optimization/DFBA-PRCP/DFBA-PRCP.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "aa78b91d-8dc7-4b5a-8f73-50d2ee80d1a6",
   "metadata": {},
   "source": [
    "# Optimization Algorithm Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c0c1402-3107-4ce0-9b51-50506ef35d0f",
   "metadata": {},
   "source": [
    " * Evaluate the performance and cost optimization algorithm on a serverless application with 6 functions and 4 types of structures"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "98e6a622-ab61-400e-a0d6-ff9c7dcabda8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import logging\n",
    "from io import BytesIO\n",
    "import time\n",
    "import zipfile\n",
    "import numpy as np\n",
    "import boto3\n",
    "from datetime import datetime, timezone\n",
    "from time import gmtime, strftime\n",
    "import json\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import math\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4976d576-017b-489f-b646-1a631072e25c",
   "metadata": {},
   "source": [
    "## Performance Profile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5d07e731-1ae4-48d1-b4b9-7ae4729dd913",
   "metadata": {},
   "outputs": [],
   "source": [
    "App6_lambda_logs = pd.read_csv('App6_Lambda_Logs.csv', low_memory=False)\n",
    "App6_lambda_logs.columns = ['Function', 'Memory_Size', 'Max_Memory_Used', 'Duration', 'Billed_Duration', 'RequestId']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e52ef9ea-cbc6-468c-b5c4-adb6b25bd85a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Function</th>\n",
       "      <th>Memory_Size</th>\n",
       "      <th>Max_Memory_Used</th>\n",
       "      <th>Duration</th>\n",
       "      <th>Billed_Duration</th>\n",
       "      <th>RequestId</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>f1</td>\n",
       "      <td>3008</td>\n",
       "      <td>62</td>\n",
       "      <td>812.05</td>\n",
       "      <td>900</td>\n",
       "      <td>b6a380ea-ee12-457e-af63-6e3919bb1f2a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>f1</td>\n",
       "      <td>3008</td>\n",
       "      <td>62</td>\n",
       "      <td>713.71</td>\n",
       "      <td>800</td>\n",
       "      <td>ff20d6e4-388f-4842-9d1e-c9aae94ca3fd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>f1</td>\n",
       "      <td>3008</td>\n",
       "      <td>62</td>\n",
       "      <td>720.99</td>\n",
       "      <td>800</td>\n",
       "      <td>a6110dab-c8e7-491e-8075-6d46bb2ad026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>f1</td>\n",
       "      <td>3008</td>\n",
       "      <td>62</td>\n",
       "      <td>711.73</td>\n",
       "      <td>800</td>\n",
       "      <td>56b4cb0a-4d1c-4912-8072-b3e0a634ec3a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>f1</td>\n",
       "      <td>3008</td>\n",
       "      <td>62</td>\n",
       "      <td>798.65</td>\n",
       "      <td>800</td>\n",
       "      <td>175ce139-3afe-45a0-8101-db5870227043</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Function  Memory_Size  Max_Memory_Used  Duration  Billed_Duration  \\\n",
       "0       f1         3008               62    812.05              900   \n",
       "1       f1         3008               62    713.71              800   \n",
       "2       f1         3008               62    720.99              800   \n",
       "3       f1         3008               62    711.73              800   \n",
       "4       f1         3008               62    798.65              800   \n",
       "\n",
       "                              RequestId  \n",
       "0  b6a380ea-ee12-457e-af63-6e3919bb1f2a  \n",
       "1  ff20d6e4-388f-4842-9d1e-c9aae94ca3fd  \n",
       "2  a6110dab-c8e7-491e-8075-6d46bb2ad026  \n",
       "3  56b4cb0a-4d1c-4912-8072-b3e0a634ec3a  \n",
