Grid_search_for_Reinforcement_learning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Grid search for Reinforcement learning.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "p18p7nKqFCAt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.model_selection import ParameterGrid"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QOWpOf0rGAAG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def run_algorithm (env, params, episodes = 100, aerl = 10):\n",
        "  \"\"\"\n",
        "  for testing reasons only returns a random number as for the average reward over the last x episodes (aerl = average episode reward length)\n",
        "  Every RL-algorithm needs to have the same input values and should return the average over the last (aerl) episodes!\n",
        "  \"\"\"\n",
        "  return np.random.randint(0,100)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WU_oRbvVF_-O",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class GridsearchRL():\n",
        "  def __init__(self, function, params, episodes= 100, aerl = 10):\n",
        "    self.function = function\n",
        "    self.episodes = episodes\n",
        "    self.aerl = aerl\n",
        "    self.params = params\n",
        "    self.grid = ParameterGrid(params)\n",
        "    self.grid_len = len(self.grid)\n",
        "    self.history = dict()\n",
        "    \n",
        "  def run_gridsearch(self, env):\n",
        "    \"\"\"\n",
        "    Run the grid search on the given environment.\n",
        "    \"\"\"\n",
        "    for idx, parameters in enumerate(self.grid):\n",
        "      result = self.function(env, parameters, self.episodes, self.aerl)\n",
        "      self.history.update({str(parameters): result})\n",
        "      print(\"\\rSearch progress [{}/{}]\".format(idx+1, self.grid_len), end = \"\", flush =True)\n",
        "  \n",
        "  \n",
        "  def results(self, figsize = (18,7)):\n",
        "    plt.figure(figsize= figsize)\n",
        "    plt.plot(*zip(*self.history.items()))\n",
        "    plt.xticks(rotation = 90)\n",
        "    plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f-XxXJzSF_75",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "search1 = GridsearchRL(run_algorithm, params,)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pocLyeb6MS1_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 33
        },
        "outputId": "37279323-65b4-4af0-ad8e-33e749666365"
      },
      "source": [
        "search1.run_gridsearch(\"a\")"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Search progress [9/9]"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VzAysJJbMSwh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 167
        },
        "outputId": "c154cf12-8a02-4000-96ed-df1f11caca59"
      },
      "source": [
        "search1.history\n"
      ],
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{\"{'a': 1, 'b': 1}\": 58,\n",
              " \"{'a': 1, 'b': 2}\": 68,\n",
              " \"{'a': 1, 'b': 3}\": 80,\n",
              " \"{'a': 2, 'b': 1}\": 16,\n",
              " \"{'a': 2, 'b': 2}\": 2,\n",
              " \"{'a': 2, 'b': 3}\": 49,\n",
              " \"{'a': 3, 'b': 1}\": 0,\n",
              " \"{'a': 3, 'b': 2}\": 17,\n",
              " \"{'a': 3, 'b': 3}\": 89}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xGbnr8bqF_3D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 340
        },
        "outputId": "9961e221-44d9-4cc0-fb65-1c5bebab22c2"
      },
      "source": [
        "search1.results()"
      ],
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAFDCAYAAAA3TgmYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rhBDcgvIPjAB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}