AI6101_Code_TengGuangWay_G2102434F.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Student Detail: \n",
    "\n",
    "Teng Guang Way G2101434F AI6101 Assignment March 2023\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "executionInfo": {
     "elapsed": 708,
     "status": "ok",
     "timestamp": 1676469964459,
     "user": {
      "displayName": "Longtao Zheng",
      "userId": "17721972102166804099"
     },
     "user_tz": -480
    },
    "id": "ehHMJ829mzJc"
   },
   "outputs": [],
   "source": [
    "from typing import List, Tuple\n",
    "import copy\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "UHqph9kA6wH7"
   },
   "source": [
    "The following class defines the grid world environment. The grid world looks like:\n",
    "\n",
    "```\n",
    "____________________________________________________________\n",
    "|___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "|_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "|_3_|___|___|___|_x_|___|___|_x_|___|___|___|___|_x_|_x_|___|\n",
    "|_4_|___|_B_|___|_x_|___|___|___|___|___|___|___|_x_|_x_|_G_|\n",
    "|_5_|_A_|___|___|_x_|___|___|___|___|___|___|___|_x_|_x_|___|\n",
    "```\n",
    "\n",
    "## NOTICE!!!:  \n",
    "\n",
    "The following environment class, CliffBoxGridWorld has been modified slightly to implement improved reward engineering. The reward engineering will encourage the agent to attempt more intelligent iterations by formulating the reward through a set of wider criteria. The reward engineering method is however, only activated in the final implementation's custom learning algorithm to further enhance its learning capability. (Faster learning rate with shortest trajectory with maximum possible cumulative reward. )\n",
    "\n",
    "Modified part of the code are marked with the comment \"NOTICE!!!\"\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "executionInfo": {
     "elapsed": 516,
     "status": "ok",
     "timestamp": 1676469970811,
     "user": {
      "displayName": "Longtao Zheng",
      "userId": "17721972102166804099"
     },
     "user_tz": -480
    },
    "id": "-RMDF0A3oKfX"
   },
   "outputs": [],
   "source": [
    "AGENT = 'A'\n",
    "BOX = 'B'\n",
    "GOAL = 'G'\n",
    "DANGER = 'x'\n",
    "GRID = '_'\n",
    "\n",
    "\n",
    "class CliffBoxGridWorld:\n",
    "    \"\"\"\n",
    "    Cliff Box Pushing Grid World.\n",
    "    \"\"\"\n",
    "    action_space = [1, 2, 3, 4]\n",
    "    forces = {\n",
    "        1: np.array([-1, 0]),\n",
    "        2: np.array([1,  0]),\n",
    "        3: np.array([0, -1]),\n",
    "        4: np.array([0,  1]),\n",
    "    }\n",
    "    world_width = 14\n",
    "    world_height = 6\n",
    "    goal_pos = np.array([4, 13])\n",
    "    init_agent_pos = np.array([5, 0])\n",
    "    init_box_pos = np.array([4, 1])\n",
    "    danger_region = [\n",
    "        [(2, 3), (5, 3)],\n",
    "        [(0, 6), (3, 6)],\n",
    "        [(0, 7), (2, 7)],\n",
    "        [(3, 11), (5, 11)],\n",
    "        [(2, 12), (5, 12)],\n",
    "    ]\n",
    "\n",
    "    def __init__(self,\n",
    "                 episode_length=100,\n",
    "                 render=False,\n",
    "                 rewardengineering=False,\n",
    "                 ):\n",
    "        \"\"\"\n",
    "        The grid world looks like:\n",
    "        ____________________________________________________________\n",
    "        |___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "        |_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "        |_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "        |_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "        |_3_|___|___|___|_x_|___|___|_x_|___|___|___|___|_x_|_x_|___|\n",
    "        |_4_|___|_B_|___|_x_|___|___|___|___|___|___|___|_x_|_x_|_G_|\n",
    "        |_5_|_A_|___|___|_x_|___|___|___|___|___|___|___|_x_|_x_|___|\n",
    "        \"\"\"\n",
    "        # Environment configurations.\n",
    "        self.episode_length = episode_length\n",
    "        self.render = render\n",
    "        self.agent_pos = self.init_agent_pos\n",
    "        self.box_pos = self.init_box_pos\n",
    "        \n",
    "        \n",
    "        # NOTICE!!! --- Custom Code BEGIN HERE ----\n",
    "        # New custom parameters for custom MDP Formulation of Reward, set to False to train RL algorithm with original formulation\n",
    "        self.rewardengineering = rewardengineering\n",
    "        self.turns_since_box_pushed = 0\n",
    "        # Here are an obvious list of major bad decisions that slow down training\n",
    "        self.make_stupid_decision_reward_walking_to_wall = -1\n",
    "        self.walking_away_from_box_stupidly = -1000\n",
    "        self.pushing_box_to_map_edge = -1000\n",
    "        # self.cliff_fall_reward = -1000\n",
    "        # --- Custom Code ENDS HERE ----\n",
    "\n",
    "        # Visualization.\n",
    "        if self.render:\n",
    "            self.world = np.chararray((self.world_height, self.world_width))\n",
    "            self.last_agent_pos = copy.deepcopy(self.agent_pos)\n",
    "            self.last_box_pos = copy.deepcopy(self.box_pos)\n",
    "            self.world[:] = GRID\n",
    "            for region in self.danger_region:\n",
    "                A, B = region\n",
    "                assert A[1] == B[1], \"A[1] != B[1]\"\n",
    "                self.world[A[0]:B[0]+1, A[1]] = DANGER\n",
    "            self.world[self.agent_pos[0], self.agent_pos[1]] = AGENT\n",
    "            self.world[self.box_pos[0], self.box_pos[1]] = BOX\n",
    "            self.world[self.goal_pos[0], self.goal_pos[1]] = GOAL\n",
    "\n",
    "    def reset(self):\n",
    "        \"\"\"\n",
    "        Resets the environment.\n",
    "\n",
    "        Returns:\n",
    "            The initial state (agent position and box position).\n",
    "        \"\"\"\n",
    "        self.timesteps = 0\n",
    "        self.action_history = []\n",
    "        self.agent_pos = self.init_agent_pos\n",
    "        self.box_pos = self.init_box_pos\n",
    "\n",
    "        return tuple([*self.agent_pos.tolist(), *self.box_pos.tolist()])\n",
    "\n",
    "    def step(self, actions: int):\n",
    "        \"\"\" \n",
    "        Args: actions (a list of int).\n",
    "\n",
    "        Returns:\n",
    "            The next state, reward, done, info.\n",
    "        \"\"\"\n",
    "        self.action_history.append(actions)\n",
    "\n",
    "        # Update the state.\n",
    "        force = self.forces[actions]\n",
    "        # check if the agent is near the box\n",
    "        if np.sum(np.abs(self.agent_pos - self.box_pos)) == 1:\n",
    "            # check if box is moved\n",
    "            if all(self.agent_pos + force == self.box_pos):\n",
    "                # check out of boundary\n",
    "                self.box_pos = self._check_pos_boundary(pos=self.box_pos + force, box_hard_boundary=True)\n",
    "        # move the agent\n",
    "        new_agent_pos = self._check_pos_boundary(self.agent_pos + force)\n",
    "        if not all(new_agent_pos == self.box_pos):\n",
    "            self.agent_pos = new_agent_pos\n",
    "        state = tuple([*self.agent_pos.tolist(), *self.box_pos.tolist()])\n",
    "\n",
    "        # Calculate the rewards\n",
    "        done = self.timesteps == self.episode_length - 1\n",
    "        # the distance between agents and box\n",
    "        dist = np.sum(np.abs(self.agent_pos - self.box_pos))\n",
    "        reward = -1  # -1 for each step\n",
    "        reward -= dist\n",
    "        # if agents or box is off the cliff\n",
    "        if self._check_off_cliff(self.agent_pos) or self._check_off_cliff(self.box_pos):\n",
    "            reward += -1000\n",
    "            done = True\n",
    "        \n",
    "        if all(self.box_pos == self.goal_pos):\n",
    "            reward += 1000\n",
    "            done = True\n",
    "        \n",
    "        reward -= np.sum(np.abs(self.box_pos - self.goal_pos))\n",
    "\n",
    "        if self.render:\n",
    "            self._update_render()\n",
    "\n",
    "        self.timesteps += 1\n",
    "        info = {}\n",
    "\n",
    "        return state, reward, done, info\n",
    "    \n",
    "\n",
    "    \n",
    "    \n",
    "    def print_world(self):\n",
    "        \"\"\"\n",
    "        Render the world in the command line.\n",
    "        \"\"\"\n",
    "        if len(self.action_history) > 0:\n",
    "            print(f'Action: {self.action_history[-1]}')\n",
    "        print(self.world)\n",
    "\n",
    "    def _check_pos_boundary(self, pos, box_hard_boundary: bool = False):\n",
    "        \"\"\"\n",
    "        Move the given position within the world bound.\n",
    "        \"\"\"\n",
    "        if pos[0] < 0:\n",
    "            pos[0] = 0\n",
    "        if pos[0] >= self.world_height:\n",
    "            pos[0] = self.world_height - 1\n",
    "        if pos[1] < 0:\n",
    "            pos[1] = 0\n",
    "        if pos[1] >= self.world_width:\n",
    "            pos[1] = self.world_width - 1\n",
    "        \n",
    "        if box_hard_boundary:\n",
    "            if pos[0] == 0:\n",
    "                pos[0] += 1\n",
    "            elif pos[0] == self.world_height - 1:\n",
    "                pos[0] = self.world_height - 2\n",
    "            if pos[1] == 0:\n",
    "                pos[1] += 1\n",
    "                    \n",
    "        return pos\n",
    "\n",
    "    def _check_off_cliff(self, pos):\n",
    "        \"\"\"\n",
    "        Check if the given position is off cliff.\n",
    "        \"\"\"\n",
    "        for region in self.danger_region:\n",
    "            A, B = region\n",
    "            assert A[1] == B[1], \"A[1] != B[1]\"\n",
    "            if A[0] <= pos[0] <= B[0] and pos[1] == A[1]:\n",
    "                return True\n",
    "        return False\n",
    "\n",
    "    def _update_render(self):\n",
    "        \"\"\"\n",
    "        Update the render information.\n",
    "        \"\"\"\n",
    "        if not all(self.last_agent_pos == self.agent_pos):\n",
    "                pos = self.last_agent_pos\n",
    "                if (pos[0] != self.goal_pos[0]) or (pos[1] != self.goal_pos[1]):\n",
    "                    self.world[pos[0], pos[1]] = GRID\n",
    "\n",
    "        if not all(self.last_box_pos == self.box_pos):\n",
    "            pos = self.last_box_pos\n",
    "            if self.world[pos[0], pos[1]].decode('UTF-8') not in {AGENT}:\n",
    "                self.world[pos[0], pos[1]] = GRID\n",
    "\n",
    "        if (self.agent_pos[0] != self.goal_pos[0]) or (self.agent_pos[1] != self.goal_pos[1]):\n",
    "            self.world[self.agent_pos[0], self.agent_pos[1]] = AGENT\n",
    "        self.world[self.box_pos[0], self.box_pos[1]] = BOX\n",
    "        self.last_box_pos = copy.deepcopy(self.box_pos)\n",
    "        self.last_agent_pos = copy.deepcopy(self.agent_pos)\n",
    "        \n",
    "    # NOTICE!!! --- Custom Code ON THE FOLLOW LINE----\n",
    "    #Modified step method to take original_step as input to calculate engineered reward if rewardengineering == True\n",
    "    def step_w_re(self, actions: int, original_state=None):\n",
    "        \"\"\" \n",