       "4  175ce139-3afe-45a0-8101-db5870227043  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "App6_lambda_logs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a76fab57-c80b-41de-8e5d-29ed033c5f21",
   "metadata": {},
   "outputs": [],
   "source": [
    "viable_mem_list = list(range(128, 3072, 192))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e1f5d1e4-04dd-462c-9608-55342c77d455",
   "metadata": {},
   "outputs": [],
   "source": [
    "performance_profile = {}\n",
    "for i in range(6):\n",
    "    dic = {}\n",
    "    for mem in viable_mem_list:\n",
    "        dic[mem] = np.array(\n",
    "            App6_lambda_logs.query(f\"Function == 'f{i + 1}' and Memory_Size == {mem} \")['Duration'].tolist()[10:190])\n",
    "    performance_profile[f\"f{i + 1}\"] = dic"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37312c2f-c4ac-4661-8336-7e8b1be67c17",
   "metadata": {},
   "source": [
    "## Application Definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ec544030-13c2-49de-b669-93a59edb0774",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('../../')\n",
    "from slappsim.Structures import *\n",
    "from slappsim.Function import Function\n",
    "from slappsim.PetriApp import *\n",
    "from slappsim.States import *\n",
    "from slapptuner.AppTuner import AppTuner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4006433c-0b96-42cf-9916-58e41ffb1412",
   "metadata": {},
   "outputs": [],
   "source": [
    "from functools import partial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6a0ffb72-9223-4093-9c17-6bb03642fd2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "rs = np.random.RandomState(64)\n",
    "random.seed(64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fdfb48dd-073f-4866-8d55-b46d1de5ba03",
   "metadata": {},
   "outputs": [],
   "source": [
    "def delay19():\n",
    "    return 19.81\n",
    "function_execution_delay_fun = partial(delay19)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7b3170dc-2eb1-4ef9-afa3-bf56e28743e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "delays = {'FunctionExecution': function_execution_delay_fun}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "08fb78ea-7f3f-4179-9902-27cfc7bb021b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def dummy_fun():\n",
    "    return 0\n",
    "dummy_pf_fun = partial(dummy_fun)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef196b82-0781-460d-82f2-f89e6db0ebcf",
   "metadata": {},
   "source": [
    "* The equation for calculating the cost of functions has been changed to the old one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b93087a7-802f-4ca6-9dc4-ceedb048b405",
   "metadata": {},
   "outputs": [],
   "source": [
    "start = Start()\n",
    "end = End()\n",
    "f1 = Function(pf_fun=dummy_pf_fun, mem=0, name='f1')\n",
    "f2 = Function(pf_fun=dummy_pf_fun, mem=0, name='f2')\n",
    "f3 = Function(pf_fun=dummy_pf_fun, mem=0, name='f3')\n",
    "f4 = Function(pf_fun=dummy_pf_fun, mem=0, name='f4')\n",
    "f5 = Function(pf_fun=dummy_pf_fun, mem=0, name='f5')\n",
    "f6 = Function(pf_fun=dummy_pf_fun, mem=0, name='f6')\n",
    "choice1 = Choice(choices=[f3, f5], probabilities=[0.8, 0.2])\n",
    "parallel_sequence1 = Sequence(actions=[choice1])\n",
    "parallel_sequence2 = Sequence(actions=[f2])\n",
    "parallel1 = Parallel(branches=[parallel_sequence1, parallel_sequence2])\n",
    "rchoice1 = Choice(choices=[f1, f4, f6], probabilities=[0.1, 0.2, 0.7], end=False)\n",
    "sequence1 = Sequence(actions=[f1, parallel1, f4, rchoice1, f6])\n",
    "structures = [parallel1, choice1, rchoice1, sequence1, parallel_sequence1, parallel_sequence2]\n",
    "i1 = InArc(place=start)\n",
    "o1 = OutArc(place=sequence1.structure_start)\n",
    "t1 = Transition(in_arcs=[i1], out_arcs=[o1])\n",
    "i2 = InArc(place=sequence1.structure_end)\n",
    "o2 = OutArc(place=end)\n",
    "t2 = Transition(in_arcs=[i2], out_arcs=[o2])\n",
    "transitions = [t1, t2]\n",
