agent_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "# Agent Analysis\n",
    "\n",
    "**Create exploratory plots & statistics of the agent (aka model, aka AI).**\n",
    "\n",
    "Specifically, we measure rewards of games (aka episodes) played by 2 models:\n",
    "- The greedy model always chooses the best action (ε=0)\n",
    "- The exploration model takes a random action with a 10% chance (ε=0.1).\n",
    "\n",
    "We measure the median reward of multiple models to limit the influence of randomness."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Overview\n",
    "\n",
    "* Setup\n",
    "* Train & Test the models\n",
    "* Reward during training progress\n",
    "* Reward of fully trained models\n",
    "* Time Metrics\n"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Setup"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "outputs": [],
   "source": [
    "import time\n",
    "import warnings\n",
    "from random import Random\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from statsmodels.nonparametric.smoothers_lowess import lowess\n",
    "\n",
    "from classes.action import Action\n",
    "from classes.agent import Agent\n",
    "from classes.game import Game\n",
    "from classes.state import State\n",
    "from classes.utils import get_track"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.165370Z",
     "end_time": "2023-12-13T21:13:10.193899Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "outputs": [],
   "source": [
    "sns.set_theme()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.195927Z",
     "end_time": "2023-12-13T21:13:10.209120Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Train & test the models"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Define some functions."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "outputs": [],
   "source": [
    "def get_reward_of_episode(episode: list[tuple[State, Action, int]]) -> int:\n",
    "    total_reward = 0\n",
    "    for (state, action, reward) in episode:\n",
    "        total_reward += reward\n",
    "    return total_reward"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.212120Z",
     "end_time": "2023-12-13T21:13:10.228142Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "outputs": [],
   "source": [
    "def simulate_train(track: np.ndarray, epsilon: float, n_models: int = 10, n_episodes: int = 1000) -> tuple[pd.DataFrame, np.ndarray, list[Agent]]:\n",
    "    \"\"\"\n",
    "    Trains multiple agents, and collect statistics of their games\n",
    "    \"\"\"\n",
    "    \n",
    "    # generate a list of random states, so that each model starts with different & reproducible seeds\n",
    "    rnd = Random(42)\n",
    "    random_states = [rnd.randint(0, 100000000) for _ in range(n_models)]\n",
    "\n",
    "    train_times = np.zeros(n_models)\n",
    "    stats_per_game_and_model: list[tuple[int, int, int, int]] = [] # dimensions: n_model, n_game, n_steps, reward\n",
    "    agents: list[Agent] = []\n",
    "\n",
    "    for i in range(n_models):\n",
    "        agent = Agent(random_state=random_states[i])\n",
    "        game = Game(racetrack=track, random_state=random_states[i])\n",
    "        start = time.time()\n",
    "        for j in range(0, n_episodes):\n",
    "            episode: list[tuple[State, Action, int]] = []\n",
    "            while not game.is_finished() and game.get_n_steps() < 1000:\n",
    "                state = game.get_state()\n",
    "                action = agent.determine_epsilon_action(state, epsilon)\n",
    "                reward = game.noisy_step(action)\n",
    "                episode.append((state, action, reward))\n",
    "            agent.learn(episode)\n",
    "            game.reset()\n",
    "\n",
    "            total_reward = get_reward_of_episode(episode)\n",
    "            stats_per_game_and_model.append((i,j,game.get_n_steps(),total_reward))\n",
    "        end = time.time()\n",
    "        train_times[i] = end - start\n",
    "        agents.append(agent)\n",
    "\n",
    "    df = pd.DataFrame(stats_per_game_and_model, columns=[\"n_model\",\"n_games\",\"n_steps\",\"reward\"])\n",
    "    df[\"epsilon\"] = epsilon\n",
    "    return df, train_times, agents"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.232143Z",
     "end_time": "2023-12-13T21:13:10.240653Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "outputs": [],
   "source": [
    "def simulate_test(track: np.ndarray, agents: list[Agent], epsilon: float, n_episodes: int = 100) -> tuple[pd.DataFrame, np.ndarray]:\n",