    "        Args: actions (a list of int).\n",
    "\n",
    "        Returns:\n",
    "            The next state, reward, done, info.\n",
    "        \"\"\"\n",
    "        self.action_history.append(actions)\n",
    "\n",
    "        # Update the state.\n",
    "        force = self.forces[actions]\n",
    "        # check if the agent is near the box\n",
    "        if np.sum(np.abs(self.agent_pos - self.box_pos)) == 1:\n",
    "            # check if box is moved\n",
    "            if all(self.agent_pos + force == self.box_pos):\n",
    "                # check out of boundary\n",
    "                self.box_pos = self._check_pos_boundary(pos=self.box_pos + force, box_hard_boundary=True)\n",
    "        # move the agent\n",
    "        new_agent_pos = self._check_pos_boundary(self.agent_pos + force)\n",
    "        if not all(new_agent_pos == self.box_pos):\n",
    "            self.agent_pos = new_agent_pos\n",
    "        state = tuple([*self.agent_pos.tolist(), *self.box_pos.tolist()])\n",
    "\n",
    "        # Calculate the rewards\n",
    "        done = self.timesteps == self.episode_length - 1\n",
    "        # the distance between agents and box\n",
    "        dist = np.sum(np.abs(self.agent_pos - self.box_pos))\n",
    "        reward = -1  # -1 for each step\n",
    "        reward -= dist\n",
    "        # if agents or box is off the cliff\n",
    "        if self._check_off_cliff(self.agent_pos) or self._check_off_cliff(self.box_pos):\n",
    "            reward += -1000\n",
    "            done = True\n",
    "        \n",
    "        if all(self.box_pos == self.goal_pos):\n",
    "            reward += 1000\n",
    "            done = True\n",
    "        \n",
    "        reward -= np.sum(np.abs(self.box_pos - self.goal_pos))\n",
    "\n",
    "        if self.render:\n",
    "            self._update_render()\n",
    "\n",
    "        self.timesteps += 1\n",
    "        info = {}\n",
    "        added_reward, done = self.reward_engineering(original_state, actions, state, done)\n",
    "        reward += added_reward\n",
    "\n",
    "        return state, reward, done, info\n",
    "\n",
    "    # New function to calculate additional rewards, given state, action, and new_state\n",
    "    def reward_engineering(self, state:tuple, actions:int, next_state:tuple, done):\n",
    "        r = 0\n",
    "        \n",
    "        boxPushed = (state[2], state[3]) != (next_state[2], next_state[3]) # True if box has moved after the action\n",
    "        agent_pos, agent_pos_x, agent_pos_y = state[0:2], state[1], state[0] # Agent pos in high level labels\n",
    "        agent_next_pos, agent_next_pos_x, agent_next_pos_y = next_state[0:2], next_state[1], state[0] # Agent new pos in high level labels\n",
    "        box_next_pos, box_next_pos_x, box_next_pos_y = next_state[2:], next_state[3], next_state[2] # Box new pos in high level labels\n",
    "        box_pos, box_pos_x, box_pos_y = state[2:], state[3], state[2] # Box pos in high level labels\n",
    "        \n",
    "        if not boxPushed:\n",
    "            self.turns_since_box_pushed += 1\n",
    "        else:\n",
    "            self.turns_since_box_pushed = 0\n",
    "                    \n",
    "        if not boxPushed:\n",
    "                r += -5\n",
    "\n",
    "        if not boxPushed and self.turns_since_box_pushed > 3: # Steps not pushing box are less valuable than steps that pushes box\n",
    "            r += -1\n",
    "\n",
    "        if abs(agent_pos_x - box_pos_x) == 1 and abs(agent_pos_y - box_pos_y) == 1 and (abs(agent_next_pos_x - box_next_pos_x) > 1 or abs(agent_next_pos_y - agent_next_pos_y) > 1) and not boxPushed:\n",
    "            r += self.walking_away_from_box_stupidly # Trying to walk away from box along diagonal axis\n",
    "            done = True\n",
    "\n",
    "        if abs(state[0] - state[2]) == 0 and abs(state[1] - next_state[1]) > 0 and not boxPushed:\n",
    "            r += self.walking_away_from_box_stupidly # Trying to walk away from box along horrizontal axis\n",
    "            done = True\n",
    "\n",
    "        if abs(state[1] - state[3]) == 0 and abs(state[0] - next_state[0]) > 0 and not boxPushed:\n",
    "            r += self.walking_away_from_box_stupidly # Trying to walk away from box along vertical axis\n",
    "            done = True\n",
    "\n",
    "        if next_state == state:\n",
    "            r += self.make_stupid_decision_reward_walking_to_wall # Trying to walk into wall\n",
    "            # r += -1\n",
    "            # done = True\n",
    "        return r, done\n",
    "    # NOTICE!!! --- Custom Code ENDS HERE ----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GKUvcFzIY4IX"
   },
   "source": [
    "Here is one example random agent class:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "executionInfo": {
     "elapsed": 348,
     "status": "ok",
     "timestamp": 1676469978101,
     "user": {
      "displayName": "Longtao Zheng",
      "userId": "17721972102166804099"
     },
     "user_tz": -480
    },
    "id": "MjhmwF8VpVki"
   },
   "outputs": [],
   "source": [
    "class RandomAgent:\n",
    "    def __init__(self, env, num_episodes):\n",
    "        self.action_space = [1, 2, 3, 4]\n",
    "        self.env = env\n",
    "        self.num_episodes = num_episodes\n",
    "\n",
    "    def act(self):\n",
    "        \"\"\"Returns a random choice of the available actions\"\"\"\n",
    "        return np.random.choice(self.action_space)\n",
    "\n",
    "    def learn(self):\n",
    "        rewards = []\n",
    "        \n",
    "        for _ in range(self.num_episodes):\n",
    "            cumulative_reward = 0 # Initialise values of each game\n",
    "            state = self.env.reset()\n",
    "            done = False\n",
    "            while not done: # Run until game terminated\n",
    "                action = self.act() \n",
    "                next_state, reward, done, info = self.env.step(action)\n",
    "                cumulative_reward += reward\n",
    "                state = next_state\n",
    "            rewards.append(cumulative_reward)\n",
    "\n",
    "        return rewards"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QYpyzvV8Y7j6"
   },
   "source": [
    "# Assignment Requirement 1: You need to complete the learn() method of the following class to implement your RL algorithm.\n",
    "\n",
    "Solution:\n",
    "The RL Learning algorithm are written within the RLAgent class. It provides learning algorithm as per the Q-Learning taught in class, as well as the modified Q-Learning which I named it pQ-Learning in this assignment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "executionInfo": {
     "elapsed": 14,
     "status": "ok",
     "timestamp": 1676463847348,
     "user": {
      "displayName": "Longtao Zheng",
      "userId": "17721972102166804099"
     },
     "user_tz": -480
    },
    "id": "vAKVvHmOptDW"
   },
   "outputs": [],
   "source": [
    "import math\n",
    "class RLAgent:\n",
    "    def __init__(self, env, num_episodes, epsilon=0.01, alpha=0.01, gamma=0.99, p=2, r = 0.1):\n",
    "        self.action_space = env.action_space\n",
    "        self.q_table = dict() # Store all Q-values in a dictionary\n",
    "        \n",
    "        # Loop through all possible grid spaces, create sub-dictionary for each\n",
    "        for agent_x in range(env.world_height):\n",
    "            for agent_y in range(env.world_width):\n",
    "                for box_x in range(env.world_height):\n",
    "                    for box_y in range(env.world_width):\n",
    "                        # Populate sub-dictionary with zero values for possible moves\n",
    "                        self.q_table[(agent_x, agent_y, box_x, box_y)] = {k: 0 for k in self.action_space}\n",
    "                        \n",
    "        # By default Q-Learning updates only compares the quality of its current state, current action vs the quality of new state and possible new actions \n",
    "        # My Q-Learning with modified terms:\n",
    "        # consider updates based on quality of past states in trajectory vs this update states\n",
    "        self.rho_matrix = np.array([r ** i for i in range(p-1,-1,-1)]).reshape(1,-1)\n",
    "        # self.rho_matrix = np.array([1 if i == 0 else r ** (i-0.75*(i-1)) for i in range(p-1,-1,-1)]).reshape(1,-1)\n",
    "        self.env = env\n",
    "        self.p = p\n",
    "        self.num_episodes = num_episodes\n",
    "        self.epsilon = epsilon\n",
    "        self.initial_epsilon = epsilon\n",
    "        self.alpha = alpha\n",
    "        self.gamma = gamma\n",
    "        \n",
    "    def act(self, state):\n",
    "        \"\"\"Returns the (epsilon-greedy) optimal action from Q-Value table.\"\"\"\n",
    "        if np.random.uniform(0,1) < self.epsilon:\n",
    "            action = self.action_space[np.random.randint(0, len(self.action_space))]\n",
    "        else:\n",
    "            q_values_of_state = self.q_table[state]\n",
    "            maxValue = max(q_values_of_state.values())\n",
    "            action = np.random.choice([k for k, v in q_values_of_state.items() if v == maxValue])\n",
    "        return action\n",
    "\n",
    "    def learn(self, learning_algorithm = \"custom q\"):\n",
    "\n",
    "        \"\"\"Updates Q-values iteratively.\"\"\"\n",
    "        print(\"Using reward engineering: \", self.env.rewardengineering)\n",
    "        rewards = []\n",
    "        if learning_algorithm == \"custom q-learning\" and self.env.rewardengineering == False:\n",
    "            for i in range(self.num_episodes):\n",
    "                H = np.zeros((1, self.p)).reshape(-1,1)            \n",
    "                cumulative_reward = 0 # Initialise values of each game\n",
    "                state = self.env.reset()\n",
    "                done = False\n",
    "                while not done: # Run until game terminated\n",
    "                    action = self.act(state)\n",
    "                    next_state, reward, done, info = self.env.step(action)\n",
    "                    cumulative_reward += reward\n",
    "                    q_values_of_new_state = self.q_table[next_state]\n",
    "                    q_value_of_this_state_action = self.q_table[state][action]\n",
    "\n",
    "                    # pop the last P th historical Q value of index 0, add the current Q value, # reshape to (p, 1)\n",
    "                    H = np.append(np.delete(H,0), q_value_of_this_state_action).reshape(-1,1)\n",
    "                    rht = self.rho_matrix.dot(H).item()\n",
    "                    self.q_table[state][action] = q_value_of_this_state_action + self.alpha * (reward + self.gamma * max(q_values_of_new_state.values()) - rht)\n",
    "                    state = next_state\n",
    "                rewards.append(cumulative_reward)\n",
    "                \n",
    "        elif learning_algorithm == \"custom q-learning\" and self.env.rewardengineering == True:\n",
    "            \n",
    "            for i in range(self.num_episodes):\n",
    "                H = np.zeros((1, self.p)).reshape(-1,1)            \n",
    "                cumulative_reward = 0 # Initialise values of each game\n",
    "                state = self.env.reset()\n",
