.desc.shape[0]\n", " self.max_x = env.desc.shape[1]\n", " self.t = None\n", " self.unit = 50\n", "\n", " def draw_box(self, x, y, fillcolor='', line_color='gray'):\n", " self.t.up()\n", " self.t.goto(x * self.unit, y * self.unit)\n", " self.t.color(line_color)\n", " self.t.fillcolor(fillcolor)\n", " self.t.setheading(90)\n", " self.t.down()\n", " self.t.begin_fill()\n", " for _ in range(4):\n", " self.t.forward(self.unit)\n", " self.t.right(90)\n", " self.t.end_fill()\n", "\n", " def move_player(self, x, y):\n", " self.t.up()\n", " self.t.setheading(90)\n", " self.t.fillcolor('red')\n", " self.t.goto((x + 0.5) * self.unit, (y + 0.5) * self.unit)\n", "\n", " def render(self):\n", " if self.t == None:\n", " self.t = turtle.Turtle()\n", " self.wn = turtle.Screen()\n", " self.wn.setup(self.unit * self.max_x + 100,\n", " self.unit * self.max_y + 100)\n", " self.wn.setworldcoordinates(0, 0, self.unit * self.max_x,\n", " self.unit * self.max_y)\n", " self.t.shape('circle')\n", " self.t.width(2)\n", " self.t.speed(0)\n", " self.t.color('gray')\n", " for i in range(self.desc.shape[0]):\n", " for j in range(self.desc.shape[1]):\n", " x = j\n", " y = self.max_y - 1 - i\n", " if self.desc[i][j] == b'S': # Start\n", " self.draw_box(x, y, 'white')\n", " elif self.desc[i][j] == b'F': # Frozen ice\n", " self.draw_box(x, y, 'white')\n", " elif self.desc[i][j] == b'G': # Goal\n", " self.draw_box(x, y, 'yellow')\n", " elif self.desc[i][j] == b'H': # Hole\n", " self.draw_box(x, y, 'black')\n", " else:\n", " self.draw_box(x, y, 'white')\n", " self.t.shape('turtle')\n", "\n", " x_pos = self.s % self.max_x\n", " y_pos = self.max_y - 1 - int(self.s / self.max_x)\n", " self.move_player(x_pos, y_pos)\n", "\n", "\n", "class CliffWalkingWapper(gym.Wrapper):\n", " def __init__(self, env):\n", " gym.Wrapper.__init__(self, env)\n", " self.t = None\n", " self.unit = 50\n", " self.max_x = 12\n", " self.max_y = 4\n", "\n", " def draw_x_line(self, y, x0, x1, color='gray'):\n", " assert x1 > x0\n", " self.t.color(color)\n", " self.t.setheading(0)\n", " self.t.up()\n", " self.t.goto(x0, y)\n", " self.t.down()\n", " self.t.forward(x1 - x0)\n", "\n", " def draw_y_line(self, x, y0, y1, color='gray'):\n", " assert y1 > y0\n", " self.t.color(color)\n", " self.t.setheading(90)\n", " self.t.up()\n", " self.t.goto(x, y0)\n", " self.t.down()\n", " self.t.forward(y1 - y0)\n", "\n", " def draw_box(self, x, y, fillcolor='', line_color='gray'):\n", " self.t.up()\n", " self.t.goto(x * self.unit, y * self.unit)\n", " self.t.color(line_color)\n", " self.t.fillcolor(fillcolor)\n", " self.t.setheading(90)\n", " self.t.down()\n", " self.t.begin_fill()\n", " for i in range(4):\n", " self.t.forward(self.unit)\n", " self.t.right(90)\n", " self.t.end_fill()\n", "\n", " def move_player(self, x, y):\n", " self.t.up()\n", " self.t.setheading(90)\n", " self.t.fillcolor('red')\n", " self.t.goto((x + 0.5) * self.unit, (y + 0.5) * self.unit)\n", "\n", " def render(self):\n", " if self.t == None:\n", " self.t = turtle.Turtle()\n", " self.wn = turtle.Screen()\n", " self.wn.setup(self.unit * self.max_x + 100,\n", " self.unit * self.max_y + 100)\n", " self.wn.setworldcoordinates(0, 0, self.unit * self.max_x,\n", " self.unit * self.max_y)\n", " self.t.shape('circle')\n", " self.t.width(2)\n", " self.t.speed(0)\n", " self.t.color('gray')\n", " for _ in range(2):\n", " self.t.forward(self.max_x * self.unit)\n", " self.t.left(90)\n", " self.t.forward(self.max_y * self.unit)\n", " self.t.left(90)\n", " for i in range(1, self.max_y):\n", " self.draw_x_line(\n", " y=i * self.unit, x0=0, x1=self.max_x * self.unit)\n", " for i in range(1, self.max_x):\n", " self.draw_y_line(\n", " x=i * self.unit, y0=0, y1=self.max_y * self.unit)\n", "\n", " for i in range(1, self.max_x - 1):\n", " self.draw_box(i, 0, 'black')\n", " self.draw_box(self.max_x - 1, 0, 'yellow')\n", " self.t.shape('turtle')\n", "\n", " x_pos = self.s % self.max_x\n", " y_pos = self.max_y - 1 - int(self.s / self.max_x)\n", " self.move_player(x_pos, y_pos)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "def env_agent_config(cfg,seed=1):\n", " '''创建环境和智能体\n", " Args:\n", " cfg ([type]): [description]\n", " seed (int, optional): 随机种子. Defaults to 1.\n", " Returns:\n", " env [type]: 环境\n", " agent : 智能体\n", " ''' \n", " env = gym.make(cfg.env_name) \n", " env = CliffWalkingWapper(env)\n", " env.seed(seed) # 设置随机种子\n", " n_states = env.observation_space.n # 状态维度\n", " n_actions = env.action_space.n # 动作维度\n", " print(f\"状态数：{n_states}，动作数：{n_actions}\")\n", " agent = Sarsa(n_actions,cfg)\n", " return env,agent" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 4、设置参数\n", "同样的参数也是一样" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "import datetime\n", "import argparse\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "def get_args():\n", " \"\"\" \n", " \"\"\"\n", " curr_time = datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\") # 获取当前时间\n", " parser = argparse.ArgumentParser(description=\"hyperparameters\") \n", " parser.add_argument('--algo_name',default='Sarsa',type=str,help=\"name of algorithm\")\n", " parser.add_argument('--env_name',default='CliffWalking-v0',type=str,help=\"name of environment\")\n", " parser.add_argument('--train_eps',default=400,type=int,help=\"episodes of training\") # 训练的回合数\n", " parser.add_argument('--test_eps',default=20,type=int,help=\"episodes of testing\") # 测试的回合数\n", " parser.add_argument('--gamma',default=0.90,type=float,help=\"discounted factor\") # 