【强化学习】深入浅出强化学习--机器人找金币

---

Grid_mdp.py

定义和初始化

首先自定义环境，自定义的环境将继承gym.env环境。在初始化的时候，可以指定环境支持的渲染模式（例如human,rgb_array,ansi)以及渲染环境的帧速率。当没有初始化的时候都有默认的渲染模式，在Grid World中将支持rgb_array和human模式，并以4FPS的速度渲染。

环境的__init__方法将接受整数大小，它决定了方形网格的大小。同时将设置一些用于渲染的变量，并定义self.observation_space和self.action_space。

在我们代码中，观测值应该提供有关代理和目标在二维网格上的位置的信息。将选择以字典的形式表示观察结果，并带有键“代理”和“目标”。观察结果可能看起来像 {“agent”： array（[1， 0]）， “target”： array（[0， 3]）}。由于我们的环境中有 4 个动作（“右”、“上”、“左”、“下”），将使用 Disparte（4） 作为动作空间。以下是GridWorldEnv的声明和__init__的实施：

import gym
from gym import spaces
import pygame
import numpy as np


class GridEnv(gym.Env):
    metadata = {
    
    "render_modes": ["human", "rgb_array"], "render_fps": 4}

    def __init__(self, render_mode=None, size=5):
        self.size = size  # The size of the square grid
        self.window_size = 512  # The size of the PyGame window

        # Observations are dictionaries with the agent's and the target's location.
        # Each location is encoded as an element of {0, ..., `size`}^2, i.e. MultiDiscrete([size, size]).
        self.observation_space = spaces.Dict(
            {
    
    
                "agent": spaces.Box(0, size - 1, shape=(2,), dtype=int),
                "target": spaces.Box(0, size - 1, shape=(2,), dtype=int),
            }
        )

        # We have 4 actions, corresponding to "right", "up", "left", "down", "right"
        self.action_space = spaces.Discrete(4)

        """
        The following dictionary maps abstract actions from `self.action_space` to 
        the direction we will walk in if that action is taken.
        I.e. 0 corresponds to "right", 1 to "up" etc.
        """
        self._action_to_direction = {
    
    
            0: np.array([1, 0]),
            1: np.array([0, 1]),
            2: np.array([-1, 0]),
            3: np.array([0, -1]),
        }

        assert render_mode is None or render_mode in self.metadata["render_modes"]
        self.render_mode = render_mode

        """
        If human-rendering is used, `self.window` will be a reference
        to the window that we draw to. `self.clock` will be a clock that is used
        to ensure that the environment is rendered at the correct framerate in
        human-mode. They will remain `None` until human-mode is used for the
        first time.
        """
        self.window = None
        self.clock = None

---

从环境状态构建观测值

我们需要在reset和step中计算观测值，因此通常可以方便地使用_get_obs私有方法将环境状态转化为观测值：

def _get_obs(self):
        return {
    
    "agent": self._agent_location, "target": self._target_location}

对于逐步返回并重置的辅助信息，机器人找金币例子中，提供agent和target之间的曼哈顿距离：

def _get_info(self):
        return {
    
    "distance": np.linalg.norm(self._agent_location - self._target_location, ord=1)}

通常，信息还将包含一些仅在步骤方法中可用的数据（例如个人奖励条款）。在这种情况下，我们将不得不更新 _get_info 按步骤返回的字典。

---

Reset

每次使用reset的方法来启动新的episode，每当发出完成信号是，都应该调用reset。可以传递seed进行重置，以将环境使用的任何随机数生成器初始化为确定性状态。在机器人找金币实例中，我们随机选择agent的位置和随机抽样的target位置，直到它与agent的位置不一致。

  def reset(self, seed=None, options=None):
        # We need the following line to seed self.np_random
        super().reset(seed=seed)

        # Choose the agent's location uniformly at random
        self._agent_location = self.np_random.integers(0, self.size, size=2, dtype=int)