    "                done = False\n",
    "                #step = 0\n",
    "                while not done: # Run until game terminated\n",
    "                    action = self.act(state)\n",
    "                    next_state, reward, done, info = self.env.step_w_re(action, state)\n",
    "                    cumulative_reward += reward\n",
    "                    q_values_of_new_state = self.q_table[next_state]\n",
    "                    q_value_of_this_state_action = self.q_table[state][action]\n",
    "\n",
    "                    # pop the last P th historical Q value of index 0, add the current Q value, # reshape to (p, 1)\n",
    "                    H = np.append(np.delete(H,0), q_value_of_this_state_action).reshape(-1,1)\n",
    "                    rht = self.rho_matrix.dot(H).item()\n",
    "                    self.q_table[state][action] = q_value_of_this_state_action + self.alpha * (reward + self.gamma * max(q_values_of_new_state.values()) - rht)\n",
    "                    state = next_state\n",
    "                rewards.append(cumulative_reward)\n",
    "        \n",
    "        else: # Normal Q-Learning\n",
    "            for i in range(self.num_episodes):          \n",
    "                cumulative_reward = 0 # Initialise values of each game\n",
    "                state = self.env.reset()\n",
    "                done = False\n",
    "                step = 0\n",
    "                while not done: # Run until game terminated\n",
    "                    action = self.act(state)\n",
    "                    next_state, reward, done, info = self.env.step(action)\n",
    "                    cumulative_reward += reward\n",
    "                    q_values_of_new_state = self.q_table[next_state]\n",
    "                    q_value_of_this_state_action = self.q_table[state][action]\n",
    "                    self.q_table[(state)][action] = q_value_of_this_state_action + self.alpha * (reward + self.gamma * max(q_values_of_new_state.values()) - q_value_of_this_state_action)\n",
    "                    state = next_state\n",
    "                rewards.append(cumulative_reward)\n",
    "        return rewards\n",
    "    \n",
    "# Original Learning Curve code shifted here for code optimization\n",
    "def plot_smooth_curve(weight:float, rewards : list, description : str):\n",
    "    # Smooth plot\n",
    "    weight=0.99\n",
    "    last = rewards[0]\n",
    "    smoothed = []\n",
    "    for v in rewards:\n",
    "        smoothed_val = last * weight + (1 - weight) * v\n",
    "        smoothed.append(smoothed_val)\n",
    "        last = smoothed_val\n",
    "\n",
    "    # Plot the learning curve\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.plot(smoothed)\n",
    "    ax.set_title(description)\n",
    "    ax.set_xticks(np.arange(0, 75000, 10000))\n",
    "    ax.set_yticks(np.arange(-2500, 700, 100))\n",
    "    ax.grid()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "T4Zp2jCTZEgw"
   },
   "source": [
    "Here is the game interface where you can manually move the agent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10398,
     "status": "ok",
     "timestamp": 1676463857733,
     "user": {
      "displayName": "Longtao Zheng",
      "userId": "17721972102166804099"
     },
     "user_tz": -480
    },
    "id": "OK-cjWJA7ymv",
    "outputId": "5ed9e3b2-c2de-4808-d867-e2cb04c9df21"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "Please input the actions (up: 1, down: 2, left: 3, right: 4): f\n"
     ]
    },
    {
     "ename": "ValueError",
     "evalue": "invalid literal for int() with base 10: 'f'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "Cell \u001b[1;32mIn[5], line 8\u001b[0m\n\u001b[0;32m      5\u001b[0m rewards \u001b[38;5;241m=\u001b[39m []\n\u001b[0;32m      7\u001b[0m \u001b[38;5;28;01mwhile\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m done:\n\u001b[1;32m----> 8\u001b[0m     action \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mint\u001b[39;49m\u001b[43m(\u001b[49m\u001b[38;5;28;43minput\u001b[39;49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mPlease input the actions (up: 1, down: 2, left: 3, right: 4): \u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m      9\u001b[0m     state, reward, done, info \u001b[38;5;241m=\u001b[39m env\u001b[38;5;241m.\u001b[39mstep(action)\n\u001b[0;32m     10\u001b[0m     rewards\u001b[38;5;241m.\u001b[39mappend(reward)\n",
      "\u001b[1;31mValueError\u001b[0m: invalid literal for int() with base 10: 'f'"
     ]
    }
   ],
   "source": [
    "env = CliffBoxGridWorld(render=True)\n",
    "env.reset()\n",
    "env.print_world()\n",
    "done = False\n",
    "rewards = []\n",
    "\n",
    "while not done:\n",
    "    action = int(input(\"Please input the actions (up: 1, down: 2, left: 3, right: 4): \"))\n",
    "    state, reward, done, info = env.step(action)\n",
    "    rewards.append(reward)\n",
    "    print(f'step: {env.timesteps}, state: {state}, actions: {action}, reward: {reward}')\n",
    "    env.print_world()\n",
    "\n",
    "print(f'rewards: {sum(rewards)}')\n",
    "print(f'action history: {env.action_history}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Example code to step random agent in the environment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x11490fa22c0>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Initialize the environment and agent\n",
    "env = CliffBoxGridWorld()\n",
    "agent = RandomAgent(env, num_episodes=1000)\n",
    "rewards = agent.learn()\n",
    "\n",
    "# Plot the learning curve\n",
    "plt.plot(rewards)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assignment Requirement 2: You need to complete the following method to visualize your training results.\n",
    "\n",
    "With the Q table learned by the RL Learning algorithm above:\n",
    "\n",
    "For each state $s(agentPos, boxPos)$  \n",
    "The policy of agent at each state in the game is described as:\n",
    "$\\pi (s)=argmax_{a}Q(s, a)$  \n",
    "$\\pi(s)$ is used to visualize the optimal policy step by step, in the following visualize function.  \n",
    "\n",
    "And the V value of each state $s(AgentPos, BoxPos)$ is:\n",
    "$v_{\\pi}(s)=q_{\\pi}(s,\\pi (s))$  \n",
    " \n",
    "\n",
    "As per requested, the V value of each agent pos state $s^{'}(agentPos)$ is the average of all $v_{\\pi}(agentPos)$ at all possible boxPos position to form the V-table:\n",
    "$v^{'}(s^{'})=average(v_{\\pi}(agentPos,boxPos))$ for all **boxPos**, when given a **agentPos**\n",
    "\n",
    "\n",
    "The Function's implementation code as follows, "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize(agent):\n",
    "    # TODO: Visualize learned V-table and policy.\n",
    "    # for ... in q_table.items():\n",
    "    #     v_table[key[0]].append(np.mean(list(value.values())))\n",
    "    #     policy[key[0]].append(actions[np.argmax([value[i] for i in range(1, 5)])])\n",
    "    q_table = agent.q_table\n",
    "    world_key = [(y, x) for y in range(agent.env.world_height) for x in range(agent.env.world_width)]\n",
    "    v_table_with_box = {}\n",
    "    policy_table = {}\n",
    "    v_table = {}\n",
    "    policy_pi = {}\n",
    "    \n",
    "    for key, value in q_table.items():\n",
    "        agent_key = key[0:2]\n",
    "        if key not in policy_pi.keys():\n",
    "            policy_pi[key] = -1\n",
    "        if key not in v_table_with_box.keys():\n",
    "            v_table_with_box[key] = 0\n",
    "        if agent_key not in v_table.keys():\n",
    "            v_table[agent_key] = []\n",
    "        #print(\"koi: \", agent_key)    \n",
    "        actions = list(q_table[key].keys())\n",
    "        policy_pi[key] = actions[np.argmax([value[i] for i in range(1, 5)])]\n",
    "        v_table[agent_key].append(q_table[(key)][policy_pi[key]])\n",
    "        #v_table[agent_key].append(np.max(list(value.values())))\n",
    "        \n",
    "    for agent_key, value in v_table.items():\n",
    "        v_table[agent_key] = np.mean(v_table[agent_key])\n",
    "    visualize_v(v_table, agent.env)\n",
    "    visualize_policy(q_table, agent.env)\n",
    "    return v_table, policy_pi\n",
    "\n",
    "def visualize_v(v_table, env):\n",
    "    V_y = {}\n",
    "    print(\"V table:\" )\n",
    "    for y in range(env.world_height):\n",
    "        V_y[y] = []\n",
    "        keys = np.array(list(v_table.keys()))\n",
    "        mask = keys[:,0] == y\n",
    "\n",
    "        interested_keys = keys[mask]\n",
    "        for k in interested_keys:\n",
    "            V_y[y].append(round(v_table[tuple(k)],2))\n",
    "        print(f\"{y} {V_y[y]}\")\n",
    "    return V_y\n",
    "    \n",
    "\n",
    "def visualize_policy(q_table, oldenv):\n",
    "    print(\"Learned policy:\" )\n",
    "    env = CliffBoxGridWorld(render=True, rewardengineering=oldenv.rewardengineering)\n",
    "    state = env.reset()\n",
    "    env.print_world()\n",
    "    done = False\n",
    "    rewards = []\n",
    "    while not done:\n",
    "        best_action = np.argmax([q_table[state][i] for i in range(1,5)])\n",
    "        action_list = list(q_table[state].keys())\n",
    "        action = action_list[best_action]\n",
    "\n",
    "        # action = int(input(\"Please input the actions (up: 1, down: 2, left: 3, right: 4): \"))\n",
    "        if oldenv.rewardengineering:\n",
    "            state, reward, done, info = env.step_w_re(action, state)\n",
    "        else:\n",
    "            state, reward, done, info = env.step(action)\n",
    "        rewards.append(reward)\n",
    "        print(f'step: {env.timesteps}, state: {state}, actions: {action}, reward: {reward}')\n",
    "        env.print_world()\n",
    "\n",
    "    print(f'rewards: {sum(rewards)}')\n",
    "    print(f'action history: {env.action_history}')        \n",
    " \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1NbkDcyfZO14"
   },
   "source": [
    "# Train your own agent!\n",
    "\n",
    "## About this assignment\n",
    "\n",
    "During the evaluation of models, a model is said to have reach optimal convergence when it achieves the optimal trajectory it can possibly achieve. This optimal trajectory might not be the best of all methods.\n",
    "\n",
    "For example:\n",
    "The optimal trajectory for the below implemented Traditional Q-Learning is:  \n",
    "```[(5, 0, 4, 1), (5, 1, 4, 1), (4, 1, 3, 1), (3, 1, 2, 1), (2, 1, 1, 1), (2, 0, 1, 1), (1, 0, 1, 1), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```  \n",
    "\n",
    "The above trajectory would yield the greatest cumalative reward within one episode. (Cumulative Reward: 642, w/o Reward Engineering ) \n",
    "It is also the best possible trajectory possible for this project's problem. However, not all model will converge with a policy table that allows us to derive the above best possible trajectory like the above.\n",
    "\n",
    "Some might end up with another optimal trajectory that might takes longer steps (Lesser Cumulative Reward) than the above such as:  \n",
    "\n",
    "```[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```\n",
    "\n",
    "The above trajectory would have yield cumulative reward of 623.\n",
    "\n",
    "# First RL Implementation - Reinforcement Learning using Traditional Q-Learning taught by Prof Bo An\n",