    "transitions += sequence1.transitions\n",
    "app = PetriApp(transitions=transitions, functions=[f1, f2, f3, f4, f5, f6], structures=structures, delays=delays)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "65502d86-e8c8-4173-801d-bcb1ae37b37f",
   "metadata": {},
   "outputs": [],
   "source": [
    "tuner_app = AppTuner(petri_model=app, performance_profile=performance_profile)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f6f6082-d828-469f-a178-4a5804dd958d",
   "metadata": {},
   "source": [
    "## Configure tasks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7a18c201-c180-4e0a-9853-754cffa8ccb9",
   "metadata": {},
   "outputs": [],
   "source": [
    "App6_BPBC_PRCP = pd.read_csv(\"App6_BPBC_PRCP.csv\")\n",
    "App6_BCPC_PRCP = pd.read_csv(\"App6_BCPC_PRCP.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "cd299064-5671-48a2-b370-e50a87d9bfeb",
   "metadata": {},
   "outputs": [],
   "source": [
    "budget_constraint = App6_BPBC_PRCP['Budget'].to_list()\n",
    "performance_constraint = App6_BCPC_PRCP['Performance_Constraint'].to_list()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d60b121-d557-4636-9e3f-9e864e2c7ee2",
   "metadata": {},
   "source": [
    "## BPBC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea465c7c-adb5-48d2-a7f6-5244319b50a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "os.mkdir(\"BPBC\")\n",
    "for constraint in budget_constraint:\n",
    "    result = {}\n",
    "    result['problem'] = 'BPBC'\n",
    "    result['constraint'] = constraint\n",
    "    result['percentile'] = None \n",
    "    configuration = tuner_app.optimize_bpbc(budget_constraint = constraint, percentile = None)\n",
    "    result['optimized_configuration'] = configuration\n",
    "    result['duration'] = tuner_app.ert_list.tolist()\n",
    "    result['cost'] = tuner_app.cost_list.tolist()\n",
    "    with open(f\"./BPBC/{round(constraint, 2)}_avg.json\", 'w') as f:\n",
    "        json.dump(result, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23b08c89-7ebd-46f6-87ea-e423cf6707ea",
   "metadata": {},
   "source": [
    "## BCPC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06e7981c-c34c-4583-9290-5250e6a090f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "os.mkdir(\"BCPC\")\n",
    "for constraint in performance_constraint:\n",
    "    result = {}\n",
    "    result['problem'] = 'BCPC'\n",
    "    result['constraint'] = constraint\n",
    "    result['percentile'] = None\n",
    "    configuration = tuner_app.optimize_bcpc(performance_constraint = constraint, percentile = None)\n",
    "    result['optimized_configuration'] = configuration\n",
    "    result['duration'] = tuner_app.ert_list.tolist()\n",
    "    result['cost'] = tuner_app.cost_list.tolist()\n",
    "    with open(f\"./BCPC/{round(constraint, 2)}_avg.json\", 'w') as f:\n",
    "        json.dump(result, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e78d167f-06ff-42c0-8912-c1bbbbb910eb",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b7b6a9a5-296d-4bd8-b004-8789b539e152",
   "metadata": {},
   "source": [
    "### DFBA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "50ab355a-ecf4-4123-bada-c5c35e504db3",
   "metadata": {},
   "outputs": [],
   "source": [
    "optimization_results = pd.DataFrame(columns=[\"problem\", \"constraint\", \"percentile\", \"optimized_configuration\", \"duration_avg\", \"duration_median\", \"duration_percentile10\", \"duration_percentile90\", \"cost_avg\", \"cost_median\", \"cost_percentile10\", \"cost_percentile90\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "061b4631-1913-4667-86bf-68c84d7108f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "for file in os.listdir(\"BPBC\"):\n",
    "    with open(f\"BPBC/{file}\" , \"r\") as f:\n",
    "        result_dict = {}\n",
    "        result = json.load(f)\n",
    "        result[\"duration_avg\"] = np.mean(result['duration'])\n",
