{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nRrn4j-O0JMM"
      },
      "source": [
        "# Practical Session: Reinforcement Learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WFZsyyDk0JMP"
      },
      "source": [
        "In"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SWKBTGvH0JMQ"
      },
      "source": [
        "# Gymnasium"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KvhmAjE30JMQ"
      },
      "source": [
        "Gymnasium - это фреймворк моделирования и библиотека задач (или модельных сред environment) разработанная в [OpenAI](https://openai.com/) с единым доступом к стандартному интерфейсу модели.\n",
        "Этот стандартный интерфейс позволяет писать общие алгоритмы обучения с подкреплением (RL) и тестировать их в нескольких средах без особых адаптаций.\n",
        "Его основным объектом является ** среда **, обычно создаваемая с помощью инструкции ``python gym.make(\"ENV_NAME\")```\n",
        "В Gymnasium  есть три ключевых метода:\n",
        "* `reset()`: этот метод сбрасывает среду и возвращает наблюдение случайного начального состояния \n",
        "* `step(a)`: этот метод выполняет действие `a` и возвращает три переменные:\n",
        "    * `observation`: наблюдение за следующим состоянием\n",
        "    * `reward`: награда, полученная после перехода из предыдущего состояния в новое при выполнении действия \"а\".\n",
        "    * `done`: логическое значение, указывающее, завершен ли эпизод.\n",
        "    * `info`: переменная, используемая для передачи любого другого вида информации\n",
        "* `render()`: метод отображает/отрисовывает текущее состояние среды"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b_AWOwwu0JMR"
      },
      "source": [
        "## Frozen Lake\n",
        "\n",
        "Ледяное озеро [Frozen lake](https://gym.openai.com/envs/FrozenLake-v0/)\n",
        "простая среда \"сеточного мира\", созданная в Gymnasium.\n",
        "Начиная с неизменяющейся начальной позиции, вы управляете агентом, цель которого - достичь цели, расположенной на противоположной стороне карты.\n",
        "По некоторым плиткам можно ходить, по другим - нет (полынья), и хождение по ним ведет к окончанию эпизода. Из-за скользкости замерзшего озера в переходы может добавляться некоторая случайность, что означает, что направление движения агента неопределенно и лишь частично зависит от выбранного направления.\n",
        "\n",
        "Вот официальное описание замерзшего озера, предоставленное Open AI: \n",
        "*Winter is here. You and your friends were tossing around a frisbee at the park when you made a wild throw that left the frisbee out in the middle of the lake. The water is mostly frozen, but there are a few holes where the ice has melted. If you step into one of those holes, you'll fall into the freezing water. At this time, there's an international frisbee shortage, so it's absolutely imperative that you navigate across the lake and retrieve the disc. However, the ice is slippery, so you won't always move in the direction you intend.*\n",
        "\n",
        "*Поверхность описывается с помощью сетчатой поверхности, например:*\n",
        "```\n",
        "SFFF       (S: starting point, safe)\n",
        "FHFH       (F: frozen surface, safe)\n",
        "FFFH       (H: hole, fall to your doom)\n",
        "HFFG       (G: goal, where the frisbee is located)\n",
        "```\n",
        "* Эпизод заканчивается, когда вы достигаете цели или падаете в полынью. Вы получаете награду в размере 1, если достигнете цели, и 0 в противном случае.*\n",
        "\n",
        "Давайте создадим Модельную Среду Ледяного озера:"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install  gymnasium"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ixBQ-gjmvr-J",
        "outputId": "0ca0cbc7-5e35-4f8f-8667-2746765bb3d2"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: gymnasium in /usr/local/lib/python3.9/dist-packages (0.28.1)\n",
            "Requirement already satisfied: numpy>=1.21.0 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (1.22.4)\n",
            "Requirement already satisfied: cloudpickle>=1.2.0 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (2.2.1)\n",
            "Requirement already satisfied: importlib-metadata>=4.8.0 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (6.1.0)\n",
            "Requirement already satisfied: jax-jumpy>=1.0.0 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (1.0.0)\n",