折扣因子\n", " parser.add_argument('--epsilon_start',default=0.95,type=float,help=\"initial value of epsilon\") # e-greedy策略中初始epsilon\n", " parser.add_argument('--epsilon_end',default=0.01,type=float,help=\"final value of epsilon\") # e-greedy策略中的终止epsilon\n", " parser.add_argument('--epsilon_decay',default=300,type=int,help=\"decay rate of epsilon\") # e-greedy策略中epsilon的衰减率\n", " parser.add_argument('--lr',default=0.1,type=float,help=\"learning rate\")\n", " parser.add_argument('--device',default='cpu',type=str,help=\"cpu or cuda\") \n", " args = parser.parse_args([]) \n", " return args\n", "\n", "def smooth(data, weight=0.9): \n", " '''用于平滑曲线，类似于Tensorboard中的smooth\n", "\n", " Args:\n", " data (List):输入数据\n", " weight (Float): 平滑权重，处于0-1之间，数值越高说明越平滑，一般取0.9\n", "\n", " Returns:\n", " smoothed (List): 平滑后的数据\n", " '''\n", " last = data[0] # First value in the plot (first timestep)\n", " smoothed = list()\n", " for point in data:\n", " smoothed_val = last * weight + (1 - weight) * point # 计算平滑值\n", " smoothed.append(smoothed_val) \n", " last = smoothed_val \n", " return smoothed\n", "\n", "def plot_rewards(rewards,cfg, tag='train'):\n", " sns.set()\n", " plt.figure() # 创建一个图形实例，方便同时多画几个图\n", " plt.title(f\"{tag}ing curve on {cfg.device} of {cfg.algo_name} for {cfg.env_name}\")\n", " plt.xlabel('epsiodes')\n", " plt.plot(rewards, label='rewards')\n", " plt.plot(smooth(rewards), label='smoothed')\n", " plt.legend()\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 5、开始训练\n", "训练结果会发现，Sarsa收敛速度更快，但收敛值会比Q-learning低" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "状态数：48，动作数：4\n", "开始训练！\n", "环境:CliffWalking-v0, 算法:Sarsa, 设备:cpu\n", "回合：1/400，奖励：-1524.0，Epsilon：0.2029722781251147\n", "回合：2/400，奖励：-1294.0，Epsilon：0.011808588201828951\n", "回合：3/400，奖励：-192.0，Epsilon：0.01050118158853445\n", "回合：4/400，奖励：-346.0，Epsilon：0.010049747911736582\n", "回合：5/400，奖励：-252.0，Epsilon：0.010009240861841986\n", "回合：6/400，奖励：-168.0，Epsilon：0.010003005072880926\n", "回合：7/400，奖励：-393.0，Epsilon：0.01000042188120369\n", "回合：8/400，奖励：-169.0，Epsilon：0.010000136281659052\n", "回合：9/400，奖励：-97.0，Epsilon：0.010000071145264558\n", "回合：10/400，奖励：-134.0，Epsilon：0.010000029022085234\n", "回合：11/400，奖励：-124.0，Epsilon：0.010000012655059554\n", "回合：12/400，奖励：-74.0，Epsilon：0.010000007701309915\n", "回合：13/400，奖励：-135.0，Epsilon：0.010000003120699265\n", "回合：14/400，奖励：-84.0，Epsilon：0.010000001776639691\n", "回合：15/400，奖励：-101.0，Epsilon：0.010000000903081117\n", "回合：16/400，奖励：-111.0，Epsilon：0.010000000429438717\n", "回合：17/400，奖励：-114.0，Epsilon：0.010000000200165738\n", "回合：18/400，奖励：-114.0，Epsilon：0.010000000093299278\n", "回合：19/400，奖励：-82.0，Epsilon：0.010000000053829002\n", "回合：20/400，奖励：-85.0，Epsilon：0.01000000003044167\n", "回合：21/400，奖励：-108.0，Epsilon：0.010000000014768242\n", "回合：22/400，奖励：-66.0，Epsilon：0.010000000009479634\n", "回合：23/400，奖励：-74.0，Epsilon：0.010000000005768887\n", "回合：24/400，奖励：-114.0，Epsilon：0.010000000002688936\n", "回合：25/400，奖励：-98.0，Epsilon：0.010000000001394421\n", "回合：26/400，奖励：-94.0，Epsilon：0.010000000000742658\n", "回合：27/400，奖励：-58.0，Epsilon：0.010000000000502822\n", "回合：28/400，奖励：-100.0，Epsilon：0.010000000000257298\n", "回合：29/400，奖励：-208.0，Epsilon：0.010000000000123995\n", "回合：30/400，奖励：-184.0，Epsilon：0.010000000000070121\n", "回合：31/400，奖励：-62.0，Epsilon：0.010000000000046227\n", "回合：32/400，奖励：-117.0，Epsilon：0.01000000000002112\n", "回合：33/400，奖励：-47.0，Epsilon：0.010000000000015387\n", "回合：34/400，奖励：-54.0，Epsilon：0.0100000000000107\n", "回合：35/400，奖励：-120.0，Epsilon：0.010000000000004792\n", "回合：36/400，奖励：-75.0，Epsilon：0.010000000000002897\n", "回合：37/400，奖励：-62.0，Epsilon：0.01000000000000191\n", "回合：38/400，奖励：-70.0，Epsilon：0.010000000000001194\n", "回合：39/400，奖励：-67.0，Epsilon：0.010000000000000762\n", "回合：40/400，奖励：-87.0，Epsilon：0.010000000000000425\n", "回合：41/400，奖励：-92.0，Epsilon：0.01000000000000023\n", "回合：42/400，奖励：-79.0，Epsilon：0.010000000000000136\n", "回合：43/400，奖励：-49.0，Epsilon：0.010000000000000097\n", "回合：44/400，奖励：-103.0，Epsilon：0.010000000000000049\n", "回合：45/400，奖励：-40.0，Epsilon：0.010000000000000037\n", "回合：46/400，奖励：-214.0，Epsilon：0.010000000000000018\n", "回合：47/400，奖励：-83.0，Epsilon：0.01000000000000001\n", "回合：48/400，奖励：-62.0，Epsilon：0.010000000000000007\n", "回合：49/400，奖励：-37.0，Epsilon：0.010000000000000005\n", "回合：50/400，奖励：-73.0，Epsilon：0.010000000000000004\n", "回合：51/400，奖励：-66.0，Epsilon：0.010000000000000002\n", "回合：52/400，奖励：-48.0，Epsilon：0.010000000000000002\n", "回合：53/400，奖励：-96.0，Epsilon：0.01\n", "回合：54/400，奖励：-189.0，Epsilon：0.01\n", "回合：55/400，奖励：-42.0，Epsilon：0.01\n", "回合：56/400，奖励：-46.0，Epsilon：0.01\n", "回合：57/400，奖励：-85.0，Epsilon：0.01\n", "回合：58/400，奖励：-52.0，Epsilon：0.01\n", "回合：59/400，奖励：-86.0，Epsilon：0.01\n", "回合：60/400，奖励：-41.0，Epsilon：0.01\n", "回合：61/400，奖励：-51.0，Epsilon：0.01\n", "回合：62/400，奖励：-59.0，Epsilon：0.01\n", "回合：63/400，奖励：-145.0，Epsilon：0.01\n", "回合：64/400，奖励：-76.0，Epsilon：0.01\n", "回合：65/400，奖励：-43.0，Epsilon：0.01\n", "回合：66/400，奖励：-49.0，Epsilon：0.01\n", "回合：67/400，奖励：-36.0，Epsilon：0.01\n", "回合：68/400，奖励：-41.0，Epsilon：0.01\n", "回合：69/400，奖励：-69.0，Epsilon：0.01\n", "回合：70/400，奖励：-38.0，Epsilon：0.01\n", "回合：71/400，奖励：-63.0，Epsilon：0.01\n", "回合：72/400，奖励：-46.0，Epsilon：0.01\n", "回合：73/400，奖励：-30.0，Epsilon：0.01\n", "回合：74/400，奖励：-45.0，Epsilon：0.01\n", "回合：75/400，奖励：-38.0，Epsilon：0.01\n", "回合：76/400，奖励：-88.0，Epsilon：0.01\n", "回合：77/400，奖励：-19.0，Epsilon：0.01\n", "回合：78/400，奖励：-40.0，Epsilon：0.01\n", "回合：79/400，奖励：-62.0，Epsilon：0.01\n", "回合：80/400，奖励：-25.0，Epsilon：0.01\n", "回合：81/400，奖励：-54.0，Epsilon：0.01\n", "回合：82/400，奖励：-41.0，Epsilon：0.01\n", "回合：83/400，奖励：-57.0，Epsilon：0.01\n", "回合：84/400，奖励：-52.0，Epsilon：0.01\n", "回合：85/400，奖励：-42.0，Epsilon：0.01\n", "回合：86/400，奖励：-51.0，Epsilon：0.01\n", "回合：87/400，奖励：-53.0，Epsilon：0.01\n", "回合：88/400，奖励：-42.0，Epsilon：0.01\n", "回合：89/400，奖励：-53.0，Epsilon：0.01\n", "回合：90/400，奖励：-31.0，Epsilon：0.01\n", "回合：91/400，奖励：-75.0，Epsilon：0.01\n", "回合：92/400，奖励：-148.0，Epsilon：0.01\n", "回合：93/400，奖励：-41.0，Epsilon：0.01\n", "回合：94/400，奖励：-47.0，Epsilon：0.01\n", "回合：95/400，奖励：-184.0，Epsilon：0.01\n", "回合：96/400，奖励：-34.0，Epsilon：0.01\n", "回合：97/400，奖励：-45.0，Epsilon：0.01\n", "回合：98/400，奖励：-52.0，Epsilon：0.01\n", "回合：99/400，奖励：-44.0，Epsilon：0.01\n", "回合：100/400，奖励：-49.0，Epsilon：0.01\n", "回合：101/400，奖励：-30.0，Epsilon：0.01\n", "回合：102/400，奖励：-49.0，Epsilon：0.01\n", "回合：103/400，奖励：-23.0，Epsilon：0.01\n", "回合：104/400，奖励：-37.0，Epsilon：0.01\n", "回合：105/400，奖励：-37.0，Epsilon：0.01\n", "回合：106/400，奖励：-44.0，Epsilon：0.01\n", "回合：107/400，奖励：-40.0，Epsilon：0.01\n", "回合：108/400，奖励：-28.0，Epsilon：0.01\n", "回合：109/400，奖励：-50.0，Epsilon：0.01\n", "回合：110/400，奖励：-46.0，Epsilon：0.01\n", "回合：111/400，奖励：-28.0，Epsilon：0.01\n", "回合：112/400，奖励：-35.0，Epsilon：0.01\n", "回合：113/400，奖励：-35.0，Epsilon：0.01\n", "回合：114/400，奖励：-45.0，Epsilon：0.01\n", "回合：115/400，奖励：-38.0，Epsilon：0.01\n", "回合：116/400，奖励：-39.0，Epsilon：0.01\n", "回合：117/400，奖励：-27.0，Epsilon：0.01\n", "回合：118/400，奖励：-49.0，Epsilon：0.01\n", "回合：119/400，奖励：-27.0，Epsilon：0.01\n", "回合：120/400，奖励：-25.0，Epsilon：0.01\n", "回合：121/400，奖励：-50.0，Epsilon：0.01\n", "回合：122/400，奖励：-41.0，Epsilon：0.01\n", "回合：123/400，奖励：-22.0，Epsilon：0.01\n", "回合：124/400，奖励：-38.0，Epsilon：0.01\n", "回合：125/400，奖励：-125.0，Epsilon：0.01\n", "回合：126/400，奖励：-25.0，Epsilon：0.01\n", "回合：127/400，奖励：-40.0，Epsilon：0.01\n", "回合：128/400，奖励：-33.0，Epsilon：0.01\n", "回合：129/400，奖励：-56.0，Epsilon：0.01\n", "回合：130/400，奖励：-32.0，Epsilon：0.01\n", "回合：131/400，奖励：-21.0，Epsilon：0.01\n", "回合：132/400，奖励：-33.0，Epsilon：0.01\n", "回合：133/400，奖励：-23.0，Epsilon：0.01\n", "回合：134/400，奖励：-33.0，Epsilon：0.01\n", "回合：135/400，奖励：-34.0，Epsilon：0.01\n", "回合：136/400，奖励：-33.0，Epsilon：0.01\n", "回合：137/400，奖励：-21.0，Epsilon：0.01\n", "回合：138/400，奖励：-40.0，Epsilon：0.01\n", "回合：139/400，奖励：-23.0，Epsilon：0.01\n", "回合：140/400，奖励：-31.0，Epsilon：0.01\n", "回合：141/400，奖励：-31.0，Epsilon：0.01\n", "回合：142/400，奖励：-26.0，Epsilon：0.01\n", "回合：143/400，奖励：-26.0，Epsilon：0.01\n", "回合：144/400，奖励：-32.0，Epsilon：0.01\n", "回合：145/400，奖励：-27.0，Epsilon：0.01\n", "回合：146/400，奖励：-33.0，Epsilon：0.01\n", "回合：147/400，奖励：-35.0，Epsilon：0.01\n", "回合：148/400，奖励：-21.0，Epsilon：0.01\n", "回合：149/400，奖励：-23.0，Epsilon：0.01\n", "回合：150/400，奖励：-33.0，Epsilon：0.01\n", "回合：151/400，奖励：-25.0，Epsilon：0.01\n", "回合：152/400，奖励：-41.0，Epsilon：0.01\n", "回合：153/400，奖励：-31.0，Epsilon：0.01\n", "回合：154/400，奖励：-28.0，Epsilon：0.01\n", "回合：155/400，奖励：-133.0，Epsilon：0.01\n", "回合：156/400，奖励：-22.0，Epsilon：0.01\n", "回合：157/400，奖励：-21.0，Epsilon：0.01\n", "回合：158/400，奖励：-33.0，Epsilon：0.01\n", "回合：159/400，奖励：-33.0，Epsilon：0.01\n", "回合：160/400，奖励：-24.0，Epsilon：0.01\n", "回合：161/400，奖励：-34.0，Epsilon：0.01\n", "回合：162/400，奖励：-20.0，Epsilon：0.01\n", "回合：163/400，奖励：-21.0，Epsilon：0.01\n", "回合：164/400，奖励：-126.0，Epsilon：0.01\n", "回合：165/400，奖励：-36.0，Epsilon：0.01\n", "回合：166/400，奖励：-18.0，Epsilon：0.01\n", "回合：167/400，奖励：-35.0，Epsilon：0.01\n", "回合：168/400，奖励：-26.0，Epsilon：0.01\n", "回合：169/400，奖励：-24.0，Epsilon：0.01\n", "回合：170/400，奖励：-33.0，Epsilon：0.01\n", "回合：171/400，奖励：-17.0，Epsilon：0.01\n", "回合：172/400，奖励：-23.0，Epsilon：0.01\n", "回合：173/400，奖励：-26.0，Epsilon：0.01\n", "回合：174/400，奖励：-23.0，Epsilon：0.01\n", "回合：175/400，奖励：-21.0，Epsilon：0.01\n", "回合：176/400，奖励：-35.0，Epsilon：0.01\n", "回合：177/400，奖励：-26.0，Epsilon：0.01\n", "回合：178/400，奖励：-17.0，Epsilon：0.01\n", "回合：179/400，奖励：-20.0，Epsilon：0.01\n", "回合：180/400，奖励：-28.0，Epsilon：0.01\n", "回合：181/400，奖励：-34.0，Epsilon：0.01\n", "回合：182/400，奖励：-27.0，Epsilon：0.01\n", "回合：183/400，奖励：-22.0，Epsilon：0.01\n", "回合：184/400，奖励：-24.0，Epsilon：0.01\n", "回合：185/400，奖励：-26.0，Epsilon：0.01\n", "回合：186/400，奖励：-20.0，Epsilon：0.01\n", "回合：187/400，奖励：-30.0，Epsilon：0.01\n", "回合：188/400，奖励：-28.0，Epsilon：0.01\n", "回合：189/400，奖励：-15.0，Epsilon：0.01\n", "回合：190/400，奖励：-30.0，Epsilon：0.01\n", "回合：191/400，奖励：-29.0，Epsilon：0.01\n", "回合：192/400，奖励：-22.0，Epsilon：0.01\n", "回合：193/400，奖励：-25.0，Epsilon：0.01\n", "回合：194/400，奖励：-21.0，Epsilon：0.01\n", "回合：195/400，奖励：-19.0，Epsilon：0.01\n", "回合：196/400，奖励：-23.0，Epsilon：0.01\n", "回合：197/400，奖励：-21.0，Epsilon：0.01\n", "回合：198/400，奖励：-32.0，Epsilon：0.01\n", "回合：199/400，奖励：-30.0，Epsilon：0.01\n", "回合：200/400，奖励：-22.0，Epsilon：0.01\n", "回合：201/400，奖励：-20.0，Epsilon：0.01\n", "回合：202/400，奖励：-27.0，Epsilon：0.01\n", "回合：203/400，奖励：-21.0，Epsilon：0.01\n", "回合：204/400，奖励：-26.0，Epsilon：0.01\n", "回合：205/400，奖励：-19.0，Epsilon：0.01\n", "回合：206/400，奖励：-17.0，Epsilon：0.01\n", "回合：207/400，奖励：-31.0，Epsilon：0.01\n", "回合：208/400，奖励：-18.0，Epsilon：0.01\n", "回合：209/400，奖励：-24.0，Epsilon：0.01\n", "回合：210/400，奖励：-17.0，Epsilon：0.01\n", "回合：211/400，奖励：-26.0，Epsilon：0.01\n", "回合：212/400，奖励：-27.0，Epsilon：0.01\n", "回合：213/400，奖励：-33.0，Epsilon：0.01\n", "回合：214/400，奖励：-16.0，Epsilon：0.01\n", "回合：215/400，奖励：-32.0，Epsilon：0.01\n", "回合：216/400，奖励：-19.0，Epsilon：0.01\n", "回合：217/400，奖励：-20.0，Epsilon：0.01\n", "回合：218/400，奖励：-15.0，Epsilon：0.01\n", "回合：219/400，奖励：-119.0，Epsilon：0.01\n", "回合：220/400，奖励：-26.0，Epsilon：0.01\n", "回合：221/400，奖励：-26.0，Epsilon：0.01\n", "回合：222/400，奖励：-22.0，Epsilon：0.01\n", "回合：223/400，奖励：-22.0，Epsilon：0.01\n", "回合：224/400，奖励：-15.0，Epsilon：0.01\n", "回合：225/400，奖励：-24.0，Epsilon：0.01\n", "回合：226/400，