    "        result[\"duration_median\"] = np.median(result['duration'])\n",
    "        result[\"duration_percentile10\"] = np.percentile(result['duration'], 10)\n",
    "        result[\"duration_percentile90\"] = np.percentile(result['duration'], 90)\n",
    "        result[\"cost_avg\"] = np.mean(result['cost']) * 1000000\n",
    "        result[\"cost_median\"] = np.median(result['cost']) * 1000000\n",
    "        result[\"cost_percentile10\"] = np.percentile(result['cost'], 10) * 1000000\n",
    "        result[\"cost_percentile90\"] = np.percentile(result['cost'], 90) * 1000000\n",
    "        result.pop(\"duration\")\n",
    "        result.pop(\"cost\")\n",
    "        optimization_results = optimization_results.append(result, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "67560f4d-be48-4ca0-85fa-035a6cf5ec8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "for file in os.listdir(\"BCPC\"):\n",
    "    with open(f\"BCPC/{file}\" , \"r\") as f:\n",
    "        result_dict = {}\n",
    "        result = json.load(f)\n",
    "        result[\"duration_avg\"] = np.mean(result['duration'])\n",
    "        result[\"duration_median\"] = np.median(result['duration'])\n",
    "        result[\"duration_percentile10\"] = np.percentile(result['duration'], 10)\n",
    "        result[\"duration_percentile90\"] = np.percentile(result['duration'], 90)\n",
    "        result[\"cost_avg\"] = np.mean(result['cost']) * 1000000\n",
    "        result[\"cost_median\"] = np.median(result['cost']) * 1000000\n",
    "        result[\"cost_percentile10\"] = np.percentile(result['cost'], 10) * 1000000\n",
    "        result[\"cost_percentile90\"] = np.percentile(result['cost'], 90) * 1000000\n",
    "        result.pop(\"duration\")\n",
    "        result.pop(\"cost\")\n",
    "        optimization_results = optimization_results.append(result, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "332c9594-c6f1-4351-b93a-378256b3fa94",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>problem</th>\n",
       "      <th>constraint</th>\n",
       "      <th>percentile</th>\n",
       "      <th>optimized_configuration</th>\n",
       "      <th>duration_avg</th>\n",
       "      <th>duration_median</th>\n",
       "      <th>duration_percentile10</th>\n",
       "      <th>duration_percentile90</th>\n",
       "      <th>cost_avg</th>\n",
       "      <th>cost_median</th>\n",
       "      <th>cost_percentile10</th>\n",
       "      <th>cost_percentile90</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BPBC</td>\n",
       "      <td>145.767517</td>\n",
       "      <td>None</td>\n",
       "      <td>{'f1': 1472, 'f2': 2240, 'f3': 2048, 'f4': 300...</td>\n",
       "      <td>2666.133869</td>\n",
       "      <td>2255.990</td>\n",
       "      <td>2042.770</td>\n",
       "      <td>3943.720</td>\n",
       "      <td>116.307133</td>\n",
       "      <td>95.1668</td>\n",
       "      <td>89.1252</td>\n",
       "      <td>172.1128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BPBC</td>\n",
       "      <td>76.887081</td>\n",
       "      <td>None</td>\n",
       "      <td>{'f1': 896, 'f2': 1280, 'f3': 1664, 'f4': 1472...</td>\n",
       "      <td>3200.129564</td>\n",
       "      <td>2656.680</td>\n",
       "      <td>2552.858</td>\n",
       "      <td>4638.300</td>\n",
       "      <td>76.653540</td>\n",
       "      <td>62.6668</td>\n",
       "      <td>60.7918</td>\n",
       "      <td>110.9836</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BPBC</td>\n",
       "      <td>91.237171</td>\n",
       "      <td>None</td>\n",
       "      <td>{'f1': 1280, 'f2': 1280, 'f3': 1664, 'f4': 224...</td>\n",
       "      <td>2830.286739</td>\n",
       "      <td>2386.890</td>\n",
       "      <td>2249.861</td>\n",
       "      <td>4169.074</td>\n",
       "      <td>91.036377</td>\n",