            "Requirement already satisfied: typing-extensions>=4.3.0 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (4.5.0)\n",
            "Requirement already satisfied: farama-notifications>=0.0.1 in /usr/local/lib/python3.9/dist-packages (from gymnasium) (0.0.4)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.9/dist-packages (from importlib-metadata>=4.8.0->gymnasium) (3.15.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 99,
      "metadata": {
        "id": "tCxJZ9Kn0JMR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "06e0ca22-6376-46dd-d4a2-b5dee8009a48"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "EnvSpec(id='FrozenLake8x8-v1', entry_point='gymnasium.envs.toy_text.frozen_lake:FrozenLakeEnv', reward_threshold=0.85, nondeterministic=False, max_episode_steps=None, order_enforce=True, autoreset=False, disable_env_checker=False, apply_api_compatibility=False, kwargs={'map_name': '8x8', 'render_mode': 'ansi', 'is_slippery': False}, namespace=None, name='FrozenLake8x8', version=1, additional_wrappers=(), vector_entry_point=None)\n"
          ]
        }
      ],
      "source": [
        "import gymnasium as gym\n",
        "from gymnasium.spaces import Discrete,MultiDiscrete\n",
        "\n",
        "env = gym.make(\"FrozenLake8x8-v1\",render_mode='ansi',is_slippery=False)  # render_mode='human' render_mode=\"rgb_array\"  ansi_list)\n",
        "env.reset()\n",
        "env.render()\n",
        "print(env.env.spec )"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JVJS3IfP0JMS"
      },
      "source": [
        "Модельная Среда предоставляет информацию о своих Действиях и Состояниях."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 100,
      "metadata": {
        "id": "d6Rn2_X10JMT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4be654c5-c5c1-4251-c835-47f73793550c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Action Space Discrete(4)\n",
            "State Space Discrete(64)\n",
            "Rewards (0, 1)\n"
          ]
        }
      ],
      "source": [
        "print(f\"Action Space {env.action_space}\")\n",
        "print(f\"State Space {env.observation_space}\")\n",
        "print(f\"Rewards {env.reward_range }\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7bVcnm0v0JMT"
      },
      "source": [
        "Среда состоит из 64 дискретных состояний, соответствующих положению агента в сети. \n",
        "Четырьмя возможными действиями являются:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 101,
      "metadata": {
        "id": "oL4IVAZs0JMT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dd8dd012-8020-4782-c810-30231b9c5cfe"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "action 0: ←\n",
            "action 1: ↓\n",
            "action 2: →\n",
            "action 3: ↑\n"
          ]
        }
      ],
      "source": [
        "action_map = {0:u'\\u2190', 1:u'\\u2193', 2:u'\\u2192', 3:u'\\u2191'}\n",
        "for k, v in action_map.items():\n",
        "    print(f\"action {k}: {v}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xJZVf4VU0JMU"
      },
      "source": [
        "Используйте методы `step` и `render`, чтобы наблюдать за воздействием действий на Окружающую Среду"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 102,
      "metadata": {
        "id": "o_-rOkxU0JMU",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "19b68b63-60a7-4e1f-cfc1-a57d2b07fdb2"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'  (Down)\\nSFFFFFFF\\n\\x1b[41mF\\x1b[0mFFFFFFF\\nFFFHFFFF\\nFFFFFHFF\\nFFFHFFFF\\nFHHFFFHF\\nFHFFHFHF\\nFFFHFFFG\\n'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 102
        }
      ],
      "source": [
        "env.step(1)\n",
        "env.render()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ck2vZNpj0JMU"
      },
      "source": [
        "Эта функция поможет визуализировать траектории движения нашего агента:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 103,
      "metadata": {
        "id": "_9aqdJ3Y0JMU"
      },
      "outputs": [],
      "source": [
        "from IPython.display import clear_output\n",
        "from time import sleep\n",
        "\n",
        "def display_trajectory(frames):\n",
        "    for i, frame in enumerate(frames):\n",
        "        clear_output(wait=True)\n",
        "        print(frame['frame'])\n",
        "        print(f\"Timestep: {i + 1}\")\n",
        "        print(f\"State: {frame['state']}\")\n",
        "        print(f\"Reward: {frame['reward']}\")\n",
        "        sleep(.2)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TGWFYqGp0JMU"