    "\n",
    "The below is a direct implementation of what was taught in Lecture 5 - Q-Learning. Learn rate $\\alpha$ was deliberately set to 0.01 to increase the duration of training so that it can be benchmarked against the customized RL model I developed later. (Even though, the Q-Learning is able to learn the optimal policy in 6000 episodes at $\\alpha=0.1$ )\n",
    "\n",
    "With $\\alpha=0.01$, the Q-Learning Model converges at **55000** episodes and completed 75000 episodes in **330** seconds.  \n",
    "The agent reached saturation at around **55000** episodes in about approx **240** seconds ($330sec \\times 55000episodes \\div 75000episodes$)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "bdc2VpjA71Kw"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using reward engineering:  False\n",
      "Total time taken:  303.97752046585083\n"
     ]
    }
   ],
   "source": [
    "# Traditional Q-Learning\n",
    "import time\n",
    "\n",
    "start = time.time()\n",
    "episodes = 75000\n",
    "env = CliffBoxGridWorld(rewardengineering=False)\n",
    "env.reset()\n",
    "agent1 = RLAgent(env, num_episodes=episodes, epsilon=0.01, alpha=0.01, gamma=0.99)\n",
    "rewards_trad= agent1.learn(learning_algorithm = \"q-learning\")\n",
    "time_taken = time.time() - start\n",
    "print(\"Total time taken: \", time_taken)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Learning Curve Visualization of traditional Q-Learning. Model converges at around 55000 episodes and completed 75000 episodes in about 330 seconds.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "description = \"Learning Progress Trad Q-Learning (Episode wards (Y axis) vs Episodes (X axis))\"\n",
    "plot_smooth_curve(weight=0.99, rewards=rewards_trad, description=description)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Learned V-table and Policy (Trad Q-Learning)\n",
    "\n",
    "The optimal converged trajectory is:  \n",
    "\n",
    "```[(5, 0, 4, 1), (5, 1, 4, 1), (4, 1, 3, 1), (3, 1, 2, 1), (2, 1, 1, 1), (2, 0, 1, 1), (1, 0, 1, 1), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```  \n",
    "\n",
    "Final cumulative reward:**642**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V table:\n",
      "0 [-35.56, -35.97, -36.93, -34.22, -30.42, -27.13, 0.0, 0.0, -29.82, -25.24, -24.34, -16.8, -14.59, -8.16]\n",
      "1 [-27.6, -25.0, -26.05, -26.07, -26.37, -25.1, 0.0, 0.0, -25.14, -15.02, -14.72, -15.3, -10.71, -6.97]\n",
      "2 [-30.8, -25.42, -31.47, 0.0, -28.48, -24.64, 0.0, 0.0, -28.07, -20.24, -15.88, -20.93, 0.0, -10.88]\n",
      "3 [-34.05, -26.0, -35.07, 0.0, -27.69, -25.04, 0.0, -29.74, -28.73, -20.26, -16.49, 0.0, 0.0, -21.4]\n",
      "4 [-34.89, -25.63, -33.9, 0.0, -24.5, -23.48, -22.1, -20.68, -19.12, -18.27, -17.86, 0.0, 0.0, -20.39]\n",
      "5 [-31.29, -28.7, -37.15, 0.0, -33.3, -33.52, -32.88, -32.25, -27.96, -22.88, -18.15, 0.0, 0.0, -19.71]\n",
      "Learned policy:\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 1, state: (5, 1, 4, 1), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 2, state: (4, 1, 3, 1), actions: 1, reward: -15\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 3, state: (3, 1, 2, 1), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 4, state: (2, 1, 1, 1), actions: 1, reward: -17\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 5, state: (2, 0, 1, 1), actions: 3, reward: -18\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 6, state: (1, 0, 1, 1), actions: 1, reward: -17\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 7, state: (1, 1, 1, 2), actions: 4, reward: -16\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 8, state: (1, 2, 1, 3), actions: 4, reward: -15\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 9, state: (1, 3, 1, 4), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 10, state: (1, 4, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 11, state: (0, 4, 1, 5), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 12, state: (0, 5, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 13, state: (1, 5, 2, 5), actions: 2, reward: -12\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 14, state: (2, 5, 3, 5), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 15, state: (3, 5, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 16, state: (3, 4, 4, 5), actions: 3, reward: -11\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 17, state: (4, 4, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 18, state: (4, 5, 4, 6), actions: 4, reward: -9\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 19, state: (4, 6, 4, 7), actions: 4, reward: -8\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 20, state: (4, 7, 4, 8), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 21, state: (4, 8, 4, 9), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 22, state: (4, 9, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 23, state: (5, 9, 4, 10), actions: 2, reward: -6\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_']]\n",
      "step: 24, state: (5, 10, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_']]\n",
      "step: 25, state: (4, 10, 3, 10), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'B' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 26, state: (3, 10, 2, 10), actions: 1, reward: -7\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 27, state: (2, 10, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 28, state: (2, 9, 1, 10), actions: 3, reward: -9\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'A' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 29, state: (1, 9, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'A' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 30, state: (1, 10, 1, 11), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'B' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 31, state: (1, 11, 1, 12), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'A' b'B' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 32, state: (1, 12, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 33, state: (0, 12, 1, 13), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 34, state: (0, 13, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 35, state: (1, 13, 2, 13), actions: 2, reward: -4\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 36, state: (2, 13, 3, 13), actions: 2, reward: -3\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 37, state: (3, 13, 4, 13), actions: 2, reward: 998\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "rewards: 642\n",
      "action history: [4, 1, 1, 1, 3, 1, 4, 4, 4, 4, 1, 4, 2, 2, 2, 3, 2, 4, 4, 4, 4, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2]\n"
     ]
    }
   ],
   "source": [
    "# Visualization of V-Table and Policy from the direct implementation of Q-Learning model taught in AI6101 Lecture\n",
    "v_table, policy = visualize(agent1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following class defines the grid world environment. The grid world looks like:\n",
    "\n",
    "\n",
    "# Problem with Traditional Q Learning\n",
    "\n",
    "\n",
    "During training, I notice that the learning process at near certain position such as $State(4,7,3,7)$ can be really inefficient with current reward scheme which is too simple to tackle certain problems quickly. Example:\n",
    "\n",
    "## Step 0:\n",
    "State Current: (5,6,4,7) Provided Action: Right\n",
    "```\n",
    "____________________________________________________________\n",
    "|___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "|_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "|_3_|___|___|___|_x_|___|___|_x_|___|___|___|___|_x_|_x_|___|\n",
    "|_4_|___|___|___|_x_|___|___|___|_B_|___|___|___|_x_|_x_|_G_|\n",
    "|_5_|___|___|___|_x_|___|___|_A_|___|___|___|___|_x_|_x_|___|\n",
    "```\n",
    "\n",
    "State New: (5,7,4,7)\n",
    "```\n",
    "____________________________________________________________\n",
    "|___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "|_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "|_3_|___|___|___|_x_|___|___|_x_|___|___|___|___|_x_|_x_|___|\n",
    "|_4_|___|___|___|_x_|___|___|___|_B_|___|___|___|_x_|_x_|_G_|\n",
    "|_5_|___|___|___|_x_|___|___|___|_A_|___|___|___|_x_|_x_|___|\n",
    "```\n",
    "\n",
    "Update $Q(S_{0},A_{0})$ based on State New. Updates is insignificant which is fine. Agent have not make a critical mistake that will ruin the game yet. However, the following step as described below is different.\n",
    "\n",
    "\n",
    "\n",
    "## Step 1:\n",
    "State Current: (5,7,4,7) Provided Action: Up $A_{1}=Up$  \n",
    "```\n",
    "____________________________________________________________\n",
    "|___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "|_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "|_3_|___|___|___|_x_|___|___|_x_|___|___|___|___|_x_|_x_|___|\n",
    "|_4_|___|___|___|_x_|___|___|___|_B_|___|___|___|_x_|_x_|_G_|\n",
    "|_5_|___|___|___|_x_|___|___|___|_A_|___|___|___|_x_|_x_|___|\n",
    "```\n",
    "\n",
    "State New: (4,7,3,7)\n",
    "```\n",
    "____________________________________________________________\n",
    "|___|_0_|_1_|_2_|_3_|_4_|_5_|_6_|_7_|_8_|_9_|_10|_11|_12|_13|\n",
    "|_0_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_1_|___|___|___|___|___|___|_x_|_x_|___|___|___|___|___|___|\n",
    "|_2_|___|___|___|_x_|___|___|_x_|_x_|___|___|___|___|_x_|___|\n",
    "|_3_|___|___|___|_x_|___|___|_x_|_B_|___|___|___|_x_|_x_|___|\n",
    "|_4_|___|___|___|_x_|___|___|___|_A_|___|___|___|_x_|_x_|_G_|\n",
    "|_5_|___|___|___|_x_|___|___|___|___|___|___|___|_x_|_x_|___|\n",
    "```\n",
    "\n",
    "The problem: \n",
    "\n",
    "Here the agent have make a move that have effectively ruined the game as the box is trapped at coordinate (3,7) now and we cant further push the box any further out of (3,7). The box is stuck indefinitely and can no longer reach the goal position, unless the game rule allow the agent to pull the box rather than just pushing.\n",
    "\n",
    "However, update $Q(S_{1},A_{1})$ based on State New is still:\n",
    "\n",
    "$Q_{new}(S_{1},A_{1}) = Q(S_{1},A_{1})+\\alpha [R+\\gamma max_{a}Q(S_{1+1},a)- Q(S_{1},A_{1})]$ is a insignificant update even though the process of $(S_{1},A_{1})$ is extremely undesirable.\n",
    "\n",
    "As per the reward formulation, since the Box nor Agent have fell into the cliff, the Agent is not significantly far from the Agent, distance between Goal and Box/Agent is fine, scheme is very insignificant sum. Therefore relying on reward term alone will takes time before the model finds out that $Q(S_{1},A_{1})$ is an extremely undesirable state/action combination that will will just lead to endless situation of increasing timesteps with no progress.\n",
    "\n",
    "Idea:\n",
    "\n",
    "Currently as per Q-Learning Bellman's Equation formulation, $Q(S_{1},A_{1})$ only rely on Step 1's Reward term, plus the difference between value at new state, $argmax_{a}Q(S_{1}, a)$ and Step 1's original current state and action.  \n",
    "\n",
    "We could make the model better by penalising $Q(S_{1},A_{1})$ greatly ASAP to speed up the learning process.  \n",
    "\n",