       "      <td>75.1669</td>\n",
       "      <td>71.5210</td>\n",
       "      <td>133.9003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BPBC</td>\n",
       "      <td>149.594208</td>\n",
       "      <td>None</td>\n",
       "      <td>{'f1': 1472, 'f2': 2240, 'f3': 2048, 'f4': 300...</td>\n",
       "      <td>2690.636497</td>\n",
       "      <td>2256.045</td>\n",
       "      <td>2042.399</td>\n",
       "      <td>3955.384</td>\n",
       "      <td>117.570038</td>\n",
       "      <td>95.2710</td>\n",
       "      <td>89.1252</td>\n",
       "      <td>173.2586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BPBC</td>\n",
       "      <td>74.017062</td>\n",
       "      <td>None</td>\n",
       "      <td>{'f1': 896, 'f2': 1280, 'f3': 1664, 'f4': 1280...</td>\n",
       "      <td>3116.321378</td>\n",
       "      <td>2581.080</td>\n",
       "      <td>2489.470</td>\n",
       "      <td>4595.533</td>\n",
       "      <td>73.654462</td>\n",
       "      <td>61.7293</td>\n",
       "      <td>60.2710</td>\n",
       "      <td>106.6961</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  problem  constraint percentile  \\\n",
       "0    BPBC  145.767517       None   \n",
       "1    BPBC   76.887081       None   \n",
       "2    BPBC   91.237171       None   \n",
       "3    BPBC  149.594208       None   \n",
       "4    BPBC   74.017062       None   \n",
       "\n",
       "                             optimized_configuration  duration_avg  \\\n",
       "0  {'f1': 1472, 'f2': 2240, 'f3': 2048, 'f4': 300...   2666.133869   \n",
       "1  {'f1': 896, 'f2': 1280, 'f3': 1664, 'f4': 1472...   3200.129564   \n",
       "2  {'f1': 1280, 'f2': 1280, 'f3': 1664, 'f4': 224...   2830.286739   \n",
       "3  {'f1': 1472, 'f2': 2240, 'f3': 2048, 'f4': 300...   2690.636497   \n",
       "4  {'f1': 896, 'f2': 1280, 'f3': 1664, 'f4': 1280...   3116.321378   \n",
       "\n",
       "   duration_median  duration_percentile10  duration_percentile90    cost_avg  \\\n",
       "0         2255.990               2042.770               3943.720  116.307133   \n",
       "1         2656.680               2552.858               4638.300   76.653540   \n",
       "2         2386.890               2249.861               4169.074   91.036377   \n",
       "3         2256.045               2042.399               3955.384  117.570038   \n",
       "4         2581.080               2489.470               4595.533   73.654462   \n",
       "\n",
       "   cost_median  cost_percentile10  cost_percentile90  \n",
       "0      95.1668            89.1252           172.1128  \n",
       "1      62.6668            60.7918           110.9836  \n",
       "2      75.1669            71.5210           133.9003  \n",
       "3      95.2710            89.1252           173.2586  \n",
       "4      61.7293            60.2710           106.6961  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "optimization_results.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec13bd9b-2246-405a-b6b2-fe8f28a5e08d",
   "metadata": {
    "tags": []
   },
   "source": [
    "### PRCP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2433070d-be4d-4fd6-ae8d-7f63a8b2da0e",
   "metadata": {},
   "outputs": [],
   "source": [
    "perf_cost_data = pd.DataFrame()\n",
    "for file in os.listdir('Exhaustive_Search'):\n",
    "    if '.csv' in file:\n",
    "        data_parts = pd.read_csv(f\"Exhaustive_Search/{file}\")\n",
    "        perf_cost_data = perf_cost_data.append(data_parts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "50efe560-3ab9-49d4-892d-bb1c6a49d401",
   "metadata": {},
   "outputs": [],
   "source": [
    "App6_BPBC_PRCP = App6_BPBC_PRCP.assign(Best_Answer_RT=lambda App6_BPBC_PRCP: App6_BPBC_PRCP[['BCR_disabled_RT', 'BCR_RT/M_RT', 'BCR_ERT/C_RT', 'BCR_MAX_RT']].min(1))\n",