      },
      "source": [
        "# Агент случайного действия"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VBSii8fD0JMU"
      },
      "source": [
        "Следующий код показывает, как запустить эпизод с агентом, выполняющим случайные действия."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 136,
      "metadata": {
        "id": "QK5gHNNR0JMV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c987ea6d-d093-4440-e208-eb6eaff1edc0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<bound method Wrapper.render of <TimeLimit<OrderEnforcing<PassiveEnvChecker<FrozenLakeEnv<FrozenLake8x8-v1>>>>>>\n",
            "Timestep: 50\n",
            "State: 29\n",
            "Reward: 0.0\n"
          ]
        }
      ],
      "source": [
        "frames = [] # for animation\n",
        "env.reset()\n",
        "while True:\n",
        "    # draw a random action from the action space\n",
        "    action = env.action_space.sample()\n",
        "    # the step method takes an action as input and returns 4 variables described in the OpenAI section\n",
        "    state, reward, done, info, loo = env.step(action)\n",
        "    frames.append({\n",
        "        'frame': env.render,\n",
        "        'state': state,\n",
        "        'reward': reward\n",
        "        })\n",
        "    #if done is True then the episode is over\n",
        "    if done:\n",
        "        break\n",
        "        \n",
        "display_trajectory(frames)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6Xy-TqHk0JMV"
      },
      "source": [
        "## Политика человека (действия игрока) "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 105,
      "metadata": {
        "id": "gB5c8UmY0JMV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "7d4c8b09-4ecd-40f2-9455-e33397631bda"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'\\n\\x1b[41mS\\x1b[0mFFFFFFF\\nFFFFFFFF\\nFFFHFFFF\\nFFFFFHFF\\nFFFHFFFF\\nFHHFFFHF\\nFHFFHFHF\\nFFFHFFFG\\n'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 105
        }
      ],
      "source": [
        "env.reset()\n",
        "env.render()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qCcN5dHi0JMV"
      },
      "source": [
        "Мы сейчас посмотрели случайную политику. Давайте теперь проверим нашу политику.\n",
        "Используя возможные действия, попытайтесь достичь цели, выполнив несколько раз следующую ячейку кода.\n",
        "Напоминание:\n",
        "action 0: ←\n",
        "action 1: ↓\n",
        "action 2: →\n",
        "action 3: ↑"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 106,
      "metadata": {
        "id": "bsiJTO480JMV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "d01daaef-0c43-4323-9e32-48a09b2ab046"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'  (Down)\\nSFFFFFFF\\n\\x1b[41mF\\x1b[0mFFFFFFF\\nFFFHFFFF\\nFFFFFHFF\\nFFFHFFFF\\nFHHFFFHF\\nFHFFHFHF\\nFFFHFFFG\\n'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 106
        }
      ],
      "source": [
        "env.step(1)\n",
        "env.render()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print('---- winning sequence ------ ')\n",
        "actions = { 'Left': 0, 'Down': 1, 'Right': 2,  'Up': 3 }\n",
        "winning_sequence = (7 * ['Right']) + (7 * ['Down'])\n",
        "print(winning_sequence)\n",
        "# env = gym.make(\"FrozenLake-v0\", is_slippery=False)\n",
        "env.reset()\n",
        "env.render()\n",
        "\n",
        "for a in winning_sequence:\n",
        "    new_state, reward, done, _, info = env.step(actions[a])\n",
        "    env.render()\n",
        "    print(f\"Reward: {reward:.2f}\")\n",
        "    print(new_state)\n",
        "    if done: break"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9qGsOYfc9-Fh",
        "outputId": "1e2ba81d-b1f4-4993-fec3-10e9ddcb16de"
      },
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "---- winning sequence ------ \n",
            "['Right', 'Right', 'Right', 'Right', 'Right', 'Right', 'Right', 'Down', 'Down', 'Down', 'Down', 'Down', 'Down', 'Down']\n",
            "Reward: 0.00\n",
            "1\n",
            "Reward: 0.00\n",
            "2\n",
            "Reward: 0.00\n",
            "2\n",
            "Reward: 0.00\n",
            "10\n",
            "Reward: 0.00\n",
            "2\n",
            "Reward: 0.00\n",
            "3\n",
            "Reward: 0.00\n",
            "3\n",
            "Reward: 0.00\n",
            "11\n",
            "Reward: 0.00\n",