    "We also know from manual observation, the update penalty of $Q(S_{1},A_{1})$ should be extremely bad as $S_{1},A_{1}$  since the situation at 2 time steps ago was much more desirable. (Refer to step 0 current state and new state). This term was not consider in the update of this Q value.  \n",
    "\n",
    "Through this assignment I realise the model can learn how to handle such scenario faster, more specifically by learning certain tricky optimal $Q$ that are harder to learn like $Q(S_{1},A_{1})$ faster. This can be done by observing the $V$ value at $S_{new}$ is significant less valuable than current $S_{1}$ just like in the original Q-Learning algorithm, **togther with past tense before recent updates' $S_{1-1}$ and, $S_{1-2}$, etc, etc.** Information related to the states comparison can be stored in a list object from previous steps and used in later steps to update Q values by comparing Q  value distant past and new states from this current states/actions. In the next section of the repoort, I will discuss more how I implement the customized version of the original Q-Learning.  \n",
    "\n",
    "\n",
    "## Solution Development of my own Custom Q-Learning - Incorporating Data-Centric AI methodology into Q-Learning Reinforcement Learning\n",
    "\n",
    "### Instead of learning with only Current state and action, learn with History Past data as training data as well\n",
    "\n",
    "Instead of just Learning the Regret from current State and Action produced Resulting Observation, Learn the Regret with both historical past and current \"States / Actions produced Resulting Observation\":  \n",
    "\n",
    "For this project I propose the idea to modify the Q-value update rule by implementing new terms to compare the historical states/actions at state 0 or/and older past, $p-1$ states/actions.\n",
    "\n",
    "For example, given $p=3$, to consider current state/action plus two historical past state/action, given a current state at step $t$:  \n",
    "If the new state/action V value, $argmax_{a}Q(S_{t+1},a)$ is way less desirable than the older steps' $Q$ such as, $Q(S_{t},A_{t})$ and/or $Q(S_{t-1},A_{t-1})$ then the update of $Q(S_{t},A_{t})$ should receive a more significant negative reward without hesistant.\n",
    "\n",
    "The Q Learning's Q Table update for a current State/Action $(S_{t},A_{t})$ with the Customized Q-Learning Bellman Equation formulation is modified as followed:\n",
    "\n",
    "$Q_{new}(S_{t},A_{t}) = Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)- P^{T}H]$\n",
    "\n",
    "Where given two new constant parameters $r$ (usually 0.1 in the experiment) is the weight decay factor, and $p$ is the number of historical pasts to consider, we form the following two vectors:  \n",
    "$P$ is a p by 1 vector where $P=[\\rho _{0}=r^{p},  \\rho _{1}=r^{p-1},  ...,   \\rho _{p-1}=r^{p-p}=1.0]^{T}$  \n",
    "$H$ is a p by 1 vector where $H=[Q(S_{t-p},A_{t-p}),Q(S_{t-p+1},A_{t-p+1}),  ...,  Q(S_{t},A_{t})]^{T}$\n",
    "\n",
    "The new hyperparameters $p$ must be: $1<=p$ and $r$ must be: $0<r<1.0$. The above 2 are essentially regularization hyperparameters to control how seriously the model refer to the past states in additional to the original Bellman's equation. $H$ is a vector that stores the historical terms of the new state which also includes the current $(Q(S_{t}, A_{t})$ term which is also present in the original Q-Learning Bellman equation. The $H$ vector is essential the training vector generated at previous episodes and steps for this current step Q update. \n",
    "\n",
    "For the rest of the assignment I refer to this model as the **pQ-Learning** RL learning algorithm, where the additional p in front stand from \"past\" or \"history\" Q-Learning. \n",
    "\n",
    "\n",
    "### DRAWBACKS:\n",
    "Even though the Custom pQ-Learning I implemented can achieve convergence way more quicker than Q-Learning, using the current Reward scheme as per the assignment's default MDP formulation, we can note that the model is unable to converge at the best possible optimal trajectory.\n",
    "\n",
    "\n",
    "During experiment, the trajectory of the converged episodes of the Custom p Q-Learning is:  \n",
    "\n",
    "```[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```  \n",
    "\n",
    "Above trajectory yields Cumulative Reward: **623**  \n",
    "This is unfortunately not the optimal shortest trajectory possible. (Refer to the optimal trajectory of Traditional Q-Learning for better optimal which has Cumulative Reward of **642**). Optimally best, the first two steps should be (Right to 5,1,4,1) followed by (Up to 4,1,3,1).\n",
    "\n",
    "First few steps in the episode might not be optimal as the Bellman's Q Value do not have any meaningful past steps Q value to take reference from last subsequent steps. Currently implementation initialize H as a 0 vector for the first step update which might not be best implementation. Nonetheless, any subsequent movement after the first 2 steps are optimally best possible.\n",
    "\n",
    "In order to achieve the best possible optimal trajectory, later in this project, I also conducted reward engineering on the current reward scheme such that it reward agent based on smart or stupid actions rather than just based on the direct distance of box/agents and the timesteps it takes to reach the goal. This will also allow us to converge at the best possible optimal trajectory.   \n",
    "\n",
    "\n",
    "### Final Algorithm Design of pQ-Learning:\n",
    "\n",
    "&nbsp; &nbsp; Algorithm parameters: step size $\\alpha$  $\\epsilon$  $[0,1]$, small $\\varepsilon$ > $0$  \n",
    "&nbsp; &nbsp; Initialize $Q(s,a)$, for all $s$ $\\epsilon$ $S^{+}$, $a$ $\\epsilon$ $A(s)$, arbitrarily except that $Q(terminal,.)=0$  \n",
    "&nbsp; &nbsp; Initialize Weight Decaying constant vector $P=[\\rho _{0}=r^{p},  \\rho _{1}=r^{p-1},  ...,   \\rho _{p-1}=r^{p-p}=1.0]^{T}$  \n",
    "\n",
    "&nbsp; &nbsp; Loop for each episode:  \n",
    "&nbsp; &nbsp; &nbsp; Initialize env $S$  \n",
    "&nbsp; &nbsp; &nbsp; Initialize Initialize $H$ as zero vector of shape $p$ by 1.  $H$ $\\epsilon$ $\\mathbb{R}^{p}$  \n",
    "&nbsp; &nbsp; &nbsp; Loop for each step of episode:  \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; Choose $A$ from $S$ using policy derived from $Q$ (From default code's $\\varepsilon$ - greedy)  \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; Update $P$ records by deleting oldest (first) record and appending the $Q$ value of current state/action:  \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; $H \\leftarrow H[1:].append(Q(S,A))$  \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; Take action $A$, observe $R$, $S^{'}$  \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; $Q(S,A)\\leftarrow Q(S,A) + \\alpha [R+ \\gamma max_{a} Q(S^{'}, a) - P^{T}H]$   \n",
    "&nbsp; &nbsp; &nbsp; &nbsp; $S\\leftarrow S^{'}$  \n",
    "&nbsp; &nbsp; &nbsp; until $S$ is terminal\n",
    "\n",
    "\n",
    "**Implementation Experiment is as followed:**\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Second RL Implementation: Testing my Custom pQ-Learning\n",
    "## We test the new Q update formulation by setting p=2 \n",
    "When $p=2$, this means we consider historical $Q(S_{t-1}, A_{t-1})$ and current which the original Q learning emcompass, $Q(S_{t}, A_{t})$ )\n",
    "\n",
    "Customized Q Learning Bellman Equation formulation:  \n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-1},A_{t-1})-r_{1}Q(S_{t},A_{t} )]$  \n",
    "where the last $r: r_{p-1}=r_{1}=1.0$ for all given p constant\n",
    "\n",
    "In the run time below, my Custom p=2 Q-Learning take approx **350 to 400** seconds (fluctuates around) to execute 75000 episodes at a rate of about **188 episodes/second**. The agent reached saturation at around **34000** episodes in about **180** seconds.  (Approx: $34000 episodes \\times 380sec \\div 75000 episodes$).\n",
    "\n",
    "We can notice that the extra computation does not result in significant increased in training duration to complete 75000 episodes. This is thanks to Modern Computing packages like Numpy. NumPy computing can use highly optimized techniques to speed up matrix operations. This helps to moderate the training duration relative to original Q-Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using reward engineering:  False\n",
      "Total time taken:  358.3728187084198\n"
     ]
    }
   ],
   "source": [
    "# Guang Way's Custom Q-Learning\n",
    "import time\n",
    "\n",
    "start = time.time()\n",
    "episodes = 75000\n",
    "p=2\n",
    "r = 0.1\n",
    "env = CliffBoxGridWorld(rewardengineering=False)\n",
    "env.reset()\n",
    "agent2 = RLAgent(env, num_episodes=episodes, epsilon=0.01, alpha=0.01, gamma=0.99, p=p, r = r)\n",
    "rewards_custom = agent2.learn(learning_algorithm = \"custom q-learning\")\n",
    "time_taken = time.time() - start\n",
    "print(\"Total time taken: \", time_taken)\n",
    "#print(f\"Speed: {episodes / (time.time() - start)} episodes/sec\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### Learning Curve: Setting p=2, \n",
    "The model converges at **35000** episodes\n",
    "\n",
    "The $Q(S,A)$ update formulation are as followed in this training:\n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-1},A_{t-1})-r_{1}Q(S_{t},A_{t} )]$  \n",
    "The agent reached saturation at around **35000** episodes in about **180** seconds \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "description = 'Learning Progress pQ-Learning (Episode rewards (Y axis) vs Episodes (X axis)) p=2 r=0.1'\n",
    "plot_smooth_curve(weight=0.99, rewards=rewards_custom, description=description)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Learned V-table and Policy (custom pQ-learning with p=2 r=0.1)\n",
    "\n",
    "The optimal converged trajectory is not the best possible (in terms of cumulative reward). It is as follows:  \n",
    "\n",
    "```[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```\n",
    "\n",
    "Final cumulative reward: **623**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V table:\n",
      "0 [-17.48, -17.45, -17.78, -17.52, -17.72, -16.9, 0.0, 0.0, -13.59, -10.01, -9.06, -3.17, -2.8, 4.83]\n",
      "1 [-17.17, -15.72, -16.15, -16.53, -16.61, -16.14, 0.0, 0.0, -10.43, -4.7, -3.37, -1.35, 1.89, 7.0]\n",
      "2 [-17.33, -16.13, -16.03, 0.0, -16.28, -16.28, 0.0, 0.0, -11.69, -9.57, -5.83, -6.29, 0.0, 2.26]\n",
      "3 [-17.42, -15.79, -15.98, 0.0, -16.6, -15.76, 0.0, -14.93, -14.09, -7.36, -8.92, 0.0, 0.0, -7.74]\n",
      "4 [-17.01, -15.82, -15.79, 0.0, -15.02, -14.72, -14.1, -13.01, -11.84, -10.41, -9.13, 0.0, 0.0, -7.05]\n",
      "5 [-16.94, -17.01, -16.91, 0.0, -17.11, -16.96, -16.46, -15.68, -13.16, -12.81, -10.14, 0.0, 0.0, -6.58]\n",