        # We will sample the target's location randomly until it does not coincide with the agent's location
        self._target_location = self._agent_location
        while np.array_equal(self._target_location, self._agent_location):
            self._target_location = self.np_random.integers(
                0, self.size, size=2, dtype=int
            )

        observation = self._get_obs()
        info = self._get_info()

        if self.render_mode == "human":
            self._render_frame()

        return observation, info

---

Step

step方法通常包括环境的大部分逻辑。它接受一个操作，在应用该操作后计算环境的状态，并返回四元组（观察、奖励、完成、信息）。一旦计算了环境的新状态，就可以检查它是否是最终状态，并相应地设置完成。由于在GridWorld中使用稀疏二进制，因此一旦知道完成，计算奖励就变得微不足道。为收集观察和信息，再次利用_get_obs和_get_info：

 def step(self, action):
        # Map the action (element of {0,1,2,3}) to the direction we walk in
        direction = self._action_to_direction[action]
        # We use `np.clip` to make sure we don't leave the grid
        self._agent_location = np.clip(
            self._agent_location + direction, 0, self.size - 1
        )
        # An episode is done iff the agent has reached the target
        terminated = np.array_equal(self._agent_location, self._target_location)
        reward = 1 if terminated else 0  # Binary sparse rewards
        observation = self._get_obs()
        info = self._get_info()

        if self.render_mode == "human":
            self._render_frame()

        return observation, reward, terminated, False, info

---

Rendering

在这里，我们使用 PyGame 进行渲染。在 Gym 附带的许多环境中都使用了类似的渲染方法：

def render(self):
        if self.render_mode == "rgb_array":
            return self._render_frame()

    def _render_frame(self):
        if self.window is None and self.render_mode == "human":
            pygame.init()
            pygame.display.init()
            self.window = pygame.display.set_mode((self.window_size, self.window_size))
        if self.clock is None and self.render_mode == "human":
            self.clock = pygame.time.Clock()

        canvas = pygame.Surface((self.window_size, self.window_size))
        canvas.fill((255, 255, 255))
        pix_square_size = (
            self.window_size / self.size
        )  # The size of a single grid square in pixels

        # First we draw the target
        pygame.draw.rect(
            canvas,
            (255, 0, 0),
            pygame.Rect(
                pix_square_size * self._target_location,
                (pix_square_size, pix_square_size),
            ),
        )
        # Now we draw the agent
        pygame.draw.circle(
            canvas,
            (0, 0, 255),
            (self._agent_location + 0.5) * pix_square_size,
            pix_square_size / 3,
        )

        # Finally, add some gridlines
        for x in range(self.size + 1):
            pygame.draw.line(
                canvas,
                0,
                (0, pix_square_size * x),
                (self.window_size, pix_square_size * x),
                width=3,
            )
            pygame.draw.line(
                canvas,
                0,
                (pix_square_size * x, 0),
                (pix_square_size * x, self.window_size),
                width=3,
            )

        if self.render_mode == "human":
            # The following line copies our drawings from `canvas` to the visible window
            self.window.blit(canvas, canvas.get_rect())
            pygame.event.pump()
            pygame.display.update()

            # We need to ensure that human-rendering occurs at the predefined framerate.
            # The following line will automatically add a delay to keep the framerate stable.
            self.clock.tick(self.metadata["render_fps"])
        else:  # rgb_array
            return np.transpose(
                np.array(pygame.surfarray.pixels3d(canvas)), axes=(1, 0, 2)
            )

---

Close

close 方法应关闭环境使用的任何开放资源。在许多情况下，通常不需要额外使用该方法。但是，在我们的示例中，render_mode可能是“人类”，我们可能需要关闭已打开的窗口：

def close(self):
        if self.window is not None:
            pygame.display.quit()
            pygame.quit()

---

注册环境

1. 将我们⾃⼰的环境⽂件（笔者创建的⽂件名为 grid_mdp.py）拷⻉到你的gym安装⽬录/gym/gym/envs/classic_control⽂件夹中（拷⻉在此⽂件夹中是因为要使⽤rendering模块。
2. 打开该⽂件夹（第⼀步中的⽂件夹）下的__init__.py⽂件，在⽂件末尾加⼊语句：

from gym.envs.classic_control.grid_mdp import GridEnv

3. 进⼊⽂件夹的gym安装⽬录/gym/gym/envs，打开该⽂件夹下
   的__init__.py⽂件，添加代码：

register(
# gym.make(‘id’)时的id
id=“GridWorld-v0”,
# 函数路口
entry_point=“gym.envs.classic_control.grid_mdp:GridEnv”,
max_episode_steps=200,
reward_threshold=100.0,
)

4. 用pycharm打开项目，解释器为安装gym环境的解释器。同时运行以下代码：

import gym

env = gym.make('GridWorld-v0', render_mode='human')
#env = gym.make('GridWorld-v0')
env.reset()
env.render()
for _ in range(1000):
    env.render()
    observation, reward, done, info, _ = env.step(env.action_space.sample())  # take a random action
    if done:
        env.reset()
env.close()

5. 代码运行后出现如下结果：
   

---

参考文章

https://www.gymlibrary.dev/content/environment_creation/