奖励：-15.0，Epsilon：0.01\n", "回合：227/400，奖励：-31.0，Epsilon：0.01\n", "回合：228/400，奖励：-24.0，Epsilon：0.01\n", "回合：229/400，奖励：-20.0，Epsilon：0.01\n", "回合：230/400，奖励：-20.0，Epsilon：0.01\n", "回合：231/400，奖励：-22.0，Epsilon：0.01\n", "回合：232/400，奖励：-15.0，Epsilon：0.01\n", "回合：233/400，奖励：-19.0，Epsilon：0.01\n", "回合：234/400，奖励：-21.0，Epsilon：0.01\n", "回合：235/400，奖励：-27.0，Epsilon：0.01\n", "回合：236/400，奖励：-15.0，Epsilon：0.01\n", "回合：237/400，奖励：-25.0，Epsilon：0.01\n", "回合：238/400，奖励：-22.0，Epsilon：0.01\n", "回合：239/400，奖励：-16.0，Epsilon：0.01\n", "回合：240/400，奖励：-18.0，Epsilon：0.01\n", "回合：241/400，奖励：-13.0，Epsilon：0.01\n", "回合：242/400，奖励：-13.0，Epsilon：0.01\n", "回合：243/400，奖励：-13.0，Epsilon：0.01\n", "回合：244/400，奖励：-23.0，Epsilon：0.01\n", "回合：245/400，奖励：-29.0，Epsilon：0.01\n", "回合：246/400，奖励：-26.0，Epsilon：0.01\n", "回合：247/400，奖励：-19.0，Epsilon：0.01\n", "回合：248/400，奖励：-21.0，Epsilon：0.01\n", "回合：249/400，奖励：-17.0，Epsilon：0.01\n", "回合：250/400，奖励：-17.0，Epsilon：0.01\n", "回合：251/400，奖励：-15.0，Epsilon：0.01\n", "回合：252/400，奖励：-20.0，Epsilon：0.01\n", "回合：253/400，奖励：-23.0，Epsilon：0.01\n", "回合：254/400，奖励：-19.0，Epsilon：0.01\n", "回合：255/400，奖励：-21.0，Epsilon：0.01\n", "回合：256/400，奖励：-19.0，Epsilon：0.01\n", "回合：257/400，奖励：-17.0，Epsilon：0.01\n", "回合：258/400，奖励：-17.0，Epsilon：0.01\n", "回合：259/400，奖励：-15.0，Epsilon：0.01\n", "回合：260/400，奖励：-21.0，Epsilon：0.01\n", "回合：261/400，奖励：-17.0，Epsilon：0.01\n", "回合：262/400，奖励：-19.0，Epsilon：0.01\n", "回合：263/400，奖励：-19.0，Epsilon：0.01\n", "回合：264/400，奖励：-15.0，Epsilon：0.01\n", "回合：265/400，奖励：-19.0，Epsilon：0.01\n", "回合：266/400，奖励：-17.0，Epsilon：0.01\n", "回合：267/400，奖励：-15.0，Epsilon：0.01\n", "回合：268/400，奖励：-19.0，Epsilon：0.01\n", "回合：269/400，奖励：-27.0，Epsilon：0.01\n", "回合：270/400，奖励：-15.0，Epsilon：0.01\n", "回合：271/400，奖励：-17.0，Epsilon：0.01\n", "回合：272/400，奖励：-17.0，Epsilon：0.01\n", "回合：273/400，奖励：-25.0，Epsilon：0.01\n", "回合：274/400，奖励：-19.0，Epsilon：0.01\n", "回合：275/400，奖励：-22.0，Epsilon：0.01\n", "回合：276/400，奖励：-23.0，Epsilon：0.01\n", "回合：277/400，奖励：-18.0，Epsilon：0.01\n", "回合：278/400，奖励：-23.0，Epsilon：0.01\n", "回合：279/400，奖励：-21.0，Epsilon：0.01\n", "回合：280/400，奖励：-21.0，Epsilon：0.01\n", "回合：281/400，奖励：-21.0，Epsilon：0.01\n", "回合：282/400，奖励：-19.0，Epsilon：0.01\n", "回合：283/400，奖励：-18.0，Epsilon：0.01\n", "回合：284/400，奖励：-15.0，Epsilon：0.01\n", "回合：285/400，奖励：-19.0，Epsilon：0.01\n", "回合：286/400，奖励：-19.0，Epsilon：0.01\n", "回合：287/400，奖励：-21.0，Epsilon：0.01\n", "回合：288/400，奖励：-15.0，Epsilon：0.01\n", "回合：289/400，奖励：-32.0，Epsilon：0.01\n", "回合：290/400，奖励：-18.0，Epsilon：0.01\n", "回合：291/400，奖励：-17.0，Epsilon：0.01\n", "回合：292/400，奖励：-15.0，Epsilon：0.01\n", "回合：293/400，奖励：-24.0，Epsilon：0.01\n", "回合：294/400，奖励：-22.0，Epsilon：0.01\n", "回合：295/400，奖励：-31.0，Epsilon：0.01\n", "回合：296/400，奖励：-17.0，Epsilon：0.01\n", "回合：297/400，奖励：-19.0，Epsilon：0.01\n", "回合：298/400，奖励：-19.0，Epsilon：0.01\n", "回合：299/400，奖励：-20.0，Epsilon：0.01\n", "回合：300/400，奖励：-21.0，Epsilon：0.01\n", "回合：301/400，奖励：-26.0，Epsilon：0.01\n", "回合：302/400，奖励：-20.0，Epsilon：0.01\n", "回合：303/400，奖励：-16.0，Epsilon：0.01\n", "回合：304/400，奖励：-20.0，Epsilon：0.01\n", "回合：305/400，奖励：-21.0，Epsilon：0.01\n", "回合：306/400，奖励：-16.0，Epsilon：0.01\n", "回合：307/400，奖励：-19.0，Epsilon：0.01\n", "回合：308/400，奖励：-24.0，Epsilon：0.01\n", "回合：309/400，奖励：-20.0，Epsilon：0.01\n", "回合：310/400，奖励：-17.0，Epsilon：0.01\n", "回合：311/400，奖励：-16.0，Epsilon：0.01\n", "回合：312/400，奖励：-25.0，Epsilon：0.01\n", "回合：313/400，奖励：-16.0，Epsilon：0.01\n", "回合：314/400，奖励：-19.0，Epsilon：0.01\n", "回合：315/400，奖励：-19.0，Epsilon：0.01\n", "回合：316/400，奖励：-27.0，Epsilon：0.01\n", "回合：317/400，奖励：-15.0，Epsilon：0.01\n", "回合：318/400，奖励：-15.0，Epsilon：0.01\n", "回合：319/400，奖励：-15.0，Epsilon：0.01\n", "回合：320/400，奖励：-19.0，Epsilon：0.01\n", "回合：321/400，奖励：-23.0，Epsilon：0.01\n", "回合：322/400，奖励：-24.0，Epsilon：0.01\n", "回合：323/400，奖励：-15.0，Epsilon：0.01\n", "回合：324/400，奖励：-20.0，Epsilon：0.01\n", "回合：325/400，奖励：-18.0，Epsilon：0.01\n", "回合：326/400，奖励：-19.0，Epsilon：0.01\n", "回合：327/400，奖励：-19.0，Epsilon：0.01\n", "回合：328/400，奖励：-26.0，Epsilon：0.01\n", "回合：329/400，奖励：-16.0，Epsilon：0.01\n", "回合：330/400，奖励：-18.0，Epsilon：0.01\n", "回合：331/400，奖励：-15.0，Epsilon：0.01\n", "回合：332/400，奖励：-15.0，Epsilon：0.01\n", "回合：333/400，奖励：-17.0，Epsilon：0.01\n", "回合：334/400，奖励：-17.0，Epsilon：0.01\n", "回合：335/400，奖励：-16.0，Epsilon：0.01\n", "回合：336/400，奖励：-24.0，Epsilon：0.01\n", "回合：337/400，奖励：-15.0，Epsilon：0.01\n", "回合：338/400，奖励：-18.0，Epsilon：0.01\n", "回合：339/400，奖励：-16.0，Epsilon：0.01\n", "回合：340/400，奖励：-15.0，Epsilon：0.01\n", "回合：341/400，奖励：-18.0，Epsilon：0.01\n", "回合：342/400，奖励：-15.0，Epsilon：0.01\n", "回合：343/400，奖励：-20.0，Epsilon：0.01\n", "回合：344/400，奖励：-18.0，Epsilon：0.01\n", "回合：345/400，奖励：-17.0，Epsilon：0.01\n", "回合：346/400，奖励：-19.0，Epsilon：0.01\n", "回合：347/400，奖励：-15.0，Epsilon：0.01\n", "回合：348/400，奖励：-15.0，Epsilon：0.01\n", "回合：349/400，奖励：-15.0，Epsilon：0.01\n", "回合：350/400，奖励：-18.0，Epsilon：0.01\n", "回合：351/400，奖励：-16.0，Epsilon：0.01\n", "回合：352/400，奖励：-16.0，Epsilon：0.01\n", "回合：353/400，奖励：-15.0，Epsilon：0.01\n", "回合：354/400，奖励：-20.0，Epsilon：0.01\n", "回合：355/400，奖励：-15.0，Epsilon：0.01\n", "回合：356/400，奖励：-17.0，Epsilon：0.01\n", "回合：357/400，奖励：-15.0，Epsilon：0.01\n", "回合：358/400，奖励：-17.0，Epsilon：0.01\n", "回合：359/400，奖励：-15.0，Epsilon：0.01\n", "回合：360/400，奖励：-16.0，Epsilon：0.01\n", "回合：361/400，奖励：-15.0，Epsilon：0.01\n", "回合：362/400，奖励：-18.0，Epsilon：0.01\n", "回合：363/400，奖励：-17.0，Epsilon：0.01\n", "回合：364/400，奖励：-22.0，Epsilon：0.01\n", "回合：365/400，奖励：-15.0，Epsilon：0.01\n", "回合：366/400，奖励：-15.0，Epsilon：0.01\n", "回合：367/400，奖励：-15.0，Epsilon：0.01\n", "回合：368/400，奖励：-16.0，Epsilon：0.01\n", "回合：369/400，奖励：-16.0，Epsilon：0.01\n", "回合：370/400，奖励：-15.0，Epsilon：0.01\n", "回合：371/400，奖励：-20.0，Epsilon：0.01\n", "回合：372/400，奖励：-15.0，Epsilon：0.01\n", "回合：373/400，奖励：-15.0，Epsilon：0.01\n", "回合：374/400，奖励：-15.0，Epsilon：0.01\n", "回合：375/400，奖励：-16.0，Epsilon：0.01\n", "回合：376/400，奖励：-15.0，Epsilon：0.01\n", "回合：377/400，奖励：-15.0，Epsilon：0.01\n", "回合：378/400，奖励：-17.0，Epsilon：0.01\n", "回合：379/400，奖励：-20.0，Epsilon：0.01\n", "回合：380/400，奖励：-17.0，Epsilon：0.01\n", "回合：381/400，奖励：-15.0，Epsilon：0.01\n", "回合：382/400，奖励：-15.0，Epsilon：0.01\n", "回合：383/400，奖励：-15.0，Epsilon：0.01\n", "回合：384/400，奖励：-15.0，Epsilon：0.01\n", "回合：385/400，奖励：-16.0，Epsilon：0.01\n", "回合：386/400，奖励：-15.0，Epsilon：0.01\n", "回合：387/400，奖励：-18.0，Epsilon：0.01\n", "回合：388/400，奖励：-15.0，Epsilon：0.01\n", "回合：389/400，奖励：-15.0，Epsilon：0.01\n", "回合：390/400，奖励：-15.0，Epsilon：0.01\n", "回合：391/400，奖励：-16.0，Epsilon：0.01\n", "回合：392/400，奖励：-18.0，Epsilon：0.01\n", "回合：393/400，奖励：-15.0，Epsilon：0.01\n", "回合：394/400，奖励：-15.0，Epsilon：0.01\n", "回合：395/400，奖励：-15.0，Epsilon：0.01\n", "回合：396/400，奖励：-20.0，Epsilon：0.01\n", "回合：397/400，奖励：-15.0，Epsilon：0.01\n", "回合：398/400，奖励：-15.0，Epsilon：0.01\n", "回合：399/400，奖励：-15.0，Epsilon：0.01\n", "回合：400/400，奖励：-15.0，Epsilon：0.01\n", "完成训练！\n" ] }, { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "开始测试！\n", "环境：CliffWalking-v0, 算法：Sarsa, 设备：cpu\n", "回合数：1/20, 奖励：-15.0\n", "回合数：2/20, 奖励：-15.0\n", "回合数：3/20, 奖励：-15.0\n", "回合数：4/20, 奖励：-15.0\n", "回合数：5/20, 奖励：-15.0\n", "回合数：6/20, 奖励：-15.0\n", "回合数：7/20, 奖励：-15.0\n", "回合数：8/20, 奖励：-15.0\n", "回合数：9/20, 奖励：-15.0\n", "回合数：10/20, 奖励：-15.0\n", "回合数：11/20, 奖励：-15.0\n", "回合数：12/20, 奖励：-15.0\n", "回合数：13/20, 奖励：-15.0\n", "回合数：14/20, 奖励：-15.0\n", "回合数：15/20, 奖励：-15.0\n", "回合数：16/20, 奖励：-15.0\n", "回合数：17/20, 奖励：-15.0\n", "回合数：18/20, 奖励：-15.0\n", "回合数：19/20, 奖励：-15.0\n", "回合数：20/20, 奖励：-15.0\n", "完成测试！\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEXCAYAAACgUUN5AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA4dUlEQVR4nO3deVxU9f7H8dcMKKKQpeKSt1CuW2qp1x0hwgUdFsXSm5WokKJmlnotxSUVJZIsJbRyuUrumguYiku4homYpdii5m6iCLhggMDM9/eHP+c6ggozCtR8no+Hj4dzzvme85nvzPA+53tmztEopRRCCCGslra0CxBCCFG6JAiEEMLKSRAIIYSVkyAQQggrJ0EghBBWToJACCGsnATBYxIUFERGRobZ7SdMmMDRo0cBGD9+PPv27XtUpQkzpaSk4OvrS/fu3fnxxx9N5t28eZMJEybg5+dH9+7d8ff35+uvvy6VOidNmkTHjh2ZOXOmRes5dOgQb775Jj169MDX15fg4GCOHz8OQGJiIr6+vgBERkYSExMDwLp163jppZd48803+e677/D09MTf358mTZrw66+/Gte9atUqGjZsyN69e43T4uLi6NWr1wNratiwIRkZGaxbt47BgwcXmB8fH8+0adMset5F9eWXX9KtWze6dOlCVFQUf+lv4ivxWDRo0EClp6eb3d7T01MdOXLkEVYkLLV+/XrVv3//QudNnjxZhYWFKYPBoJRS6tKlS8rDw0Pt3bu3BCu8rWHDhiolJcWidRw4cEB5eHio5ORk47TY2FjVpk0blZ6ervbv3698fHwKtAsICFAxMTFKKaXGjh2r5syZo5RSatiwYWrRokXG5QYPHqwGDx6sJk+ebJw2ceJEFRUV9cC67nyu1q5dq4KDgy15ihbZtWuX6tGjh/rzzz9VTk6OeuONN9SmTZtKrR5L2ZZ2EP0dhYSEANC/f3/mzZuHVqslNDSUlJQU8vLy8PHxYciQIeTn5zN16lQOHTpEuXLl+Mc//kF4eDjz5s0jNTWV0aNHExERwYwZM3jjjTdo2rQpAwYMwMPDg8OHD3P9+nVGjhyJt7c32dnZTJo0icOHD+Po6Ei9evUA+Oijj0xqy8/P5+OPP2bXrl3Y2NjQokULJk2axNy5c7l69SoffPABAFFRUcbHAQEBVK5cmVOnTvHqq6/y+eefs3fvXsqXL49er8fT05OFCxdSo0YNwsLCOH78OHl5ebRv3573338fW1vTt1lmZiZTpkzht99+Q6PR4O7uzqhRo7C1teX5558nODiYhIQEUlNT6devHwMGDCjQx4cPH2batGlkZ2dTrlw53n//fdq3b0/jxo3p378/iYmJZGVlMWrUKLy8vFi3bh1bt25l7ty5AAUe323VqlUsWbIErVZLtWrVmDhxIpcvX2bWrFlkZmYSEBDAkiVLTNpcuXKFqlWrkpeXR/ny5alRowZRUVE8+eSTAOzcuZO5c+eSm5tLRkYG/v7+jBgxgsTERMLCwqhYsSJZWVksW7aM8ePHc/bsWbRaLU2aNCE0NBSADz/8kMOHD/Pnn3+ilGLatGm0bNnSpI7XX38dpRSDBg1i0qRJVK5cmdDQUK5du4ZGoyEoKAh/f/8C212zZg3ly5c3ruezzz7jrbfeomnTpsZp3bt3x87ODr1eb7LNsWPHUr9+fS5fvkxycjIXLlzgypUrxMfHY2dnR2ZmJi+++CK7du1iwIAB5OTkcOTIERYvXszAgQOZNGkSAPv37+eTTz7h9OnThIaGkpWVRWpqKo0aNWLWrFnY2dkVeK0AtmzZwowZM5g3bx4//fST8XUNCAigefPmHDp0iJSUFFq2bMn06dPRarWsW7eOefPmUaFCBdq1a8fixYv55ZdfTNZrMBjw9PRk9uzZPP/88wCMHDmS1q1b88svv+Dr60vFihUBePnll9mwYQPe3t6F1ljmlXYS/V3dfUQQEBCg4uPjlVJK5eTkqICAALVp0yaVlJSkunXrZtyLjIiIUD/88INSyvSIoG/fviouLk6dP39eNWjQQO3YsUMppdSWLVvUSy+9pJRSasaMGWrUqFFKr9erzMxM5efnp8aMGVOgrq+++kq98cYbKjs7W+n1evXuu++q9evXq88++0xNmTLFuNzdj/v27atCQkKM89544w0VFxenlLq9Z9SnTx+l1O09wMWLFyullMrPz1ejR49W8+bNK1DD+++/r6ZOnaoMBoO6deuWCgoKUnPnzjX225IlS5RSSiUnJ6umTZuqnJwck/a5ubmqQ4cOaufOncblfH19lV6vVw0aNFBffPGFUkqpX3/9VbVs2bLQPcj77VHu27dPde7c2fjarV27Vul0OmUwGB64F/rrr78qLy8v1aJFCxUUFKRmz56tTp06pZRSymAwqL59+6rTp08rpW4fLTz33HPGPetGjRqpCxcuKKVuH3UEBQUZ+3D8+PHqzJkz6tChQ2r48OFKr9crpZSaO3euGjx4cKG13Hnv5eXlqU6dOqmtW7cat+vu7q4OHTpUYLv3at68uTpx4kSh85RSJkcEY8aMUQsWLFBK/e+9eu/0lJQU1aZNG6XX61V8fLwaNmyYUkqprl27qp9//ln98ccfytXVVRkMBvXRRx8Zjypyc3OVr6+v2rJli8lzu/NabNiwQfn4+KiLFy8aX687r1Hfvn3VO++8Y/xMuLm5qe+//16dOHFCtW/f3njUFBUVpRo0aFDo84yMjDR+Dq5du6batGmjbty4oYKCgtTGjRuNyyUkJCh/f//79ldZJ0cEj1lWVhZJSUlcv36dyMhI47TffvsNNzc3bGxs6N27N25ubnTt2pUXXnjhgesrV64cHh4eADRu3Jhr164BsHv3bkJCQtBqtTg4ONCzZ0+OHTtWoP2+ffvo0aMHFSpUAGDWrFnA7SOAB2nVqpXx/71792b9+vV069aNdevW0bt3bwB27dpFcnIya9asASAnJ6fQde3Zs4cVK1ag0WgoX748ffr04auvviI4OBiATp06AdCkSRNyc3PJysoy2Rs8fvw4Wq2Wl156CYCmTZvyzTffGOf37dsXgEaNGtGgQQOSkpIe+NzutnfvXry9valSpQpwe08vLCyMCxcuPLBdo0aN2LJlCz///DNJSUkkJCTw5ZdfEhkZSceOHfnyyy/ZtWsXGzdu5OTJkyilyM7OBqBWrVrUrl0bgJYtWzJz5kwCAgJwdXWlf//+ODs74+zsTOXKlVm5ciXnz58nMTGRSpUqPbCmM2fOcOvWLby8vACoUaMGXl5e7N27l7Zt25ps915arRaDwVDkfnuYmjVrUr16dY4dO8bOnTuNr52npyffffcd1apV48UXX0Sj0fDee++RkJDA/PnzOXPmDKmpqWRlZRVYZ3JyMnv37mXcuHHUqlWr0O16enoaPxPOzs5cv36d3377jQ4dOlCzZk3g9vvlfu//V155hV69ejF27Fg2btyIp6cnjo6OhZ4P0Gr/uqdcJQgeM4PBgFKKlStXYm9vD0BGRgZ2dnZUqlSJ2NhYDh06xP79+xkxYsR9h0LuKFeunPENp9FojNNtbW1N3pz3e1PeO0yTlpaGwWBAo9GYtM/LyzNZ7s4hMEC3bt0IDw/n5MmTJCUlGYefDAYDkZGR/POf/wTgxo0bJjXe3Sf3Ps7Pzzc+vvNH/07bez90NjY2BdZ7/PhxXFxcjPPvXved5R/0/O4o7AOulDKp7175+flMmTKF//znPzRt2pSmTZsSGBjI559/zqpVq2jXrh09