Example_6.7_maximization_bias_double_Q.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#######################################################################\n",
    "# Copyright (C)                                                       #\n",
    "# 2016-2018 Shangtong Zhang(zhangshangtong.cpp@gmail.com)             #\n",
    "# 2016 Kenta Shimada(hyperkentakun@gmail.com)                         #\n",
    "# Permission given to modify the code as long as you keep this        #\n",
    "# declaration at the top                                              #\n",
    "#######################################################################\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "這邊參考的主要來源是[ShangtongZhang](https://github.com/ShangtongZhang/reinforcement-learning-an-introduction/blob/master/chapter06/windy_grid_world.py)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Example 6.7: Maximization Bias Example\n",
    "\n",
    "![](https://i.imgur.com/eIvX9qL.png)\n",
    "Figure 6.5：在一個簡單的episodic MDP上比較Q-learning與Double Q-learning(如插圖所示)。Q-learning一開始學習到採用left的頻率比right還要來的多次， 而且機率明顯高過於$\\epsilon=0.1$所強制的最小5%的機率。相比之下，Double Q-learning就不受maximization bias的影響。這些資料是執行10,000次的平均。初始的action-value為0。Any ties in $\\epsilon$-greedy action selection were broken randomly. (這句真的不會翻，不過意思大致就是，當你遇到兩個action的機率一致的時候就隨機選一個吧。)\n",
    "\n",
    "Figure 6.5那個小小的MDP插圖提供簡單的範例，說明maximization bias如何對TD control algorithms造成傷害：\n",
    "1. 有一個MPD，它有兩個non-terminal states $A$、$B$\n",
    "2. episode都是從$A$開始\n",
    "3. 兩個actions，左、右\n",
    "4. 如果選擇右，那就會馬上進入終止狀態(terminal state)，然後reward、return都是0\n",
    "5. 如果選擇左，那就轉移至$B$，然後reward也是0，然後接續會有很多actions都會導致立即終止，然後會從一個均值為-0.1、方差為1.0的均勻分佈中得到一個reward\n",
    "\n",
    "好，現在我們知道，對於任意一個trajectory是從左這個action開始的expected return都是-0.1，這樣子，你在$A$選擇左這個action就是錯的。\n",
    "\n",
    "儘管如此，我們的control methods也許會喜歡\"左\"這個action，因為maximization bias讓$B$明顯的有positive value。所以你看一下上面的圖也不難發現，用著$\\epsilon-$greedy的Q-learning一開始的學習是非常愛好\"左\"這個action。即使在漸近線處，選擇\"左\"的機率也還是比我們參數設置($\\epsilon=0.1, \\alpha=0.1, \\gamma=1$)的最佳值高了大約5%。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "載入需求套件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "import copy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "根據說明，會有兩個nonterminal states"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "STATE_A = 0\n",
    "STATE_B = 1\n",
    "STATE_TERMINAL = 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "根據說明，每一個episode都是從A開始"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "START_STATE = STATE_A"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "定義環境參數"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPSILON = 0.1\n",
    "STEP_SIZE = 0.1\n",
    "DISCOUNT_RATE = 1.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "定義STATE A的action space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "ACTION_A_LEFT = 1\n",
    "ACTION_A_RIGHT = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "定義B的action space，根據說明，進入B之後你有超級多個actions給你選擇"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "ACTION_B = range(0, 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "完整的action space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[0, 1], range(0, 10)]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ACTION_SPACE = [[ACTION_A_RIGHT, ACTION_A_LEFT], ACTION_B]\n",
    "ACTION_SPACE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "初始化Q values，根據說明，全部初始化為0，多一個是用來記錄terminal state"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0., 0.]), array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]), array([0.])]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "INIT_Q = [np.zeros(2), np.zeros(len(ACTION_B)), np.zeros(1)]\n",
    "INIT_Q"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "狀態的轉移，如果是state A的話，那就會是往右，也就是terminal state，或是往左，也就是進入B，這要搭配上面設置的action。如果是B的話就會是好多個選擇，但是都還是會進入terminal state"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[2, 1], [2, 2, 2, 2, 2, 2, 2, 2, 2, 2]]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TRANSITION = [[STATE_TERMINAL, STATE_B], [STATE_TERMINAL] * len(ACTION_B)]\n",
    "TRANSITION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "定義一個選擇action的函數"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def select_action(q_values, state):\n",
    "    if np.random.binomial(1, EPSILON) == 1:\n",
    "        return np.random.choice(ACTION_SPACE[state])\n",
    "    else:\n",
    "        _values = q_values[state]\n",