      "Learned policy:\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 1, state: (4, 0, 4, 1), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'A' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 2, state: (4, 1, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'A' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 3, state: (5, 1, 4, 2), actions: 2, reward: -14\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 4, state: (5, 2, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 5, state: (4, 2, 3, 2), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 6, state: (3, 2, 2, 2), actions: 1, reward: -15\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 7, state: (2, 2, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 8, state: (2, 1, 1, 2), actions: 3, reward: -17\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 9, state: (1, 1, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 10, state: (1, 2, 1, 3), actions: 4, reward: -15\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 11, state: (1, 3, 1, 4), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 12, state: (1, 4, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 13, state: (0, 4, 1, 5), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 14, state: (0, 5, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 15, state: (1, 5, 2, 5), actions: 2, reward: -12\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 16, state: (2, 5, 3, 5), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 17, state: (3, 5, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 18, state: (3, 4, 4, 5), actions: 3, reward: -11\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 19, state: (4, 4, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 20, state: (4, 5, 4, 6), actions: 4, reward: -9\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 21, state: (4, 6, 4, 7), actions: 4, reward: -8\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 22, state: (4, 7, 4, 8), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 23, state: (4, 8, 4, 9), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 24, state: (4, 9, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 25, state: (5, 9, 4, 10), actions: 2, reward: -6\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_']]\n",
      "step: 26, state: (5, 10, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_']]\n",
      "step: 27, state: (4, 10, 3, 10), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'B' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 28, state: (3, 10, 2, 10), actions: 1, reward: -7\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 29, state: (2, 10, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 30, state: (2, 9, 1, 10), actions: 3, reward: -9\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'A' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 31, state: (1, 9, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'A' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 32, state: (1, 10, 1, 11), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'B' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 33, state: (1, 11, 1, 12), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'A' b'B' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 34, state: (1, 12, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 35, state: (0, 12, 1, 13), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 36, state: (0, 13, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 37, state: (1, 13, 2, 13), actions: 2, reward: -4\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 38, state: (2, 13, 3, 13), actions: 2, reward: -3\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 39, state: (3, 13, 4, 13), actions: 2, reward: 998\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "rewards: 623\n",
      "action history: [1, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2, 3, 2, 4, 4, 4, 4, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2]\n"
     ]
    }
   ],
   "source": [
    "# Visualization of V-Table and Policy from the direct implementation of custom pQ-Learning with p=2, r=0.1\n",
    "v_table2, policy2 = visualize(agent2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Learning with Custom Q where p=3, \n",
    "Setting $p = 3$, we notice tiny convergence rate improvement as relative to p=2.  \n",
    "\n",
    "Customized Q Learning Bellman Equation formulation: \n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-2},A_{t-2}) -r_{1}Q(S_{t-1},A_{t-1})-r_{2}Q(S_{t},A_{t} )]$  \n",
    "where the last $r: r_{p-1}=r_{2}=1.0$ for all given p constant\n",
    "\n",
    "Converges at **35000** episodes  and completed 75000 episodes also in about **350-400** seconds. The agent reached saturation at around **35000** episodes in about **180** seconds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using reward engineering:  False\n",
      "Total time taken:  348.51920676231384\n"
     ]
    }
   ],
   "source": [
    "# Guang Way's Custom Q-Learning\n",
    "import time\n",
    "\n",
    "start = time.time()\n",
    "episodes = 75000\n",
    "p=3\n",
    "r = 0.1\n",
    "env = CliffBoxGridWorld(rewardengineering=False)\n",
    "env.reset()\n",
    "agent3 = RLAgent(env, num_episodes=episodes, epsilon=0.01, alpha=0.01, gamma=0.99, p=p, r = r)\n",
    "rewards_custom3 = agent3.learn(learning_algorithm = \"custom q-learning\")\n",
    "time_taken = time.time() - start\n",
    "print(\"Total time taken: \", time_taken)\n",
    "# print(f\"Speed: {episodes / (time.time() - start)} episodes/sec\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Learning Curve: Setting p=3 also can allow us to achieve a tiny improvement. \n",
    "\n",
    "Q value update formulation:  \n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-2},A_{t-2}) -r_{1}Q(S_{t-1},A_{t-1})-r_{2}Q(S_{t},A_{t} )]$  \n",
    "\n",
    "The improvement achieved is not significant. The model converged in about **35000** episodes and completed 75000 episodes in **350-400** seconds. This indicate that: Given time step $t$, the historical record $t-2$ or older is nowhere as significant as $t-1$. The runtime of this run is the same as the case in $p=2$ and $p=4$. Increasing p constant moderately does not affect the overall runtime significantly. The agent reached saturation at around **35000** episodes in about **180** seconds "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "description = 'Learning Progress pQ-Learning (Episode rewards (Y axis) vs Episodes (X axis)) p=3 r=0.1'\n",
    "plot_smooth_curve(weight=0.99, rewards=rewards_custom3, description=description)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Learned V-table and Policy (custom pQ-learning with p=3 r=0.1)\n",
    "\n",
    "The optimal converged trajectory is not the best possible as well(in terms of cumulative reward). It is as follows:  \n",
    "\n",
    "```[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```\n",
    "\n",
    "Final cumulative reward: **623**\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V table:\n",
      "0 [-16.44, -16.47, -16.79, -16.61, -16.79, -16.22, 0.0, 0.0, -12.61, -9.07, -8.17, -2.45, -2.01, 5.75]\n",
      "1 [-16.25, -14.91, -15.29, -15.71, -15.82, -15.35, 0.0, 0.0, -8.68, -4.46, -2.65, -0.64, 3.3, 7.24]\n",
      "2 [-16.31, -15.22, -15.2, 0.0, -15.48, -15.49, 0.0, 0.0, -12.11, -8.87, -5.38, -5.63, 0.0, 3.08]\n",
      "3 [-16.35, -14.93, -15.08, 0.0, -15.76, -15.01, 0.0, -14.02, -13.37, -6.48, -8.25, 0.0, 0.0, -6.92]\n",
      "4 [-16.01, -14.86, -14.88, 0.0, -14.38, -14.2, -13.38, -12.28, -11.21, -9.85, -8.3, 0.0, 0.0, -6.28]\n",
      "5 [-15.97, -16.05, -15.94, 0.0, -16.23, -15.98, -15.54, -14.87, -12.42, -12.07, -9.26, 0.0, 0.0, -5.84]\n",
      "Learned policy:\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 1, state: (4, 0, 4, 1), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'A' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 2, state: (4, 1, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'A' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 3, state: (5, 1, 4, 2), actions: 2, reward: -14\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 4, state: (5, 2, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 5, state: (4, 2, 3, 2), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 6, state: (3, 2, 2, 2), actions: 1, reward: -15\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 7, state: (2, 2, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 8, state: (2, 1, 1, 2), actions: 3, reward: -17\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 9, state: (1, 1, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 10, state: (1, 2, 1, 3), actions: 4, reward: -15\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 11, state: (1, 3, 1, 4), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 12, state: (1, 4, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 13, state: (0, 4, 1, 5), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 14, state: (0, 5, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 15, state: (1, 5, 2, 5), actions: 2, reward: -12\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 16, state: (2, 5, 3, 5), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 17, state: (3, 5, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 18, state: (3, 4, 4, 5), actions: 3, reward: -11\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 19, state: (4, 4, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 20, state: (4, 5, 4, 6), actions: 4, reward: -9\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 21, state: (4, 6, 4, 7), actions: 4, reward: -8\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 22, state: (4, 7, 4, 8), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 23, state: (4, 8, 4, 9), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 24, state: (4, 9, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 25, state: (5, 9, 4, 10), actions: 2, reward: -6\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_']]\n",
      "step: 26, state: (5, 10, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_']]\n",
      "step: 27, state: (4, 10, 3, 10), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'B' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 28, state: (3, 10, 2, 10), actions: 1, reward: -7\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 29, state: (2, 10, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 30, state: (2, 9, 1, 10), actions: 3, reward: -9\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'A' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 31, state: (1, 9, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'A' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 32, state: (1, 10, 1, 11), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'B' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 33, state: (1, 11, 1, 12), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'A' b'B' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 34, state: (1, 12, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 35, state: (0, 12, 1, 13), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 36, state: (0, 13, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 37, state: (1, 13, 2, 13), actions: 2, reward: -4\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 38, state: (2, 13, 3, 13), actions: 2, reward: -3\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 39, state: (3, 13, 4, 13), actions: 2, reward: 998\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "rewards: 623\n",
      "action history: [1, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2, 3, 2, 4, 4, 4, 4, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2]\n"
     ]
    }
   ],
   "source": [
    "# Visualization of V-Table and Policy from the direct implementation of custom pQ-Learning model with p=3 r=0.1\n",
    "v_table3, policy3 = visualize(agent3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Learning with Custom pQ-Learning where p=4\n",