e/akc+fOtGrVildeeYVvv/3WuK27+/aZZ55h+/btJCYmsn//fgIDA5kwYQIVKlQgLCyMwMBAOnXqhIuLCxs2bLhvTXee+4Oey93bvVfz5s05fPgwDRo0MJk+ZcoUunTpYtLHReXu7s6BAwfYvXs37777LgAeHh5ER0fzxBNPGHcARo0ahV6vR6fT8dJLL5GSklLo6+Lo6Mgnn3zCiBEjeOmll/jHP/5RYJk7OzyA8T1gY2Njsr67n8ugQYNITU0F4J133qFTp040btyYXbt2sW7dOsaNGwfcDu8rV64Y212+fNkYLH9Ff90IK+NsbGzIz8/HwcGB5s2bs2jRIuD2H8fXXnuN+Ph4du7cyYABA2jRogXDhw/H39+f3377zaR9UXl4eLB27VoMBgPZ2dls3Lix0D/C7du3Z+PGjeTm5mIwGJg8eTKbNm3iqaee4ueff0YpRVZWFt999919t2VnZ4ePjw9jx47Fy8vLGHBubm5ER0ejlCI3N5ehQ4eydOnSAu3d3NxYtmyZcbnVq1fj6upa5Ofq4uKCRqMhISEBgJ9//pn+/fsb//Dd+QbLzz//zOnTp2ndujVVqlThxIkT3Lp1i/z8fHbu3Fnout3c3Ni8ebPxG19r167lySefxNnZ+b712NracubMGT7//HNjwOTn53P+/HkaN27M2bNnuXnzJiNGjKBjx44cOHDA2P/3Wr58OSEhIbi5ufHee+/h5ubGiRMnSEhIwNPTk9dff53nn3+eb7/9tsBY/b3q1q1LuXLl2LZtG3D7j9XWrVuL1NdDhw5l9uzZxm+uwf/Oq9wbDkX14osvsnbtWqpXr061atWA20eax48f58cff6RDhw4AfPfddwwbNgxvb280Gg2HDx8u9LnWqVOH9u3bExAQwJgxY4p8BOPm5sb333/P5cuXAUy+3TV//nxiY2OJjY01BtO///1v5s+fT05OjvGcTKdOndiwYQNZWVnk5uaybt06OnfubFa/lAVyRPCYdOnShddff53PP/+cGTNmMHXqVPz8/MjNzTV+BVGv17Nnzx7jSafKlSszdepUADp37szIkSOL/FW4wYMHExoaip+fH46OjlStWtVkb+iOPn368Mcff/Dyyy+jlKJNmzYEBASQnZ3N3r178fLyokaNGrRo0eKBX4fr3bs3S5cuZfLkycZp48ePJywsDD8/P/Ly8nB1dWXgwIEF2k6YMIFp06YZl3N3d2fIkCFFep4A5cuXJyoqig8//JCIiAjKlStHVFSU8WTnoUOHWL16NQaDgZkzZ1K5cmU6dOhA69at0el0ODk50bZt20KHzjp06MCAAQOMwVKlShXmzp370MP+yMhIPv74Y7p27Yq9vT1KKTp37sywYcOMw1g6nY4nnniCZ599lnr16nH27FmTE7QA/v7+HDhwAG9vb+zt7Xn66afp168faWlpjB49Gj8/P2xsbGjVqhXbtm3DYDDct7Zy5crx+eefM23aNKKiotDr9QwbNox27dqRmJj4wOfTqlUrpk2bRlhYGFlZWeTl5fHss8+yePFiqlWrxsmTJx/YvjAtW7bkwoULBAUFGafd+YLAtWvXcHBwAG6fkB02bBiVK1fG3t6e1q1bc+7cufuud8iQIezYsYMFCxYYA+ZB6tatS0hICG+++Sbly5fnueeeM+7MFKZjx45MmTKFQYMGmUw7fvw4vXv3Ji8vj06dOuHv71+EXiibNOpBn3bxl7Fp0yYcHBzw8PDAYDAwfPhwOnTowOuvv17apZWohg0b8v333xvH+IW41/nz54mNjeWtt95Cq9Wybds25s+fX2q/+ygL5Ijgb6J+/fp88MEHfPrpp+Tl5dG2bVvjSVwhxP/UrFmT1NRU49GVo6MjH374YWmXVarkiEAIIaycnCwWQggrJ0EghBBWToJACCGsnMVBEBkZWeiv8i5dukSbNm0K/UWmXq9n0qRJ+Pr64uPjQ3R0tKVlCCGEMJPZ3xrKzMwkPDycTZs2FfiuuMFgYPz48ff99ea6deu4du0aGzZsICcnh169etG6dWuaNGlS5O1fvfonBkPxz3NXrepAevrNYrcrKVKfZaQ+y0h9linL9Wm1Gp56qvDLkpgdBPHx8dSpU4fAwMAC8xYsWICrqyunT58utG39+vVp3rw5Wq2WihUr8swzz5CSklKsIDAYlFlBcKdtWSb1WUbqs4zUZ5myXl9hzA6CO7+iu3dY6OjRoyQmJjJ//nyWLVtWaNvmzZsb/3/o0CGOHDlCREREsbZftapDsZa/m5OTo9ltS4LUZxmpzzJSn2XKen2FeWgQxMXFER4ebjLNxcWl0HH97OxsQkNDmTVrVpGuxHfgwAFGjRrFjBkzqFy5ctGrBtLTb5qVvE5Ojly5klnsdiVF6rOM1GcZqc8yZbk+rVZz3x3ohwaBTqdDp9MVaUMHDx4kLS2NoUOHApCamkpwcDCzZ882Xhnyjm3btjF58mRmzpxJ27Zti7R+IcTfg1KKq1evkJubAxR9hy419dFeHvtRK/36NJQvX4GnnnIq9KKT9/NILzHh7u7Ojh07jI87duzIvHnzClwe9siRI0yePJmFCxfSqFGjR1mCEOIv4ObN62g0GmrU+AcaTdG/vGhrqyU/v+wGQWnXp5SBa9fSuHnzOo6OTxa5XYn9jiA5Odl49b4vvvgCvV7PmDFj6NGjBz169CA+Pr6kShFClLLs7Js4Oj5ZrBAQD6fRaHF0fIrs7OJ9c+kve60hOUdQOqQ+y0h9t126dJYaNZ4t1vAFlP4e98OUhfqUUly+fI6aNU3vofGgcwQSx0KIUlHcEBBFY06/ShAIIUQp6tXLj5SUi6VagwSBEEJYObkxjRDC6h06dJAvvvgMvd5ArVq1sLevyKlTJzEYDLzxRj86duxCjx7dWL06hooVKzF0aBAdOrxI374D+Pbbrfz0048MHfo206dP4/Lly6SlXaF58xZMmBDKjz/+YFy3i8s/eeedUYSGTiQ19TJ16riQm5sLwO+/nyAiIgy9Xk/58uUZN24SzzzzbIk8fwkCIUSpS0hO4bsjKQ9dTqOB4n69xe2FWnR4vtZDlzt//hxr1mxkyZJFVKvmxIQJU/jzz5sMGRJE48ZNadmyFT/+eIgWLVqSkpLCTz8dom/fAezfv49Onbqwb9931K/fgNDQj8jLy6Nv394cO/abybodHBz49NPpNGjQiBkzPuOnnw6xY8d2AFavXk6fPn3p2LEz8fHb+PnnZAkCIYQoSc8844yDgwMHDx7g1q0cNm3aAEBOTg6nT5+ifXs3fvjhAFqtBi8vHfHx28jPz+fw4Z94771x2NnZcezYL6xevZwzZ05z/fp1srOzTNYN8OOPPzB58u1bYzZv/i+efro2AO3bd+DTTyNITNyHq6s7L73UqcSeuwSBEKLUdXi+aHvtj/PrmXZ2dgAYDHomTpxKw4a3f+yakZHOE09UJjMzk5Url2FjY0vLlq05d+4MGzfG4OLigp2dHWvWrGT37h34+fWkV682nD59kjvfzr+zbrj9rZ67f31sY2MDgKdnZ5o2fYGEhL18/fUK9u9PYMyYCY/lud5LThYLIcRd/vWv1sTErAEgLS2N/v1f4/LlSzz11FPY2dmRkLCHF15ozr/+1Zro6P/i6uoOQFJSIv7+r+DlpQM0nDhxvNDLTbRq1YZt2+IA+PXXn/njj9v3bPnggxB++eVn/P1fYeDAIcZhpZIgQSCEEHcJChrErVu3CAj4N+++O4S33nqH2rVvXyanffsOODg4UrFiRVq2bE1a2hVcXd0A+Pe/X+e//51HUNAbfPrpdJo2faHQr4W++eZg/vjjAn37/pulS6ONQ0MBAYEsWbKIoKA3mDNnFsOHjyyx5yy/LC5jpD7LSH2WKclfFt/7y9eiKAu/3H2QslJfYf0rvywWQghxXxIEQghh5SQIhBDCykkQCCGElZMgEEIIKydBIIQQVk6CQAghrJzFQRAZGUlUVFSB6ZcuXaJNmzZcuHDhge3feeedQtsLIcTfzcWLfxAeHgrcvuLp228Hm72uzZu/ISxs8iOpy+wgyMzMZNy4cSxcuLDAPIPBwPjx48nLy3vgOtasWUNiYqK5JQghxF/KpUspxktKlCVmX3QuPj6eOnXqEBgYWGDeggULcHV15fTp0/dtf/bsWdavX0+fPn3MLUEI8TeRdzyBvGN7HrqcRqOhuBdDKNfwRco16PDAZVJTLxMaOpHs7Gy0Wg3vvvsekyePo2PH25eXtrGxYfDgYaxcuZQLF84zbNgIOnXqQkZGOh99NJXLly9hY2PD0KFv07p1e3Jycpg+fRq//34crVZLnz590el8iYycwcWLf/DJJ9Px9OzEtWvXGD36Hf744wLPPuvM1KnTKV++PHFxG/n66xUYDIqGDRsxatQY7Ozs2LJlE1999V8qVXKgZs2a2NtXLFZf3I/ZRwT+/v4EBwcbr5x3x9GjR0lMTCw0IO7Iz89nwoQJTJkyBVtbuQCqEKJ0bdwYi6urG//97xKGDn2HI0d+AqBaNSeWLl1Nw4aNWLo0mk8/nc3EiaEsXboIgJkzP+Zf/2rFV1+tZOrU6YSFTSEjI52FC+dSuXJllixZTWTklyxcOJ/ffz/Bu++OpmHD5/jPf8YAcPnyJUaNGsOyZWvIyEjn4MEDnDp1km++ieGLLxYSHb2cp56qwooVS0hLu8IXX3zGnDnz+fLLhWRlZT2y5//Qv8JxcXGEh4ebTHNxcSE6OrrAstnZ2YSGhjJr1iy02vtnTFRUFF26dKFevXrFr/j/3e+aGUXh5ORodtuSIPVZRuqzTEnUl5qqxdb2f38jbBu7Y9/Y/bFv937atm3H2LGj+f3347i6uvHqq31Yt241bm5u2NpqqVWrFtWrV6dChfLUrl2bzMxMbG21HDp0kPHjJ2Jrq8XZ+VmaNHme33775f+nT8LWVku1alXw8PDg8OFD1KtXH41Gg62tFhsbLfXrN+DZZ58BoG5dFzIzr3P58kUuXDjPkCG3d6bz8vJo2LARv/ySzPPPN6N6dScAdDpvkpKSTPrxDq1WW6zX8aFBoNPp0Ol0RVrZwYMHSUtLY+jQoQCkpqYSHBzM7NmzcXFxMS63detWypcvz9q1a0lLSwPA3t6egQMHFrlwuehc6ZD6LCP13WYwGMy6ONvjuqhbkyYvsHTpavbt+47t27eycePtm9JoNDbk5xswGBQaze1t6/W3t397uoH8fGWsSSlFXl6e8fndma7X356u1xtQShnXo9Vq72oLer2BvDw9HTt2ZsSI9wDIyspCr9fzww8H0Ovv7jetcV33MhgMBV7HErvonLu7Ozt27CA2NpbY2FiqV6/OvHnzTEIAYMuWLWzYsIHY2Fj69OlDnz59ihUCQgjxKH3+eSRbt25Gp/Nl5MgxHD9+rEjtWrZsxcaNMQD88ccFjhz5iSZNXuBf/2rNpk2xAFy7do29e3fRokUrbGxs0ev1D1xnixYt2bNnF1evZqCU4pNPwlm9ejkvvNCcX35J5sqVVAwGg/EWl49CiQ3QJycn89lnnzF//vyS2qQQQhTJK6+8ypQpE9i8eSNarZb//GcsX3zx2UPbjRjxHhERYWze/A0ajYZx4z6gWrVqBAYO5JNPptOv36sYDAb69QuiYcNGXL9+jZs3M5k6dSI+Pj0KXWf9+g0IDBzEO+8MQSlF/foN6dt3AHZ2dowY8R4jRrxFhQr21KlT95E9f7kfQRkj9VlG6rOM3I/AMmWlPrkfgRBCiGKRIBBCCCsnQSCEKBV/0VHpMs+cfpUgEEKUOK3WBr0+v7TL+FvS6/PRam0evuBdJAiEECXO3t6BzMxrKFX6J1b/TpQykJl5FXv74v3gVq7vIIQocQ4Olbl69QqXL18Aij6UodVqMRjKbniUfn0aypevgIND5WK1kiAQQpQ4jUZDlSrVi91Ovn77eMjQkBBCWDkJAiGEsHISBEIIYeUkCIQQwspJEAghhJWTIBBCCCsnQSCEEFZOgkAIIaycBIEQQlg5CQIhhLByFgdBZGQkUVFRBaZfunSJNm3acOHChULbrV69mp49e9K1a1e5faUQQpQis4MgMzOTcePGsXDhwgLzDAYD48ePJy8vr9C2Bw8eZOHChSxevJh169bx9ddf8/vvv5tbihBCCAuYfdG5+Ph46tSpQ2BgYIF5CxYswNXVldOnTxfaNi4ujtdffx1HR0cAFi5cyJNPPmluKUIIISxg9hGBv78/wcHB2NiY3gDh6NGjJCYmFhoQd5w9e5aMjAz69u1Ljx492LlzJw4Oxbt+thBCiEfjoUcEcXFxhIeHm0xzcXEhOjq6wLLZ2dmEhoYya9YstNr7Z4xer+fQoUPMnTuX/Px8+vbtS/369WnXrl2RC69a1fzgcHJyNLttSZD6LCP1WUbqs0xZr68wDw0CnU6HTqcr0soOHjxIWloaQ4cOBSA1NZXg4GBmz56Ni4uLcblq1arRpEkTKlWqBIC7uzvJycnFCoL09JsYDMW/N2dZv1641GcZqc8yUp9lynJ9Wq3mvjvQj/Tro+7u7uzYsYPY2FhiY2OpXr068+bNMwkBAE9PT7Zv305ubi45OTns37+fpk2bPspShBBCFFGJ/Y4gOTmZQYMGAeDt7Y27uzs9e/akR48edO7cmfbt25dUKUIIIe6iUUoVf3ylDJChodIh9VlG6rOM1Ge+EhsaEkII8dcjQSCEEFZOgkAIIaycBIEQQlg5CQIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggrJ0EghBBWToJACCGsnASBEEJYOQkCIYSwchIEQghh5SQIhBDCykkQCCGElZMgEEIIKydBIIQQVs7iIIiMjCQqKqrA9EuXLtGmTRsuXLhQaLsFCxbg7e2Nt7c3ixYtsrQMIYQQZjI7CDIzMxk3bhwLFy4sMM9gMDB+/Hjy8vIKbXv27FmWL1/OunXrWLNmDYsXL+bs2bPmliKEEMICZgdBfHw8derUITAwsMC8BQsW4OrqylNPPVVoW4PBQF5eHrdu3SI3NxelFLa2tuaWIoQQwgJm//X19/cHKDAsdPToURITE5k/fz7Lli0rtG3dunXx9fXF09MTpRS9e/emdu3axdp+1aoOZtUN4OTkaHbbkiD1WUbqs4zUZ5myXl9hHhoEcXFxhIeHm0xzcXEhOjq6wLLZ2dmEhoYya9YstNr7H2zs2bOHo0ePsnfvXpRSDBo0iM2bN+Pt7V3kwtPTb2IwqCIvf4eTkyNXrmQWu11JkfosI/VZRuqzTFmuT6vV3HcH+qFBoNPp0Ol0RdrQwYMHSUtLY+jQoQCkpqYSHBzM7NmzcXFxMS63c+dOunbtSqVKlQDw9fUlKSmpWEEghBDi0XikA/Pu7u7s2LHD+Lhjx47MmzePf/zjHybLNWrUiG3btvHaa69hMBjYs2cPXbt2fZSlCCGEKKIS+x1BcnIygwYNAqB3797Uq1cPHx8fevbsSb169ejZs2dJlSKEEOIuGqVU8QfaywA5R1A6pD7LSH2WkfrM96BzBPLLYiGEsHISBEIIYeUkCIQQwspJEAghhJWTIBBCCCsnQSCEEFZOgkAIIaycBIEQQlg5CQIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggrJ0EghBBWToJACCGsnASBEEJYOQkCIYSwchbfszgyMhKtVsvw4cMBSEpK4u2336ZmzZoANG7cmPDwcJM2ubm5jB8/nqNHj1KhQgVmzJjBP//5T0tLEUIIYQazgyAzM5Pw8HA2bdrEwIEDjdOTk5MJCgpi8ODB9227ZMkS7O3tiYuLIykpibFjx/L111+bW4oQQggLmD00FB8fT506dQgMDDSZnpycTEJCAv7+/gwZMoSUlJQCbXft2kX37t0BaN26NVevXuXixYvmliKEEMICZh8R+Pv7AxAVFWUy3dHRER8fHzp37syKFSsYOXIkK1euNFkmNTUVJycn42MnJycuXbrE008/bW45RfLzjk3YnvkeZcZN70vKMa1G6rOA1GcZqc8yj7s+vYsrTTr6PPL1PjQI4uLiCozxu7i4EB0dXejyoaGhxv+/9tprfPLJJ2RmZuLo6PjA7Wi1xTs4qVrVoVjLA9hXKEceoNFqit22JEl9lpH6LCP1WeZx1mdfoRxOTg/+W2qOhwaBTqdDp9MVaWUGg4G5c+cSHByMjY3N/zZia7qZ6tWrc+XKFZydnQG4cuUK1atXL07dpKffxFDM5HVx9cKpxytcuZJZrHYlycnJUeqzgNRnGanPMiVRn7nr12o1992BfqRfH9VqtWzfvp2tW7cCEBMTQ7NmzbC3tzdZzsPDg9jYWAAOHjyInZ3dYx8WEkIIUbhH/juC6dOns3jxYnx8fFi7di3Tpk0DYMWKFURGRgIQEBBAbm4uPj4+hIWFERER8ajLEEIIUUQapVTZPfPyAOYMDYEcWlpK6rOM1GcZqc98JTY0JIQQ4q9HgkAIIaycBIEQQlg5CQIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggrJ0EghBBWToJACCGsnASBEEJYOQkCIYSwchIEQghh5SQIhBDCykkQCCGElZMgEEIIKydBIIQQVu6hN69/mMjISLRaLcOHDwcgKSmJt99+m5o1awLQuHFjwsPDTdqkpqYSEhJCWloaWq2W999/n/bt21taihBCCDOYHQSZmZmEh4ezadMmBg4caJyenJxMUFAQgwcPvm/biIgIPD096du3L6dOnSIgIIA9e/ZgY2NjbjlCCCHMZHYQxMfHU6dOHQIDA02mJycnk56eTlxcHDVr1mTSpEnUqlXLZBkvLy/atm0LgLOzM7du3SIrKwtHR0dzyxFCCGEms88R+Pv7ExwcXGAv3tHRkX79+hETE4OHhwcjR44s0NbLy4vKlSsD8N///pfnnntOQkAIIUqJRimlHrRAXFxcgTF+FxcXoqOjAYiKigIwniO4V6tWrdi5c2ehf+ijo6NZsmQJS5cuLXDUIIQQomQ8dGhIp9Oh0+mKtDKDwcDcuXMLHCnY2hbcTEREBLt372bZsmXGE8vFkZ5+E4PhgRlWKCcnR65cySx2u5Ii9VlG6rOM1GeZslyfVquhalWHwuc92g1p2b59O1u3bgUgJiaGZs2aYW9vb7JcdHQ0iYmJrFixwqwQEEII8ehY/PXRe02fPp2JEycyZ84cqlSpQkREBAArVqwgNTWVd955hzlz5uDg4EBAQICx3bx586hRo8ajLkcIIcRDPPQcQVklQ0OlQ+qzjNRnGanPfCU2NCSEEOKvR4JACCGsnASBEEJYOQkCIYSwchIEQghh5SQIhBDCykkQCCGElZMgEEIIKydBIIQQVk6CQAghrJwEgRBCWDkJAiGEsHISBEIIYeUkCIQQwspJEAghhJWTIBBCCCsnQSCEEFZOgkAIIaycxUEQGRlJVFSU8XFSUhJt27alR48e9OjRg5CQkPu2vXnzJp07dyYxMdHSMoQQQpjJ7JvXZ2ZmEh4ezqZNmxg4cKBxenJyMkFBQQwePPih65g6dSo3btwwtwQhhBCPgNlHBPHx8dSpU4fAwECT6cnJySQkJODv78+QIUNISUkptP3mzZupVKkSDRs2NLcEIYQQj4DZRwT+/v4AJsNCAI6Ojvj4+NC5c2dWrFjByJEjWblypckyFy9e5KuvvuKrr75i0KBBZm2/alUHs9oBODk5mt22JEh9lpH6LCP1Waas11eYhwZBXFwc4eHhJtNcXFyIjo4udPnQ0FDj/1977TU++eQTMjMzcXS83TkGg4Hx48czceJEKlSoYHbh6ek3MRhUsds5OTly5Uqm2dt93KQ+y0h9lpH6LFOW69NqNffdgX5oEOh0OnQ6XZE2ZDAYmDt3LsHBwdjY2PxvI7b/28ypU6c4deoU48ePB+DcuXNMmDCBqVOn0q5duyJtRwghxKNj9tBQYbRaLdu3b8fZ2Rlvb29iYmJo1qwZ9vb2xmXq1avH7t27jY8DAgJ4++23adu27aMsRQghRBE98t8RTJ8+ncWLF+Pj48PatWuZNm0aACtWrCAyMvJRb04IIYSFNEqp4g+0lwFyjqB0SH2WkfosI/WZ70HnCOSXxUIIYeUkCIQQwspJEAghhJWTIBBCCCsnQSCEEFZOgkAIIaycBIEQQlg5CQIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggrJ0EghBBWToJACCGsnASBEEJYOQkCIYSwchIEQghh5Sy+Z3FkZCRarZbhw4cDkJSUxNtvv03NmjUBaNy4MeHh4SZtcnNziYiI4ODBg+Tl5RESEoKbm5ulpQghhDCD2UGQmZlJeHg4mzZtYuDAgcbpycnJBAUFMXjw4Pu2XbBgAVevXmX9+vX8/vvvBAUFsWfPHjQajbnlCCGEMJPZQ0Px8fHUqVOHwMBAk+nJyckkJCTg7+/PkCFDSElJKdA2Li6OQYMGodFoqF+/PosWLeIveutkIYT4yzM7CPz9/QkODsbGxsZkuqOjI/369SMmJgYPDw9GjhxZoO3Zs2dJSkri5Zdf5tVXXyUtLQ2tVk5XCCFEadCoh+yKx8XFFRjjd3FxITo6GoCoqCgA4zmCe7Vq1YqdO3fi6OhonNakSRMGDBjA6NGjOXbsGAMHDiQuLs5kGSGEECXjoecIdDodOp2uSCszGAzMnTu3wJGCra3pZqpVq4aPjw8ajYZGjRpRs2ZNTp8+zQsvvFDkwtPTb2IwFH84ycnJkStXMovdrqRIfZaR+iwj9VmmLNen1WqoWtWh8HmPdkNatm/fztatWwGIiYmhWbNm2Nvbmyzn6enJ5s2bATh//jwpKSnUrVv3UZYihBCiiB75wPz06dNZvHgxPj4+rF27lmnTpgGwYsUKIiMjARg9ejSpqan4+PgwZMgQpk2bJsNCQghRSh56jqCskqGh0iH1WUbqs4zUZ74SGxoSQgjx1yNBIIQQVk6CQAghrJwEgRBCWDkJAiGEsHISBEIIYeUkCIQQwspJEAghhJWTIBBCCCsnQSCEEFZOgkAIIaycBIEQQlg5CQIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggr99Cb1z9MZGQkWq2W4cOHA5CUlMTbb79NzZo1AWjcuDHh4eEmbXJzcwkJCeH48eNotVrGjBmDq6urpaUIIYQwg9lBkJmZSXh4OJs2bWLgwIHG6cnJyQQFBTF48OD7to2NjcVgMPDNN99w7NgxBg0axJ49e8wtRQghhAXMHhqKj4+nTp06BAYGmkxPTk4mISEBf39/hgwZQkpKSoG2BoOB7Oxs9Ho92dnZVKhQwdwyhBBCWMjsIPD39yc4OBgbGxuT6Y6OjvTr14+YmBg8PDwYOXJkgbY9e/bk2rVruLu707dvX0aPHm1uGUIIISykUUqpBy0QFxdXYIzfxcWF6OhoAKKiogCM5wju1apVK3bu3Imjo6Nx2qeffkpubi5jxozhzJkzDBgwgOXLl1O7dm1LnosQQggzPPQcgU6nQ6fTFWllBoOBuXPnFjhSsLU13Ux8fDwzZ85Eo9FQt25dmjVrxpEjR4oVBOnpNzEYHphhhXJycuTKlcxityspUp9lpD7LSH2WKcv1abUaqlZ1KHSexd8aMt2Qlu3bt+Ps7Iy3tzcxMTE0a9YMe3t7k+UaNWrEt99+S4MGDcjIyODo0aOMGjWqmNvSWFCn+W1LgtRnGanPMlKfZcpqfQ+q66FDQw9z79DQiRMnmDhxIpmZmVSpUoWIiAhq1arFihUrSE1N5d133yUtLY2JEydy7tw5tFotgwcPxtfX15IyhBBCmMniIBBCCPHXJr8sFkIIKydBIIQQVk6CQAghrJwEgRBCWDkJAiGEsHISBEIIYeUkCIQQwspJEAghhJX72wbBN998g7e3N126dGHZsmUF5v/666+88sordO3alfHjx5Ofn1+i9c2ePRsfHx98fHyIiIgodL6npyc9evSgR48ehT6Hx6lfv374+PgYt3/48GGT+fv27cPPzw8vLy9mzpxZorV9/fXXxrp69OhBy5YtCQ0NNVmmtPrv5s2b+Pr6cuHCBaBo/XTx4kXeeOMNunXrxtChQ/nzzz9LrL5Vq1bh6+uLn58fISEh5ObmFmgTExODm5ubsS8f5+t9b30hISF4eXkZt719+/YCbUrys3x3fbt37zZ5H7Zr167Q+7CUZP+ZTf0NXbp0SXl6eqqrV6+qP//8U/n5+akTJ06YLOPj46N+/PFHpZRSISEhatmyZSVWX0JCgnr11VfVrVu3VG5ururXr5/atm2byTKDBw9Whw4dKrGa7mYwGFSHDh1UXl5eofOzs7OVh4eHOnfunMrLy1NBQUFq165dJVzlbcePH1ddunRR6enpJtNLo/9++ukn5evrq5o0aaLOnz9f5H4KDg5WGzduVEopNXv2bBUREVEi9Z06dUp16dJFZWZmKoPBoN5//321aNGiAu1CQ0PVN99881hqelB9Sinl6+urLl++/MB2JfVZLqy+O1JTU1WnTp3U6dOnC7Qrqf6zxN/yiGDfvn20a9eOJ598kooVK9K1a1e2bNlinP/HH3+Qk5ND8+bNAXj55ZdN5j9uTk5OjB07lvLly1OuXDn++c9/cvHiRZNljh49yvz58/Hz8yM0NJRbt26VWH2nTp1Co9EwaNAgunfvztKlS03mHzlyBGdnZ5555hlsbW3x8/Mr0f672+TJkxk5ciRVqlQxmV4a/bd69WomTZpE9erVgaL1U15eHklJSXTt2hV4vO/Fe+srX748kydPxsHBAY1GQ4MGDQq8D+H2zaZiYmLo3r07o0eP5vr16yVSX1ZWFhcvXmTixIn4+fnx2WefYTAYTNqU5Gf53vruFhERQZ8+fahTp06BeSXVf5b4WwZBamoqTk5OxsfVq1fn8uXL953v5ORkMv9xq1+/vvGNe+bMGTZv3oyHh4dx/p9//slzzz3HmDFjWL9+PTdu3ODzzz8vsfpu3LhB+/btmTNnDtHR0axcuZKEhATj/If1b0nZt28fOTk5BS6TXlr9FxYWRqtWrYyPi9JPV69excHBwXip9sf5Xry3vtq1axvvFZ6RkcGyZcvo1KlTgXZOTk4MHz6c2NhYatWqVWAY7nHVl56eTrt27fjwww9ZvXo1Bw8eZM2aNSZtSvKzfG99d5w5c4YDBw7Qr1+/QtuVVP9Z4m8ZBKqQ6+hpNJoizy8pJ06cICgoiDFjxpjsSVSqVIn58+fj7OyMra0tQUFB7N69u8TqatGiBREREVSsWJEqVarQq1cvk+2Xlf5buXJlgVulQun33x1F6aey0JeXL1+mf//+vPLKK7Rt27bA/Dlz5tCsWTM0Gg0DBw4ssfuLP/PMM8yZM4eqVatib29PQEBAgdexLPTfqlWreP311ylfvnyh80ur/4rjbxkENWrUIC0tzfg4NTXV5HDu3vlXrlwp9HDvcfrhhx8YMGAA//nPf+jZs6fJvIsXL5rs+SilCtzc53E6ePAg33///X23/7D+LQm5ubkkJSXRsWPHAvNKu//uKEo/ValShZs3b6LX64GSfy+ePHmS1157jZ49ezJs2LAC8zMzM413I4SS7ctjx46xdevWB267LHyW4+Pj8fb2LnReafZfcfwtg8DV1ZXvv/+ejIwMsrOz2bZtGy+++KJxfu3atbGzs+OHH34Abp/Vv3v+45aSksKwYcOYMWMGPj4+BeZXqFCBjz/+mPPnz6OUYtmyZXTp0qXE6svMzCQiIoJbt25x8+ZN1q9fb7L9Zs2acfr0ac6ePYter2fjxo0l2n9w+49EnTp1qFixYoF5pd1/dxSln8qVK0erVq3YvHkzULLvxZs3b/Lmm2/y7rvvEhQUVOgyFStWZMGCBcZvjS1durTE+lIpxYcffsj169fJy8tj1apVBbZd2p/ljIwMcnJyeOaZZwqdX5r9VyylcIK6RGzYsEH5+PgoLy8vNW/ePKWUUgMHDlRHjhxRSin166+/qldeeUV169ZNjRo1St26davEaps6dapq3ry56t69u/Hf8uXLTerbsmWLsf6xY8eWaH1KKTVz5kzVrVs35eXlpaKjo5VSSnXv3l1dunRJKaXUvn37lJ+fn/Ly8lJhYWHKYDCUaH2bNm1SI0aMMJlWVvrP09PT+K2S+/XTuHHj1LfffquUUurChQuqb9++SqfTqaCgIHXt2rUSqW/RokWqSZMmJu/DWbNmFagvKSlJ+fv7q27duqkhQ4aoGzdulEh9Sim1dOlSpdPpVJcuXdTHH39sXKY0P8t313f48GHVu3fvAsuUZv+ZQ25MI4QQVu5vOTQkhBCi6CQIhBDCykkQCCGElZMgEEIIKydBIIQQVk6CQAgLDBo0iN9//71YbQYPHsy6deseU0VCFF/Z+4mbEH8h8+fPL+0ShLCYBIGwSjt27OCLL74gLy+PChUqMGbMGL777jtOnDhBWloa6enpNGrUiLCwMBwcHFi+fDkrV66kXLly2NnZERoaSr169ejYsSORkZE8//zzrFq1iiVLlqDVaqlWrRoTJ06kbt26XL58mbFjx5KamsrTTz9Nenq6sY6TJ08SFhbGtWvX0Ov1BAQE0KtXL/78809CQkI4e/YsWq2WJk2aEBoailYrB/HiMSjtX7QJUdJOnz6tfH19VUZGhlLq9j0NOnTooD766CP14osvqitXrii9Xq9GjRqlPvroI5Wfn6+aNGlivC7++vXr1cqVK5VSt39leuTIEbVv3z7VuXNn430R1q5dq3Q6nTIYDOqtt95SM2fOVEopdebMGdW8eXO1du1alZeXp7y9vdXRo0eVUkrduHFD6XQ69eOPP6r169eroKAgpZRS+fn5avz48erMmTMl2U3CisgRgbA6CQkJpKamMmDAAOM0jUbDuXPn6NatG9WqVQOgV69efPjhh4wZM4Zu3brRp08fXnrpJTp06ICfn5/JOvfu3Yu3t7fxvggvv/wyYWFhXLhwgX379jFmzBgAnJ2djVf4PHPmDOfOnWPcuHHG9eTk5PDLL7/g7u7OzJkzCQgIwNXVlf79++Ps7Pw4u0VYMQkCYXUMBgPt27dn1qxZxmkpKSmsWrXK5FaNBoPBOBQzY8YMjh8/zr59+5g/fz5r1qzhiy++MC6rCrlSi1KK/Px8NBqNyfw7V5/U6/U88cQTxMbGGuelpaXh6OiInZ0d27dvJzExkf379xMYGMiECRPo1q3bI+sHIe6QAUdhddq1a0dCQgInT54EYPfu3XTv3p1bt24RHx9PZmYmBoOB1atX4+npSUZGBh4eHjz55JMMGDCAESNGcOzYMZN1urm5sXnzZjIyMgBYu3YtTz75JM7Ozri7u7Nq1Srg9iWyExMTAahbty52dnbGIEhJScHX15ejR4+yfPlyQkJCcHNz47333sPNzY0TJ06UVBcJKyMXnRNWKS4uji+//NJ4ffhx48bx/fffs3//fvR6PVevXqV169ZMmDCBChUqsHLlShYvXkyFChWwsbFh5MiRuLq6mpwsXrZsGStXrsRgMFClShU++OAD6tevT0ZGBiEhIZw7d46aNWuSn59Pz549efnll/ntt9+MJ4vz8/Pp168fr732GllZWYwbN45jx45hb2/P008/TVhYGJUrVy7trhN/QxIEQvy/qKgorl69ygcffFDapQhRomRoSAghrJwcEQghhJWTIwIhhLByEgRCCGHlJAiEEMLKSRAIIYSVkyAQQggrJ0EghBBW7v8Ab4ItqHBRUq8AAAAASUVORK5CYII=", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 获取参数\n", "cfg = get_args() \n", "# 训练\n", "env, agent = env_agent_config(cfg)\n", "res_dic = train(cfg, env, agent)\n", " \n", "plot_rewards(res_dic['rewards'], cfg, tag=\"train\") \n", "# 测试\n", "res_dic = test(cfg, env, agent)\n", "plot_rewards(res_dic['rewards'], cfg, tag=\"test\") # 画出结果" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3.7.12 ('rl_tutorials')", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.7.12" }, "orig_nbformat": 4, "vscode": { "interpreter": { "hash": "4f613f1ab80ec98dc1b91d6e720de51301598a187317378e53e49b773c1123dd" } } }, "nbformat": 4, "nbformat_minor": 2 }