    "        return np.random.choice([_action for _action, _value in enumerate(_values) if _value == np.max(_values)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "根據說明，進入state B之後，就會從均值為-0.1、方差為1的均勻分佈中得到一個reward，也就是在這邊會產生positive bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_normal_reward(state, action):\n",
    "    if state == STATE_A:\n",
    "        return 0\n",
    "    return np.random.normal(-0.1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.3850611273429807"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_normal_reward(1, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_normal_reward(0, 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "現在讓我們來定義learning algorithm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下面給出完整的演算法做為參考：  \n",
    "![](https://i.imgur.com/lZFOlUq.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def q_learning(q1, q2=None):\n",
    "    \"\"\"Q-learning\n",
    "    如果存在q2那就使用double q learning\n",
    "    \"\"\"\n",
    "    state = START_STATE\n",
    "    # 因為我們要跟著範例確認在state A的時候left被執行幾次，所以這邊要設個參數來記錄\n",
    "    action_left_count = 0\n",
    "    while state != STATE_TERMINAL:\n",
    "        # 如果不是double q，那就直接由q1取出一個action\n",
    "        # 如果是，那就兩個value function加總起來做選擇\n",
    "        if q2 is None:\n",
    "            action = select_action(q1, state)\n",
    "        else:\n",
    "            action = select_action([_q1 + _q2 for _q1, _q2 in zip(q1, q2)], state)\n",
    "        \n",
    "        if state == STATE_A and action == ACTION_A_LEFT:\n",
    "            action_left_count += 1\n",
    "        # 如果state是A那reward就是0，不然就是從normal distribution中取得reward\n",
    "        reward = get_normal_reward(state, action)\n",
    "        next_state = TRANSITION[state][action]\n",
    "        \n",
    "        if q2 is None:\n",
    "            active_q = q1\n",
    "            target = np.max(active_q[next_state])\n",
    "        else:\n",
    "            # 如果是double q learning，那就50/50的機率\n",
    "            if np.random.binomial(1, 0.5) == 1:\n",
    "                active_q = q1\n",
    "                target_q = q2\n",
    "            else:\n",
    "                active_q = q2\n",
    "                target_q = q1\n",
    "                \n",
    "            # 根據公式，我們會從action q中取得最大value的那個action\n",
    "            best_action = np.random.choice([action_ for action_, value_ in enumerate(active_q[next_state]) if value_ == np.max(active_q[next_state])])\n",
    "            # 再用target q來評估這個action有多好\n",
    "            target = target_q[next_state][best_action]\n",
    "        \n",
    "        # 更新\n",
    "        active_q[state][action] += STEP_SIZE * (\n",
    "            reward + DISCOUNT_RATE * target - active_q[state][action]\n",
    "        )\n",
    "        state = next_state\n",
    "    return action_left_count"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "現在我們可以來試著執行了"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "首先定義執行次數"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "episodes = 300\n",
    "runs = 1000"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "設置參數來記錄在state A執行left的次數"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "left_counts_q = np.zeros((runs, episodes))\n",
    "left_counts_double_q = np.zeros((runs, episodes))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "執行"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:44<00:00, 22.41it/s]\n"
     ]
    }
   ],
   "source": [
    "for run in tqdm(range(runs)):\n",
    "    q1 = copy.deepcopy(INIT_Q)\n",
    "    q2 = copy.deepcopy(INIT_Q)\n",
    "    q3 = copy.deepcopy(INIT_Q)\n",
    "    \n",
    "    for episode in range(episodes):\n",
    "        left_counts_q[run, episode] = q_learning(q1)\n",
    "        left_counts_double_q[run, episode] = q_learning(q2, q3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "計算均值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "left_counts_q = left_counts_q.mean(axis=0)\n",
    "left_counts_double_q = left_counts_double_q.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "繪圖"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16, 9))\n",
    "plt.plot(left_counts_q, label='Q-learning')\n",
    "plt.plot(left_counts_double_q, label='Double Q-learning')\n",
    "plt.plot(np.ones(episodes) * 0.05, label='Optimal')\n",
    "plt.xlabel('episodes')\n",
    "plt.ylabel('% left actions from A')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}