    "\n",
    "Customized Q Learning Bellman Equation formulation:\n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-3},A_{t-3})-r_{1}Q(S_{t-2},A_{t-2}) -r_{2}Q(S_{t-1},A_{t-1})-r_{3}Q(S_{t},A_{t} )]$\n",
    "where the last $r: r_{p-1}=r_{3}=1.0$ for all given p constant\n",
    "\n",
    "Setting p=4 does not result in no significant improvement, the learning outcome wasn't harmed as well. The model converge at **34000** episodes and completed 75000 episodes in **350-400** seconds. The agent reached saturation at around **34000** episodes in about **176.9** seconds \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using reward engineering:  False\n",
      "Total time taken:  360.4449496269226\n"
     ]
    }
   ],
   "source": [
    "# Guang Way's Custom Q-Learning\n",
    "import time\n",
    "\n",
    "start = time.time()\n",
    "episodes = 75000\n",
    "p=4\n",
    "r = 0.1\n",
    "env = CliffBoxGridWorld(rewardengineering=False)\n",
    "env.reset()\n",
    "agent4 = RLAgent(env, num_episodes=episodes, epsilon=0.01, alpha=0.01, gamma=0.99, p=p, r = r)\n",
    "rewards_custom4 = agent4.learn(learning_algorithm = \"custom q-learning\")\n",
    "time_taken = time.time() - start\n",
    "print(\"Total time taken: \", time_taken)\n",
    "# print(f\"Speed: {episodes / (time.time() - start)} episodes/sec\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Learning Curve Visualization when setting p=4, \n",
    "\n",
    "At $p=4$ can be noted the result doesnt differ alot from $p=3$\n",
    "\n",
    "Customized Q Learning Bellman Equation formulation:\n",
    "$Q(S_{t},A_{t})_{new}=Q(S_{t},A_{t})+\\alpha [R+\\gamma max_{a}Q(S_{t+1},a)-r_{0}Q(S_{t-3},A_{t-3})-r_{1}Q(S_{t-2},A_{t-2}) -r_{2}Q(S_{t-1},A_{t-1})-r_{3}Q(S_{t},A_{t} )]$\n",
    "\n",
    "It appears that the historical record from Time Step $T-3$ has no direct implication on Time Step $T$ for this specific problem. However neither does increasing $p$ constant beyond optimal does any harm to the convergence rate. The agent reached saturation at around **34000** episodes in about **176.9** seconds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "description = 'Learning Progress pQ-Learning (Episode rewards (Y axis) vs Episodes (X axis)) p=4 r=0.1'\n",
    "plot_smooth_curve(weight=0.99, rewards=rewards_custom4, description=description)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Learned V-table and Policy (custom pQ-learning with p=4 r=0.1)\n",
    "The optimal converged trajectory is not the best possible as well(in terms of cumulative reward). It is as follows:  \n",
    "\n",
    "```[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```\n",
    "\n",
    "Final cumulative reward: **623**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V table:\n",
      "0 [-16.28, -16.29, -16.63, -16.44, -16.59, -15.96, 0.0, 0.0, -12.44, -8.83, -8.21, -2.7, -1.93, 6.01]\n",
      "1 [-16.05, -14.76, -15.13, -15.55, -15.56, -15.27, 0.0, 0.0, -9.29, -3.91, -2.34, -0.68, 2.79, 7.37]\n",
      "2 [-16.16, -15.08, -15.08, 0.0, -15.31, -15.26, 0.0, 0.0, -11.07, -8.65, -5.27, -5.32, 0.0, 3.16]\n",
      "3 [-16.23, -14.81, -14.99, 0.0, -15.59, -14.97, 0.0, -13.96, -13.01, -6.45, -7.82, 0.0, 0.0, -6.83]\n",
      "4 [-15.88, -14.73, -14.78, 0.0, -14.22, -14.05, -13.27, -12.33, -11.19, -9.8, -8.09, 0.0, 0.0, -6.19]\n",
      "5 [-15.83, -15.92, -15.79, 0.0, -15.97, -15.79, -15.46, -14.62, -12.2, -11.92, -9.14, 0.0, 0.0, -5.73]\n",
      "Learned policy:\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 1, state: (4, 0, 4, 1), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'A' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 2, state: (4, 1, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'A' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 3, state: (5, 1, 4, 2), actions: 2, reward: -14\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 4, state: (5, 2, 4, 2), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 5, state: (4, 2, 3, 2), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 6, state: (3, 2, 2, 2), actions: 1, reward: -15\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 7, state: (2, 2, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 8, state: (2, 1, 1, 2), actions: 3, reward: -17\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 9, state: (1, 1, 1, 2), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 10, state: (1, 2, 1, 3), actions: 4, reward: -15\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 11, state: (1, 3, 1, 4), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 12, state: (1, 4, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 13, state: (0, 4, 1, 5), actions: 1, reward: -14\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 14, state: (0, 5, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 15, state: (1, 5, 2, 5), actions: 2, reward: -12\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 16, state: (2, 5, 3, 5), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 17, state: (3, 5, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 18, state: (3, 4, 4, 5), actions: 3, reward: -11\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 19, state: (4, 4, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 20, state: (4, 5, 4, 6), actions: 4, reward: -9\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 21, state: (4, 6, 4, 7), actions: 4, reward: -8\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 22, state: (4, 7, 4, 8), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 23, state: (4, 8, 4, 9), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 24, state: (4, 9, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 25, state: (5, 9, 4, 10), actions: 2, reward: -6\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_']]\n",
      "step: 26, state: (5, 10, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_']]\n",
      "step: 27, state: (4, 10, 3, 10), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'B' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 28, state: (3, 10, 2, 10), actions: 1, reward: -7\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 29, state: (2, 10, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 30, state: (2, 9, 1, 10), actions: 3, reward: -9\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'A' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 31, state: (1, 9, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'A' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 32, state: (1, 10, 1, 11), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'B' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 33, state: (1, 11, 1, 12), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'A' b'B' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 34, state: (1, 12, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 35, state: (0, 12, 1, 13), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 36, state: (0, 13, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 37, state: (1, 13, 2, 13), actions: 2, reward: -4\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 38, state: (2, 13, 3, 13), actions: 2, reward: -3\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 39, state: (3, 13, 4, 13), actions: 2, reward: 998\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "rewards: 623\n",
      "action history: [1, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2, 3, 2, 4, 4, 4, 4, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2]\n"
     ]
    }
   ],
   "source": [
    "# Visualization of V-Table and Policy from the direct implementation of custom pQ-Learning with p=4, r=0.1\n",
    "v_table4, policy4 = visualize(agent4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Development of Improved Agent that learns using the Custom pQ-Learning with reformulated Reward scheme\n",
    "As per discussed prior to the above implementation/visualization, the Custom pQ-Learning has the trajectory of converged episodes as follows:\n",
    "\n",
    "```\n",
    "[(5, 0, 4, 1), (4, 0, 4, 1), (4, 1, 4, 2), (5, 1, 4, 2), (5, 2, 4, 2), (4, 2, 3, 2), (3, 2, 2, 2), (2, 2, 1, 2), (2, 1, 1, 2), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]\n",
    "```\n",
    "\n",
    "Above trajectory yields Cumulative Reward: **623**  \n",
    "This is unfortunately not the optimal shortest trajectory possible like the one Traditional Q-Learning achieved.  \n",
    "To resolve the issue, for the last part of my RL implementation, I conducted reward engineering on the current reward scheme such that it reward agent based on smart or stupid actions rather than just based on the direct distance of box/agents and the timesteps it takes to reach the goal. \n",
    "\n",
    "For example:  \n",
    "1. Movement that does not involves pushing the box are penalized with negative rewards, so the RL agent will not be able to achieve decent cumulative reward if it spends too much time pondering around the box without any intention to move it when its near the goal post.  \n",
    "\n",
    "2. Movement that involves moving away from the box further when Agent is already next to the box is also penalized heavily as the goal is actually as the following 2:  \n",
    "    1. To move Agent position to the Box  \n",
    "    2. To move the Box until it reach the Goal position  \n",
    "\n",
    "    Therefore any steps that involves moving away from the box would be consider breach of the first goal and definitely not the optimal trajectory anymore and hence ends the game. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using Custom p=2 Q-Learning Incorporated with Reward Engineering to further speed up learning\n",
    "Model converges and achieved optimal trajectory at about episode **21000** at the best trajectory possible. (Similar to Traditional Q-Learning)  \n",
    "```[(5, 0, 4, 1), (5, 1, 4, 1), (4, 1, 3, 1), (3, 1, 2, 1), (2, 1, 1, 1), (2, 0, 1, 1), (1, 0, 1, 1), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```  \n",
    "\n",
    "The above optimal trajectory yields cumulative reward of **577**. The main trick to ensure that the new modified Custom pQ-Learning achieve best possible optimal trajectory is to set a greater penalty reward for movement that does not involves pushing the box, so that the Model reduces focus on the Box/Agent's distance to the goal. Reason being is that, in the early phase of the game, such attributes are less important than being able to get the box to start moving first.\n",
    "\n",
    "The model completed running **75,000** episodes in in about **300** seconds. Since it takes about **20000** episodes to converges, convergence is achieved at approx **80** seconds. About **2 times faster than the original Custom pQ-Learning and 3 times faster than the original Traditional Q-Learning**, when both are without Reward engineering. (approx **180** and **240** seconds respectively)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using reward engineering:  True\n",
      "Total time taken:  347.53932452201843\n"
     ]
    }
   ],
   "source": [
    "# Guang Way's Custom Q-Learning with Reward Engineering\n",
    "import time\n",
    "\n",
    "start = time.time()\n",
    "episodes = 75000\n",
    "p=2\n",
    "r = 0.1\n",
    "env = CliffBoxGridWorld(rewardengineering=True) # Set this to True to activate my custom implemented reward engineering\n",