    "    \"\"\"\n",
    "    Tests multiple trained agents, and collect statistics of their games\n",
    "    \"\"\"\n",
    "\n",
    "    # generate a list of random states\n",
    "    rnd = Random(43)\n",
    "    random_states = [rnd.randint(0, 100000000) for _ in range(len(agents))]\n",
    "\n",
    "    test_times = np.zeros(shape=(len(agents),n_episodes))\n",
    "    stats_per_game_and_model: list[tuple[int, int, int, int]] = [] # dimensions:  n_game, n_steps, reward\n",
    "\n",
    "    for i, agent in enumerate(agents):\n",
    "        game = Game(racetrack=track, random_state=random_states[i])\n",
    "        for j in range(0, n_episodes):\n",
    "            start = time.time()\n",
    "            episode: list[tuple[State, Action, int]] = []\n",
    "            while not game.is_finished() and game.get_n_steps() < 1000:\n",
    "                state = game.get_state()\n",
    "                action = agent.determine_best_action(state)\n",
    "                reward = game.step(action)\n",
    "                episode.append((state, action, reward))\n",
    "            game.reset()\n",
    "            total_reward = get_reward_of_episode(episode)\n",
    "            stats_per_game_and_model.append((i,j,game.get_n_steps(),total_reward))\n",
    "            end = time.time()\n",
    "            test_times[i][j] = end - start\n",
    "\n",
    "    df = pd.DataFrame(stats_per_game_and_model, columns=[\"n_model\",\"n_games\",\"n_steps\",\"reward\"])\n",
    "    df[\"epsilon\"] = epsilon\n",
    "    return df, test_times"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.245162Z",
     "end_time": "2023-12-13T21:13:10.260680Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "outputs": [],
   "source": [
    "def lowess_grouped(data: pd.DataFrame, x_name: str | tuple[str,...], y_name: str | tuple[str,...], group_name: str | tuple[str,...] | None, frac: float) -> pd.DataFrame:\n",
    "    \"\"\"\n",
    "    Apply groupwise lowess smoothing to a dataframe\n",
    "    \"\"\"\n",
    "\n",
    "    df = data.copy()\n",
    "\n",
    "    if type(y_name) is str:\n",
    "        y_name_smoothed = y_name + \"_smooth\"\n",
    "    elif type(y_name) is tuple:\n",
    "        y_name_smoothed = (y_name[0]+\"_smooth\",) + (y_name[1:])\n",
    "    else:\n",
    "        raise ValueError(\"Type of y_name not supported\")\n",
    "\n",
    "    if group_name is not None:\n",
    "        groups = df[group_name].unique().tolist()\n",
    "        smoothed_dfs = []\n",
    "        for group in groups:\n",
    "            df_by_select_group = df[df[group_name] == group]\n",
    "            smoothed_df = lowess(df_by_select_group[y_name],df_by_select_group[x_name], frac=frac)\n",
    "            smoothed_df = pd.DataFrame(smoothed_df, columns=[x_name, y_name_smoothed])\n",
    "            smoothed_df[group_name] = group\n",
    "            smoothed_dfs.append(smoothed_df)\n",
    "        return pd.merge(df,pd.concat(smoothed_dfs), how=\"left\", on=[x_name,group_name])\n",
    "    else:\n",
    "        smoothed_df = lowess(df[y_name], df[x_name], frac=frac)\n",
    "        smoothed_df = pd.DataFrame(smoothed_df, columns=[x_name, y_name_smoothed])\n",
    "        return pd.merge(df,smoothed_df, how=\"left\", on=x_name)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.261680Z",
     "end_time": "2023-12-13T21:13:10.273189Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Select the racetrack."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [],
   "source": [
    "track = get_track(track_number=None, track_random_seed=81)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2023-12-13T21:13:10.274698Z",
     "end_time": "2023-12-13T21:13:10.290730Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Train and test models using epsilon of 0% and 10%."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "outputs": [],
   "source": [
    "stats_train_e0, train_times_e0, agents_e0 = simulate_train(track, epsilon=0, n_models=3, n_episodes=50000)\n",
    "stats_train_e10, train_times_e10, agents_e10 = simulate_train(track, epsilon=0.1, n_models=3, n_episodes=50000)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:45:37.691444500Z",
     "start_time": "2023-10-11T09:08:08.888885Z"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "outputs": [],
   "source": [
    "stats_test_e0, test_times_e0 = simulate_test(track, agents_e0, epsilon=0)\n",