            "10\n",
            "Reward: 0.00\n",
            "11\n",
            "Reward: 0.00\n",
            "19\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jw5w7yKG0JMW"
      },
      "source": [
        "В модели есть вероятность достижения желаемой цели, что связано со стохастичностью окружающей среды.\n",
        "Модельная среда предоставляет полное описание модели переходов:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 125,
      "metadata": {
        "id": "7Od24Qgu0JMW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "162eb569-8934-4c62-9503-9fecf5668478"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{0: [(0.3333333333333333, 1, 0.0, False),\n",
              "  (0.3333333333333333, 0, 0.0, False),\n",
              "  (0.3333333333333333, 9, 0.0, False)],\n",
              " 1: [(0.3333333333333333, 0, 0.0, False),\n",
              "  (0.3333333333333333, 9, 0.0, False),\n",
              "  (0.3333333333333333, 2, 0.0, False)],\n",
              " 2: [(0.3333333333333333, 9, 0.0, False),\n",
              "  (0.3333333333333333, 2, 0.0, False),\n",
              "  (0.3333333333333333, 1, 0.0, False)],\n",
              " 3: [(0.3333333333333333, 2, 0.0, False),\n",
              "  (0.3333333333333333, 1, 0.0, False),\n",
              "  (0.3333333333333333, 0, 0.0, False)]}"
            ]
          },
          "metadata": {},
          "execution_count": 125
        }
      ],
      "source": [
        "state = 1\n",
        "env.env.P[state]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8oNz3tP-0JMX"
      },
      "source": [
        "Здесь видны вероятности переходов для каждого возможного действия.\n",
        "Например, выполнение действия 2 имеет:\n",
        "* 33% шансов привести к состоянию 9\n",
        "* 33% шансов привести к состоянию 2\n",
        "* 33% шансов привести к состоянию 1\n",
        "\n",
        "Каждая строка составлена следующим образом: (вероятность, следующее состояние, награда, конец эпизода)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "state = 62\n",
        "env.env.P[state]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i81eMF1-EahC",
        "outputId": "b1f68517-0448-49be-fd60-e4f783e41688"
      },
      "execution_count": 129,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{0: [(0.3333333333333333, 54, 0.0, True),\n",
              "  (0.3333333333333333, 61, 0.0, False),\n",
              "  (0.3333333333333333, 62, 0.0, False)],\n",
              " 1: [(0.3333333333333333, 61, 0.0, False),\n",
              "  (0.3333333333333333, 62, 0.0, False),\n",
              "  (0.3333333333333333, 63, 1.0, True)],\n",
              " 2: [(0.3333333333333333, 62, 0.0, False),\n",
              "  (0.3333333333333333, 63, 1.0, True),\n",
              "  (0.3333333333333333, 54, 0.0, True)],\n",
              " 3: [(0.3333333333333333, 63, 1.0, True),\n",
              "  (0.3333333333333333, 54, 0.0, True),\n",
              "  (0.3333333333333333, 61, 0.0, False)]}"
            ]
          },
          "metadata": {},
          "execution_count": 129
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qc1cA1rn0JMX"
      },
      "source": [
        "#Расчет политики (правил действий)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YE6pGM_c0JMX"
      },
      "source": [
        "Let's now try to solve the FrozenLake problem using the value policy algorithm.\n",
        "![](https://github.com/DavidBert/N7-techno-IA/blob/master/code/reinforcement_learning/images/policy_iter.png?raw=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kJEgMDCN0JMX"
      },
      "source": [
        "Запишем реализацию расчета алгоритма политики (правила):\n",
        "(параметр:  ```env.env.P[s][a]``` возвратит список  \n",
        "$$[(p_1, s'_1, r_1, done)\\\\  \n",
        "...\\\\\n",
        "(p_n, s'_n, r_n, done)]$$\n",
        "Выполним перебор по этому списку с расчетом $\\sum_{s',r}p(s',r|s,a)[r+\\gamma V(s')]$)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 130,
      "metadata": {
        "id": "t-7tCF8v0JMX"
      },
      "outputs": [],
      "source": [
        "def compute_sum(env, V, s, a, gamma):\n",
        "    # V is a list containing the estimated value for every state\n",
        "    # len(V) = nb_states\n",
        "    total = 0  # state value for state s\n",
        "    for p, s_prime, r, _ in env.env.P[s][a]:\n",
        "        total += p * (r + gamma * V[s_prime])\n",