    "env.reset()\n",
    "agent5 = RLAgent(env, num_episodes=episodes, epsilon=0.01, alpha=0.01, gamma=0.99, p=p, r=r)\n",
    "rewards_custom_re = agent5.learn(learning_algorithm = \"custom q-learning\")\n",
    "print(\"Total time taken: \", time.time() - start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Learning Curve: Setting p=2, \n",
    "##### The model converges at 21000 episodes in approx 92.4 seconds. \n",
    "This is the best model we achieved by the custom innovation. (In terms of training speed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "description = 'Learning Progress pQ-Learning w/ Reward Engineering (Episode rewards (Y) vs Episodes (X) p=2 r=0.1'\n",
    "plot_smooth_curve(weight=0.99, rewards=rewards_custom_re, description=description)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Learned V-table and Policy (custom pQ-learning with p=2 r=0.1 w/ Reward engineering)\n",
    "\n",
    "Now the model converged at the best possible trajectory like the original q-learning did:\n",
    "\n",
    "```[(5, 0, 4, 1), (5, 1, 4, 1), (4, 1, 3, 1), (3, 1, 2, 1), (2, 1, 1, 1), (2, 0, 1, 1), (1, 0, 1, 1), (1, 1, 1, 2), (1, 2, 1, 3), (1, 3, 1, 4), (1, 4, 1, 5), (0, 4, 1, 5), (0, 5, 1, 5), (1, 5, 2, 5), (2, 5, 3, 5), (3, 5, 4, 5), (3, 4, 4, 5), (4, 4, 4, 5), (4, 5, 4, 6), (4, 6, 4, 7), (4, 7, 4, 8), (4, 8, 4, 9), (4, 9, 4, 10), (5, 9, 4, 10), (5, 10, 4, 10), (4, 10, 3, 10), (3, 10, 2, 10), (2, 10, 1, 10), (2, 9, 1, 10), (1, 9, 1, 10), (1, 10, 1, 11), (1, 11, 1, 12), (1, 12, 1, 13), (0, 12, 1, 13), (0, 13, 1, 13), (1, 13, 2, 13), (2, 13, 3, 13)]```\n",
    "\n",
    "Final cumulative reward **642**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V table:\n",
      "0 [-15.09, -15.39, -15.77, -15.69, -14.98, -14.27, 0.0, 0.0, -9.05, -5.21, -3.86, -2.67, 2.14, 3.82]\n",
      "1 [-14.94, -14.02, -14.49, -14.41, -13.69, -13.55, 0.0, 0.0, -8.9, -4.1, -2.77, -0.95, 8.21, 5.1]\n",
      "2 [-15.11, -13.99, -14.61, 0.0, -13.13, -13.96, 0.0, 0.0, -10.46, -5.7, -1.21, -0.16, 0.0, 6.86]\n",
      "3 [-15.21, -14.49, -14.83, 0.0, -13.48, -12.62, 0.0, -10.47, -9.87, -3.34, -6.18, 0.0, 0.0, -3.45]\n",
      "4 [-15.48, -14.3, -14.5, 0.0, -12.52, -12.61, -11.53, -10.38, -9.5, -6.02, -6.64, 0.0, 0.0, -3.04]\n",
      "5 [-15.51, -15.58, -15.74, 0.0, -13.28, -13.31, -12.61, -11.58, -10.66, -8.56, -7.45, 0.0, 0.0, -2.76]\n",
      "Learned policy:\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 1, state: (5, 1, 4, 1), actions: 4, reward: -19\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 2, state: (4, 1, 3, 1), actions: 1, reward: -15\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 3, state: (3, 1, 2, 1), actions: 1, reward: -16\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 4, state: (2, 1, 1, 1), actions: 1, reward: -17\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 5, state: (2, 0, 1, 1), actions: 3, reward: -23\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'A' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 6, state: (1, 0, 1, 1), actions: 1, reward: -22\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 7, state: (1, 1, 1, 2), actions: 4, reward: -16\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 8, state: (1, 2, 1, 3), actions: 4, reward: -15\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 9, state: (1, 3, 1, 4), actions: 4, reward: -14\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 10, state: (1, 4, 1, 5), actions: 4, reward: -13\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 11, state: (0, 4, 1, 5), actions: 1, reward: -19\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 12, state: (0, 5, 1, 5), actions: 4, reward: -18\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 13, state: (1, 5, 2, 5), actions: 2, reward: -12\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 14, state: (2, 5, 3, 5), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 15, state: (3, 5, 4, 5), actions: 2, reward: -10\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 16, state: (3, 4, 4, 5), actions: 3, reward: -16\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 17, state: (4, 4, 4, 5), actions: 2, reward: -15\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'A' b'B' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 18, state: (4, 5, 4, 6), actions: 4, reward: -9\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'A' b'B' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 19, state: (4, 6, 4, 7), actions: 4, reward: -8\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'A' b'B' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 20, state: (4, 7, 4, 8), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'B' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 21, state: (4, 8, 4, 9), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'A' b'B' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 22, state: (4, 9, 4, 10), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 23, state: (5, 9, 4, 10), actions: 2, reward: -11\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'A' b'_' b'x' b'x' b'_']]\n",
      "step: 24, state: (5, 10, 4, 10), actions: 4, reward: -10\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'B' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'_']]\n",
      "step: 25, state: (4, 10, 3, 10), actions: 1, reward: -6\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'B' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'A' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 26, state: (3, 10, 2, 10), actions: 1, reward: -7\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'A' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 27, state: (2, 10, 1, 10), actions: 1, reward: -8\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 28, state: (2, 9, 1, 10), actions: 3, reward: -14\n",
      "Action: 3\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'A' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 29, state: (1, 9, 1, 10), actions: 1, reward: -13\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'A' b'B' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 30, state: (1, 10, 1, 11), actions: 4, reward: -7\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'A' b'B' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 31, state: (1, 11, 1, 12), actions: 4, reward: -6\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'A' b'B' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 32, state: (1, 12, 1, 13), actions: 4, reward: -5\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 33, state: (0, 12, 1, 13), actions: 1, reward: -11\n",
      "Action: 1\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'A' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 34, state: (0, 13, 1, 13), actions: 4, reward: -10\n",
      "Action: 4\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 35, state: (1, 13, 2, 13), actions: 2, reward: -4\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 36, state: (2, 13, 3, 13), actions: 2, reward: -3\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'G']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "step: 37, state: (3, 13, 4, 13), actions: 2, reward: 998\n",
      "Action: 2\n",
      "[[b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'_' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'x' b'_' b'_' b'_' b'_' b'x' b'_']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'x' b'x' b'A']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'B']\n",
      " [b'_' b'_' b'_' b'x' b'_' b'_' b'_' b'_' b'_' b'_' b'_' b'x' b'x' b'_']]\n",
      "rewards: 577\n",
      "action history: [4, 1, 1, 1, 3, 1, 4, 4, 4, 4, 1, 4, 2, 2, 2, 3, 2, 4, 4, 4, 4, 4, 2, 4, 1, 1, 1, 3, 1, 4, 4, 4, 1, 4, 2, 2, 2]\n"
     ]
    }
   ],
   "source": [
    "# Visualization of V-Table and Policy from the direct implementation of custom pQ-Learning with p=2, r=0.1\n",
    "v_table5, policy5 = visualize(agent5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion\n",
    "\n",
    "The assignment ends here after we have trained the above pQ-Learning RL agent with reward engineering. Through this assignment I learned to implement the Q-Learning RL algorithm taught in class. I also discovered that some modification can be done on the Bellman's Equation to achieve greater convergence rate by remembering historical data to train and learn the optimal action for later steps.  \n",
    "\n",
    "It is proven that for this assignment's application, the custom technique learns faster. To prove so, in this assignment the learnrate $\\alpha $ was reduced from optimally 0.1 to 0.01 to slow down the Learning Algorithms' learning curve so that we can benchmark performance against each other. the pQ-Learning model of all $p$ and $r$ settings significannt out performances. (Converging with about 2 to 3 times lesser episodes.) This comes with a cost of just about 0-25% longer duration to cover 75,000 episodes.   \n",
    "\n",
    "\n",
    "# Possible future upgrade on Algorithm\n",
    "Similar implementation might also be possible with other RL algorithm that was taught in AI6101 such as SARSA which is relatively similar to Q-Learning. \n",
    "\n",
    "Some of the formulation in this implementation such as weight decay vector $P$, and how $H$ was initialized at the start of every episode may be improved too. \n",
    "\n",
    "However due to tight deadlines together with many other assignments in the Masters Program, I did not attempt further investigation. Possible furture work may consider these aspects.   \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[NbConvertApp] Converting notebook .\\AI6101_Code_TengGuangWay_G2102434F.ipynb to pdf\n",
      "[NbConvertApp] Support files will be in AI6101_Code_TengGuangWay_G2102434F_files\\\n",
      "[NbConvertApp] Making directory .\\AI6101_Code_TengGuangWay_G2102434F_files\n",
      "[NbConvertApp] Making directory .\\AI6101_Code_TengGuangWay_G2102434F_files\n",
      "[NbConvertApp] Making directory .\\AI6101_Code_TengGuangWay_G2102434F_files\n",
      "[NbConvertApp] Making directory .\\AI6101_Code_TengGuangWay_G2102434F_files\n",
      "[NbConvertApp] Making directory .\\AI6101_Code_TengGuangWay_G2102434F_files\n",
      "[NbConvertApp] Writing 234122 bytes to notebook.tex\n",
      "[NbConvertApp] Building PDF\n",
      "[NbConvertApp] Running xelatex 3 times: ['xelatex', 'notebook.tex', '-quiet']\n",
      "[NbConvertApp] Running bibtex 1 time: ['bibtex', 'notebook']\n",
      "[NbConvertApp] WARNING | b had problems, most likely because there were no citations\n",
      "[NbConvertApp] PDF successfully created\n",
      "[NbConvertApp] Writing 375750 bytes to AI6101_Code_TengGuangWay_G2102434F.pdf\n"
     ]
    }
   ],
   "source": [
    "!jupyter nbconvert --to pdf .\\AI6101_Code_TengGuangWay_G2102434F.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyP9ohMoS+Jycd5V49qgyi7u",
   "provenance": [
    {
     "file_id": "1IDOYe2na06C1FbbHae02fQ1NwoG3NUDg",
     "timestamp": 1676470026430
    },
    {
     "file_id": "1bRDhABxWGe-LIeWgzQl8GGkBmf9iKyuz",
     "timestamp": 1676444626866
    }
   ]
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.9"
  },
  "vscode": {
   "interpreter": {
    "hash": "2372fb78870ca5ddfbb77edbd31465c7991d171e5767f3094f475408094f996f"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}