notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7fPc-KWUi2Xd"
      },
      "outputs": [],
      "source": [
        "# Copyright 2021 Google LLC\n",
        "#\n",
        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eoXf8TfQoVth"
      },
      "source": [
        "# Optimizing multiple objectives with Vertex AI Vizier\n",
        "\n",
        "<table align=\"left\">\n",
        "\n",
        "  <td>\n",
        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb\">\n",
        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
        "    </a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb\">\n",
        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
        "      View on GitHub\n",
        "    </a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb\">\n",
        "        <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
        "      Open in Vertex AI Workbench\n",
        "    </a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b397c59391b1"
      },
      "source": [
        "## Overview\n",
        "\n",
        "This tutorial demonstrates [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier/overview) multi-objective optimization. Multi-objective optimization is concerned with mathematical optimization problems involving more than one objective function to be optimized simultaneously.\n",
        "\n",
        "Learn more about [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier/overview)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AksIKBzZ-nre"
      },
      "source": [
        "### Objective\n",
        "\n",
        "In this tutorial, you learn how to use `Vertex AI Vizier` to optimize a multi-objective study.\n",
        "\n",
        "The goal is to __`minimize`__ the objective metric:\n",
        "   ```\n",
        "   y1 = r*sin(theta)\n",
        "   ```\n",
        "\n",
        "and simultaneously __`maximize`__ the objective metric:\n",
        "   ```\n",
        "   y2 = r*cos(theta)\n",
        "   ```\n",
        "\n",
        "that you will evaluate over the parameter space:\n",
        "\n",
        "   - __`r`__ in [0,1],\n",
        "\n",
        "   - __`theta`__ in [0, pi/2]\n",
        "\n",
        "### Costs\n",
        "\n",
        "This tutorial uses billable components of Google Cloud:\n",
        "\n",
        "* Vertex AI\n",
        "\n",
        "Learn about [Vertex AI\n",
        "pricing](https://cloud.google.com/vertex-ai/pricing) and use the [Pricing\n",
        "Calculator](https://cloud.google.com/products/calculator/)\n",
        "to generate a cost estimate based on your projected usage.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iMHz63rPbq6P"
      },
      "source": [
        "## Installation\n",
        "\n",
        "Install the packages required for executing this notebook."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "b6f3dc43494b"
      },
      "outputs": [],
      "source": [
        "! pip3 install --quiet google-vizier==0.0.4\n",
        "! pip3 install --upgrade --quiet google-cloud-aiplatform"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "58707a750154"
      },
      "source": [
        "### Colab only: Uncomment the following cell to restart the kernel."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "f200f10a1da3"
      },
      "outputs": [],
      "source": [
        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
        "# import IPython\n",
        "\n",
        "# app = IPython.Application.instance()\n",
        "# app.kernel.do_shutdown(True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WReHDGG5g0XY"
      },
      "source": [
        "#### Set your project ID\n",
        "\n",
        "**If you don't know your project ID**, try the following:\n",
        "* Run `gcloud config list`.\n",
        "* Run `gcloud projects list`.\n",
        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oM1iC_MfAts1"
      },
      "outputs": [],
      "source": [
        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
        "\n",
        "# Set the project id\n",
        "! gcloud config set project {PROJECT_ID}"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "region"
      },
      "source": [
        "#### Set the region\n",
        "\n",
        "**Optional**: Update the 'REGION' variable to specify the region that you want to use. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nsN5NJKSu-GU"
      },
      "outputs": [],
      "source": [
        "REGION = \"us-central1\"  # @param {type: \"string\"}"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sBCra4QMA2wR"
      },
      "source": [
        "### Authenticate your Google Cloud account\n",