        "    return total"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 131,
      "metadata": {
        "id": "pXsNtJGO0JMY"
      },
      "outputs": [],
      "source": [
        "import numpy as np \n",
        "\n",
        "def policy_iteration(env, gamma, theta):\n",
        "    nb_states = env.observation_space.n\n",
        "    nb_actions = env.action_space.n\n",
        "    # 1. Initialization\n",
        "    V = np.zeros(nb_states)\n",
        "    pi = np.zeros(nb_states)\n",
        "    \n",
        "    while True:\n",
        "    \n",
        "        # 2. Policy Evaluation\n",
        "        while True:\n",
        "            delta = 0\n",
        "            for s in range(nb_states):\n",
        "                v = V[s]\n",
        "                V[s] = compute_sum(env, V=V, s=s, a=pi[s], gamma=gamma)\n",
        "                delta = max(delta, abs(v - V[s]))\n",
        "            if delta < theta: break\n",
        "\n",
        "        # 3. Policy Improvement\n",
        "        policy_stable = True\n",
        "        for s in range(nb_states):\n",
        "            old_action = pi[s]\n",
        "            pi[s] = np.argmax([compute_sum(env, V=V, s=s, a=a, gamma=gamma) for a in range(nb_actions)])\n",
        "            if old_action != pi[s]: policy_stable = False\n",
        "        if policy_stable: break\n",
        "    return V, pi"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ae9B6bYE0JMY"
      },
      "source": [
        "Запустим расчет ледяного озера с параметрами ( $\\gamma=1$ и $\\theta=1e-6$)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 132,
      "metadata": {
        "id": "zP1rZGJC0JMY"
      },
      "outputs": [],
      "source": [
        "V, pi = policy_iteration(env, gamma=1.0, theta=1e-6)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uYIDHOVo0JMY"
      },
      "source": [
        "Выведем на экран полученные значения политики (правила)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 133,
      "metadata": {
        "id": "VN7HmF3W0JMZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        },
        "outputId": "d8a8477f-2ff3-4ec5-ca41-bfa0c7e7760b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[['↓' '→' '↓' '→' '→' '→' '→' '→']\n",
            " ['↑' '↑' '↑' '↑' '↑' '↑' '↑' '→']\n",
            " ['←' '←' '←' '←' '→' '↑' '↑' '→']\n",
            " ['←' '←' '←' '↓' '←' '←' '→' '→']\n",
            " ['←' '↑' '←' '←' '→' '↓' '↑' '→']\n",
            " ['←' '←' '←' '↓' '↑' '←' '←' '→']\n",
            " ['←' '←' '↓' '←' '←' '←' '←' '→']\n",
            " ['←' '↓' '←' '←' '↓' '→' '↓' '←']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import seaborn as sns\n",
        "sns.color_palette(\"YlOrBr\", as_cmap=True)\n",
        "sns.heatmap(V.reshape([8, -1]), cmap=\"coolwarm\", annot=True)\n",
        "policy = np.array([action_map[x] for x in pi]).reshape([-1, 8])\n",
        "print((policy))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A3_DC3Sr0JMZ"
      },
      "source": [
        "Можно заметить, что правило для состояний 55(плитка (6,7)) и 62(плитка (6,6)) отличается от той, которую мы ожидали.\n",
        "Используя ``env.env.P[s]`` для проверки переходов среды, можете ли вы объяснить это поведение и почему состояние 62 имеет низкое значение?"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 134,
      "metadata": {
        "id": "0_AsY3-M0JMZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "26c73a27-2eba-4c2d-fafc-18c0c71c0ddf"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[[180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        ...,\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230]],\n",
              "\n",
              "       [[180, 200, 230],\n",
              "        [204, 230, 255],\n",
              "        [204, 230, 255],\n",
              "        ...,\n",
              "        [204, 230, 255],\n",
              "        [204, 230, 255],\n",
              "        [180, 200, 230]],\n",
              "\n",
              "       [[180, 200, 230],\n",
              "        [235, 245, 249],\n",
              "        [204, 230, 255],\n",
              "        ...,\n",
              "        [204, 230, 255],\n",
              "        [204, 230, 255],\n",
              "        [180, 200, 230]],\n",
              "\n",
              "       ...,\n",
              "\n",
              "       [[180, 200, 230],\n",
              "        [235, 245, 249],\n",
              "        [235, 245, 249],\n",