    "stats_test_e10, test_times_e10 = simulate_test(track, agents_e10, epsilon=0.1)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:45:42.739922500Z",
     "start_time": "2023-10-11T09:45:37.731437200Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Aggregate and smooth train and test data"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "outputs": [],
   "source": [
    "# ignore warnings from function lowess_grouped()\n",
    "with warnings.catch_warnings():\n",
    "    warnings.simplefilter('ignore')\n",
    "\n",
    "    stats_test = pd.concat([stats_test_e0,stats_test_e10])\n",
    "    stats_test = stats_test.groupby([\"epsilon\",\"n_games\"]).aggregate({'reward':['median','min','max']}).reset_index()\n",
    "    stats_test = lowess_grouped(stats_test, y_name=(\"reward\",\"median\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)\n",
    "    stats_test = lowess_grouped(stats_test, y_name=(\"reward\",\"min\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)\n",
    "    stats_test = lowess_grouped(stats_test, y_name=(\"reward\",\"max\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)\n",
    "\n",
    "    stats_train = pd.concat([stats_train_e0,stats_train_e10])\n",
    "    stats_train = stats_train.groupby([\"epsilon\",\"n_games\"]).aggregate({'reward':['median','min','max']}).reset_index()\n",
    "    stats_train = lowess_grouped(stats_train, y_name=(\"reward\",\"median\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)\n",
    "    stats_train = lowess_grouped(stats_train, y_name=(\"reward\",\"min\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)\n",
    "    stats_train = lowess_grouped(stats_train, y_name=(\"reward\",\"max\"), x_name=(\"n_games\",\"\"), group_name=\"epsilon\", frac=0.07)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:50:56.164374Z",
     "start_time": "2023-10-11T09:45:42.763921600Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Reward during training progress"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Let's look how the reward (aka performance) of the models change, as we train them for more episodes"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.lineplot(stats_train,x=(\"n_games\",\"\"),y=(\"reward_smooth\",\"median\"),hue=\"epsilon\")\n",
    "g.set(title='Reward per Episode during Training', xlabel='Number of Episodes', ylabel='Median Reward');"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:02.964252400Z",
     "start_time": "2023-10-11T09:50:56.175917100Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Reward of fully trained models"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Let's look how the reward of the trained models compares."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(pd.concat([stats_test_e0,stats_test_e10]), x=\"epsilon\", y=\"reward\", showfliers=False);"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:03.336790400Z",
     "start_time": "2023-10-11T09:52:02.939765Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Show outliners in boxplot"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(pd.concat([stats_test_e0,stats_test_e10]), x=\"epsilon\", y=\"reward\", showfliers=True);"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:03.340790400Z",
     "start_time": "2023-10-11T09:52:03.128445400Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "In some cases violinplots can help understanding the distribution better."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.violinplot(pd.concat([stats_test_e0,stats_test_e10]), x=\"epsilon\", y=\"reward\");"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:03.536842400Z",
     "start_time": "2023-10-11T09:52:03.300300200Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Time Metrics"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Training Times"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_times = pd.DataFrame({\"e0\": train_times_e0, \"e10\": train_times_e10})\n",
    "sns.boxplot(train_times).set(ylabel=\"train time (s)\");"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:03.728953800Z",
     "start_time": "2023-10-11T09:52:03.536842400Z"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Testing Times"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_times = pd.DataFrame({\"e0\": test_times_e0.flatten(), \"e10\": test_times_e10.flatten()})\n",
    "sns.boxplot(test_times);"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2023-10-11T09:52:03.929038900Z",
     "start_time": "2023-10-11T09:52:03.724950600Z"
    }
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}