        "\n",
        "To authenticate your Google Cloud account, follow the instructions for your Jupyter environment:\n",
        "\n",
        "**1. Vertex AI Workbench**\n",
        "<br>You are already authenticated.\n",
        "\n",
        "**2. Local JupyterLab instance**\n",
        "<br>Uncomment and run the following code:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "254614fa0c46"
      },
      "outputs": [],
      "source": [
        "# ! gcloud auth login"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ef21552ccea8"
      },
      "source": [
        "**3. Colab**\n",
        "<br>Uncomment and run the following code:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "603adbbf0532"
      },
      "outputs": [],
      "source": [
        "# from google.colab import auth\n",
        "\n",
        "# auth.authenticate_user()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c13224697bfb"
      },
      "source": [
        "**4. Service account or other**\n",
        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Dax2zrpTi2Xy"
      },
      "source": [
        "### Import libraries and define constants"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xD60d6Q0i2X0"
      },
      "outputs": [],
      "source": [
        "import datetime\n",
        "\n",
        "from google.cloud import aiplatform\n",
        "from google.cloud.aiplatform.vizier import Study, pyvizier"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CWuu4wmki2X3"
      },
      "source": [
        "## Tutorial\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KyEjqIdnad0w"
      },
      "source": [
        "This section defines some parameters and util methods to call Vertex Vizier APIs. Please fill in the following information to get started."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8HCgeF8had77"
      },
      "outputs": [],
      "source": [
        "# These will be automatically filled in.\n",
        "STUDY_DISPLAY_NAME = \"{}_study_{}\".format(\n",
        "    PROJECT_ID.replace(\"-\", \"\"), datetime.datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
        ")\n",
        "PARENT = \"projects/{}/locations/{}\".format(PROJECT_ID, REGION)\n",
        "\n",
        "print(\"REGION: {}\".format(REGION))\n",
        "print(\"PARENT: {}\".format(PARENT))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8NBduXsEaRKr"
      },
      "source": [
        "### Create the study configuration\n",
        "\n",
        "The following is a sample study configuration, built as a hierarchical python dictionary. It is already filled out. Run the cell to configure the study."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "s-AHfPOASXXW"
      },
      "outputs": [],
      "source": [
        "# Parameter Configuration\n",
        "problem = pyvizier.StudyConfig()\n",
        "problem.algorithm = pyvizier.Algorithm.RANDOM_SEARCH\n",
        "\n",
        "# Objective Metrics\n",
        "problem.metric_information.append(\n",
        "    pyvizier.MetricInformation(name=\"y1\", goal=pyvizier.ObjectiveMetricGoal.MINIMIZE)\n",
        ")\n",
        "problem.metric_information.append(\n",
        "    pyvizier.MetricInformation(name=\"y2\", goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE)\n",
        ")\n",
        "\n",
        "# Defines the parameters configuration.\n",
        "root = problem.search_space.select_root()\n",
        "root.add_float_param(\"r\", 0, 1.0, scale_type=pyvizier.ScaleType.LINEAR)\n",
        "root.add_float_param(\"theta\", 0, 1.57, scale_type=pyvizier.ScaleType.LINEAR)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uyXG_RKha7Kb"
      },
      "source": [
        "### Create the study\n",
        "\n",
        "Next, create the study, which you will subsequently run to optimize the two objectives."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jgskzqZX0Mkt"
      },
      "outputs": [],
      "source": [
        "aiplatform.init(project=PROJECT_ID, location=REGION)\n",
        "study = Study.create_or_load(display_name=STUDY_DISPLAY_NAME, problem=problem)\n",
        "\n",
        "STUDY_ID = study.name\n",
        "print(\"STUDY_ID: {}\".format(STUDY_ID))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dKOMBKmtkcJb"
      },
      "source": [
        "### Metric evaluation functions\n",
        "\n",
        "Next, define some functions to evaluate the two objective metrics."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Xnl1uqnyz3Qp"
      },
      "outputs": [],
      "source": [
        "import math\n",
        "\n",
        "\n",
        "# r * sin(theta)\n",
        "def Metric1Evaluation(r, theta):\n",