              "        ...,\n",
              "        [204, 230, 255],\n",
              "        [235, 245, 249],\n",
              "        [180, 200, 230]],\n",
              "\n",
              "       [[180, 200, 230],\n",
              "        [235, 245, 249],\n",
              "        [235, 245, 249],\n",
              "        ...,\n",
              "        [204, 230, 255],\n",
              "        [204, 230, 255],\n",
              "        [180, 200, 230]],\n",
              "\n",
              "       [[180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        ...,\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230],\n",
              "        [180, 200, 230]]], dtype=uint8)"
            ]
          },
          "metadata": {},
          "execution_count": 134
        }
      ],
      "source": [
        "env.env.s = 62\n",
        "env.render()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 118,
      "metadata": {
        "id": "DB9HlvpG0JMZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d2a52c28-0e7c-4c8f-fc22-4ac7a9788f4b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{0: [(1.0, 61, 0.0, False)],\n",
              " 1: [(1.0, 62, 0.0, False)],\n",
              " 2: [(1.0, 63, 1.0, True)],\n",
              " 3: [(1.0, 54, 0.0, True)]}"
            ]
          },
          "metadata": {},
          "execution_count": 118
        }
      ],
      "source": [
        "env.env.P[62]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S_PxqHvk0JMZ"
      },
      "source": [
        "# Расчет ценности:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GWVPdW150JMZ"
      },
      "source": [
        "Решим задачу ледяного озера по алгоритму расчета ценности.\n",
        "![](https://github.com/DavidBert/N7-techno-IA/blob/master/code/reinforcement_learning/images/value_iteration.png?raw=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lxQZhVJB0JMa"
      },
      "source": [
        "Код в ячейке реализует алгоритм расчета ценности:  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 119,
      "metadata": {
        "id": "aFXjvo670JMa"
      },
      "outputs": [],
      "source": [
        "def value_iteration(env, gamma, theta):\n",
        "    nb_states = env.observation_space.n\n",
        "    nb_actions = env.action_space.n\n",
        "    V = np.zeros(nb_states)\n",
        "    \n",
        "    while True:\n",
        "        delta = 0\n",
        "        for s in range(nb_states):\n",
        "            v = V[s]\n",
        "            V[s] = max([compute_sum(env, V=V, s=s, a=a, gamma=gamma) for a in range(nb_actions)])            \n",
        "            delta = max(delta, abs(v - V[s]))\n",
        "    \n",
        "        if delta < theta: break\n",
        "\n",
        "    # Output a deterministic policy\n",
        "    pi = np.zeros(nb_states)\n",
        "    for s in range(nb_states):\n",
        "        pi[s] = np.argmax([compute_sum(env, V=V, s=s, a=a, gamma=gamma) for a in range(nb_actions)])\n",
        "    \n",
        "    return V, pi"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wy-pDmZ30JMa"
      },
      "source": [
        "запустим расчет (со значениями $\\gamma=1$ и $\\theta=1e-6$)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 120,
      "metadata": {
        "id": "oEn5DenG0JMa"
      },
      "outputs": [],
      "source": [
        "V, pi = value_iteration(env, gamma=1.0, theta=1e-6)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qx9l4EZ60JMa"
      },
      "source": [
        "Выведем на консоль полученные значения и действия"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 121,
      "metadata": {
        "id": "lXgW0Vov0JMa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        },
        "outputId": "433913c1-eccd-481d-e6ea-5835676dbd87"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[['←' '←' '←' '←' '←' '←' '←' '←']\n",
            " ['←' '←' '←' '←' '←' '←' '←' '←']\n",
            " ['←' '←' '←' '←' '↓' '←' '←' '←']\n",
            " ['←' '←' '←' '←' '←' '←' '↓' '←']\n",
            " ['←' '←' '←' '←' '↓' '←' '←' '←']\n",
            " ['←' '←' '←' '↓' '←' '←' '←' '↓']\n",
            " ['←' '←' '↓' '←' '←' '↓' '←' '↓']\n",
            " ['←' '←' '←' '←' '↓' '←' '←' '←']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "sns.color_palette(\"YlOrBr\", as_cmap=True)\n",
        "sns.heatmap(V.reshape([8, -1]), cmap=\"coolwarm\", annot=True)\n",
        "policy = np.array([action_map[x] for x in pi]).reshape([-1, 8])\n",