        "    \"\"\"Evaluate the first metric on the trial.\"\"\"\n",
        "    return r * math.sin(theta)\n",
        "\n",
        "\n",
        "# r * cos(theta)\n",
        "def Metric2Evaluation(r, theta):\n",
        "    \"\"\"Evaluate the second metric on the trial.\"\"\"\n",
        "    return r * math.cos(theta)\n",
        "\n",
        "\n",
        "def CreateMetrics(trial_id, r, theta):\n",
        "    print((\"=========== Start Trial: [{}] =============\").format(trial_id))\n",
        "\n",
        "    # Evaluate both objective metrics for this trial\n",
        "    y1 = Metric1Evaluation(r, theta)\n",
        "    y2 = Metric2Evaluation(r, theta)\n",
        "    print(\n",
        "        \"[r = {}, theta = {}] => y1 = r*sin(theta) = {}, y2 = r*cos(theta) = {}\".format(\n",
        "            r, theta, y1, y2\n",
        "        )\n",
        "    )\n",
        "    measurement = pyvizier.Measurement()\n",
        "    measurement.metrics[\"y1\"] = y1\n",
        "    measurement.metrics[\"y2\"] = y2\n",
        "\n",
        "    # Return the results for this trial\n",
        "    return measurement"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qzn5lVpRq05U"
      },
      "source": [
        "### Set configuration parameters for running trials\n",
        "\n",
        "__`client_id`__: The identifier of the client that is requesting the suggestion. If multiple SuggestTrialsRequests have the same `client_id`, the service will return the identical suggested trial if the trial is `PENDING`, and provide a new trial if the last suggested trial was completed.\n",
        "\n",
        "__`suggestion_count_per_request`__: The number of suggestions (trials) requested in a single request.\n",
        "\n",
        "__`max_trial_id_to_stop`__: The number of trials to explore before stopping. It is set to 4 to shorten the time to run the code, so don't expect convergence. For convergence, it would likely need to be about 20 (a good rule of thumb is to multiply the total dimensionality by 10).\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5usXaZA5qvUZ"
      },
      "outputs": [],
      "source": [
        "worker_id = \"worker1\"  # @param {type: 'string'}\n",
        "suggestion_count_per_request = 3  # @param {type: 'integer'}\n",
        "max_trial_id_to_stop = 6  # @param {type: 'integer'}\n",
        "\n",
        "print(\"worker_id: {}\".format(worker_id))\n",
        "print(\"suggestion_count_per_request: {}\".format(suggestion_count_per_request))\n",
        "print(\"max_trial_id_to_stop: {}\".format(max_trial_id_to_stop))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UnV2SJNskm7V"
      },
      "source": [
        "### Run Vertex Vizier trials\n",
        "\n",
        "Run the trials."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "opmuTntW4-eS"
      },
      "outputs": [],
      "source": [
        "while len(study.trials()) < max_trial_id_to_stop:\n",
        "    trials = study.suggest(count=suggestion_count_per_request, worker=worker_id)\n",
        "\n",
        "    for suggested_trial in trials:\n",
        "        measurement = CreateMetrics(\n",
        "            suggested_trial.name,\n",
        "            suggested_trial.parameters[\"r\"].value,\n",
        "            suggested_trial.parameters[\"theta\"].value,\n",
        "        )\n",
        "        suggested_trial.add_measurement(measurement=measurement)\n",
        "        suggested_trial.complete(measurement=measurement)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i5ZTqgqBiRsq"
      },
      "source": [
        "### List the optimal solutions\n",
        "\n",
        "list_optimal_trials returns the pareto-optimal Trials for multi-objective Study or the optimal Trials for single-objective Study. In the case, we define mutliple-objective in previeous steps, pareto-optimal trials will be returned."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Or2PL1YxTr33"
      },
      "outputs": [],
      "source": [
        "optimal_trials = study.optimal_trials()\n",
        "print(\"optimal_trials: {}\".format(optimal_trials))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KAxfq9Fri2YV"
      },
      "source": [
        "## Cleaning up\n",
        "\n",
        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial. You can also manually delete resources that you created by running the following code."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zQlLDfvlzYde"
      },
      "outputs": [],
      "source": [
        "study.delete()"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "gapic-vizier-multi-objective-optimization.ipynb",
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}