        "print((policy))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pXcs1Yg40JMa"
      },
      "source": [
        "На этом практическом занятии мы обучили агента решать проблему ледяного озера, даже не взаимодействуя с ним.\n",
        "Алгоритмы итерации политики и итерации значений являются алгоритмами, основанными на модели. Они могут быть вычислены только в том случае, если мы знаем вероятности переходов, используемые в окружающей среде. \n",
        "Но это не всегда так.\n",
        "На следующем примере обучим агента RL взаимодействию со средой с использованием алгоритма без модели (model-free algorithm)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 122,
      "metadata": {
        "id": "oWCqAJqi0JMb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "outputId": "ba5cf1c2-6404-4eb2-d295-2dd24a86904a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "State: 0\n",
            "Action: 1\n",
            "Reward: 0.0\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "env = gym.make(\"FrozenLake8x8-v1\",render_mode='rgb_array')\n",
        "env.reset()\n",
        "# Sample random action\n",
        "action = env.action_space.sample()\n",
        "next_state, reward, done, foo, loo = env.step(action)\n",
        "# Print output\n",
        "print(f\"State: {next_state}\")\n",
        "print(f\"Action: {action}\")\n",
        "print(f\"Reward: {reward}\")\n",
        "# Render and plot an environment frame\n",
        "frame = env.render()\n",
        "plt.imshow(frame)\n",
        "plt.axis(\"off\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Модель агента случайного выбора"
      ],
      "metadata": {
        "id": "BY9B5_0ugtfm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "epoch = 0\n",
        "num_failed = 0\n",
        "experience_buffer = []\n",
        "acum_reward = 0\n",
        "done = False\n",
        "scene = 1\n",
        "step = 0\n",
        "env.reset()\n",
        "\n",
        "while not done:\n",
        "    # Sample random action\n",
        "    action = env.action_space.sample()\n",
        "    state, reward, danet, foo, loo = env.step(action)\n",
        "    acum_reward += reward\n",
        "    # Store experience in dictionary\n",
        "    experience_buffer.append({\n",
        "        \"frame\": env.render(),\n",
        "        \"episode\": scene,\n",
        "        \"step\": step,\n",
        "        \"state\": state,\n",
        "        \"action\": action,\n",
        "        \"reward\": acum_reward,\n",
        "        }\n",
        "    )\n",
        "    if danet and reward < 1:\n",
        "        num_failed += 1\n",
        "        scene += 1\n",
        "        env.reset()\n",
        "    step += 1\n",
        "    if reward ==1 or step >5000: done =True\n",
        "\n",
        "# Run animation and print console output\n",
        "# run_animation(experience_buffer)\n",
        "print(f\"# steps: {step}\")\n",
        "print(f\"# failed: {num_failed}\")\n",
        "print(f\"# reward: {acum_reward}\")\n",
        "print(f\"# scene: {scene}\")\n",
        "# Render and plot an environment frame\n",
        "frame = env.render()\n",
        "plt.imshow(frame)\n",
        "plt.axis(\"off\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 335
        },
        "id": "JCkXT_F1gxC3",
        "outputId": "73c4b29c-955f-4274-fa41-16752d2dd5a3"
      },
      "execution_count": 172,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0\n",
            "# steps: 2622\n",
            "# failed: 89\n",
            "# reward: 1.0\n",
            "# scene: 90\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "s=0\n",
        "pim = np.zeros(111) #nb_states\n",
        "if reward ==1:\n",
        "  for sce in experience_buffer:\n",
        "    if sce[\"episode\"]==scene:\n",
        "      pim[s]=sce[\"action\"]\n",
        "      s +=1\n",
        "  print (pim[:s])\n",
        "print (s)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bwaBlDFhliOe",
        "outputId": "a399e365-4bd0-4774-912e-9f3aa356071b"
      },
      "execution_count": 173,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0. 1. 3. 1. 0. 1. 3. 0. 3. 1. 2. 2. 2. 3. 1. 1. 0. 2. 0. 0. 1. 2. 2. 0.\n",
            " 3. 2. 2. 2. 1. 1. 1. 0. 1. 2. 1. 2.]\n",
            "36\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Вариант политики\n",
        "\n",
        "[0. 1. 3. 1. 0. 1. 3. 0. 3. 1. 2. 2. 2. 3. 1. 1. 0. 2. 0. 0. 1. 2. 2. 0. 3. 2. 2. 2. 1. 1. 1. 0. 1. 2. 1. 2.]"
      ],
      "metadata": {
        "id": "yqvQflosnH4b"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from scipy.special import softmax  # импортируем функцию softmax\n",
        "import math\n",
        "\"\"\"Training the agent\"\"\"\n",
        "nb_states=env.observation_space.n\n",
        "nb_actions=env.action_space.n\n",
        "q_table = np.zeros([nb_states, nb_actions])  # начальные значения 0\n",
        "\n",
        "# Hyperparameters\n",
        "lm = 0.1  # Learning rate - гиперпараметр (скорость обучения)\n",
        "gamma = 1.0  # Discount rate\n",
        "episodes = 7000  # Number of episodes\n",
        "eps1, eps2, decays = 1, 0.001, 5000  # параметры epsilon\n",
        "epsilon = eps1 # Exploration rate\n",
        "\n",
        "q_table  = np.random.uniform( size=(nb_states, nb_actions) )\n",
        "aQ = q_table.mean(axis=1, keepdims=True) # средние значения (num_states, 1)\n",
        "pimc = softmax(q_table - aQ, axis=1)     # условные вероятности pi(a|s)\n",
        "  \n",
        "s  = 1                             # номер состояния\n",
        "a  = np.random.choice(np.arange(nb_actions), p=pimc[s]) \n",
        "decay = math.exp(math.log(eps2/eps1)/decays)\n",
        "\n",
        "def policy(s):                             # epsilon-жадная политика\n",
        "    if np.random.random() < epsilon:       # случайно любое действие\n",
        "        return np.random.randint(nb_actions)\n",
        "    return np.argmax(q_table[s])                 # иначе лучшее\n",
        "\n",
        "def run_episode(ticks=5000):    \n",
        "    s0, _ = env.reset()\n",
        "    a0 = policy(s0)\n",
        "    for t in range(ticks):                \n",
        "        s1, r1, done, foo, _ = env.step(a0)\n",
        "        a1 = policy(s1)\n",
        "        # \n",
        "        q_table[s0, a0] += lm * (r1 + gamma * q_table[s1, a1] - q_table[s0, a0])\n",
        "        # \n",
        "        if done:  return\n",
        "        # \n",
        "        s0, a0 = s1, a1\n",
        "\n",
        "for episode in range(episodes):        # проходы по эпизодам\n",
        "    run_episode(1111)\n",
        "    epsilon *= decay                   # становимся жаднее\n",
        "    if epsilon < eps2: epsilon = 0\n",
        "\n",
        "pim = np.zeros(nb_states)\n",
        "for s in range(nb_states):\n",
        "    pim[s] = np.argmax(q_table[s,0:4])"
      ],
      "metadata": {
        "id": "zNBcWtF43laa"
      },
      "execution_count": 161,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print (pim)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Mfx1JxdiN7Ca",
        "outputId": "02320901-e4aa-424e-91e6-8bd47b768089"
      },
      "execution_count": 162,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0. 0. 0. 3. 2. 2. 2. 0. 0. 0. 0. 3. 3. 3. 2. 2. 2. 0. 0. 1. 1. 1. 0. 3.\n",
            " 1. 2. 0. 1. 1. 3. 2. 2. 3. 3. 0. 0. 0. 3. 3. 3. 2. 2. 0. 1. 0. 3. 1. 0.\n",
            " 3. 3. 2. 3. 0. 3. 2. 0. 3. 0. 0. 2. 3. 1. 0. 3.]\n"
          ]
        }
      ]
    }
  ],
  "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": "3.6.12"
    },
    "colab": {
      "provenance": []
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}