    "best_rt = [perf_cost_data.query(\"Cost<={}\".format(item))['RT'].min() for item in App6_BPBC_PRCP['Budget']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "5ed0274f-012f-4fb9-9446-016fa2c0c9fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Budget</th>\n",
       "      <th>BCR_disabled_RT</th>\n",
       "      <th>BCR_RT/M_RT</th>\n",
       "      <th>BCR_ERT/C_RT</th>\n",
       "      <th>BCR_MAX_RT</th>\n",
       "      <th>BCR_disabled_Cost</th>\n",
       "      <th>BCR_RT/M_Cost</th>\n",
       "      <th>BCR_ERT/C_Cost</th>\n",
       "      <th>BCR_MAX_Cost</th>\n",
       "      <th>BCR_disabled_Config</th>\n",
       "      <th>BCR_RT/M_Config</th>\n",
       "      <th>BCR_ERT/C_Config</th>\n",
       "      <th>BCR_MAX_Config</th>\n",
       "      <th>BCR_disabled_Iterations</th>\n",
       "      <th>BCR_RT/M_Iterations</th>\n",
       "      <th>BCR_ERT/C_Iterations</th>\n",
       "      <th>BCR_MAX_Iterations</th>\n",
       "      <th>BCR_threshold</th>\n",
       "      <th>Best_Answer_RT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>59.666971</td>\n",
       "      <td>10868.5009</td>\n",
       "      <td>10868.5009</td>\n",
       "      <td>25433.083814</td>\n",
       "      <td>10868.500900</td>\n",
       "      <td>59.599543</td>\n",
       "      <td>59.599543</td>\n",
       "      <td>59.666971</td>\n",
       "      <td>59.599543</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 128, 2: 128, 3: 128, 4: 128, 5: 128, 6: 128}</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>10868.500900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>60.623644</td>\n",
       "      <td>10772.8808</td>\n",
       "      <td>10772.8808</td>\n",
       "      <td>13629.086743</td>\n",
       "      <td>11727.572671</td>\n",
       "      <td>60.328543</td>\n",
       "      <td>60.328543</td>\n",
       "      <td>60.447400</td>\n",
       "      <td>59.980800</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 1856}</td>\n",
       "      <td>{1: 896, 2: 320, 3: 320, 4: 128, 5: 512, 6: 1472}</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>10772.880800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>61.580317</td>\n",
       "      <td>10772.8808</td>\n",
       "      <td>10772.8808</td>\n",
       "      <td>13629.086743</td>\n",
       "      <td>11631.952571</td>\n",
       "      <td>60.328543</td>\n",
       "      <td>60.328543</td>\n",
       "      <td>60.447400</td>\n",
       "      <td>60.709800</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 1856}</td>\n",
       "      <td>{1: 896, 2: 320, 3: 320, 4: 128, 5: 512, 6: 1856}</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>10772.880800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>62.536990</td>\n",
       "      <td>10757.7975</td>\n",
       "      <td>10772.8808</td>\n",
       "      <td>13614.003443</td>\n",
       "      <td>6811.492243</td>\n",
       "      <td>61.893543</td>\n",
       "      <td>60.328543</td>\n",
       "      <td>62.012400</td>\n",
       "      <td>62.397943</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 2...</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...</td>\n",
       "      <td>{1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 2048}</td>\n",
       "      <td>{1: 896, 2: 320, 3: 320, 4: 320, 5: 512, 6: 1472}</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>6811.492243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>63.493662</td>\n",
       "      <td>10754.2255</td>\n",
       "      <td>10738.5096</td>\n",
       "      <td>13614.003443</td>\n",
       "      <td>6715.872143</td>\n",
       "      <td>63.458543</td>\n",
       "      <td>63.181114</td>\n",
       "      <td>62.012400</td>\n",
       "      <td>63.126943</td>\n",
       "      <td>{1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 2...</td>\n",
       "      <td>{1: 1088, 2: 896, 3: 1088, 4: 128, 5: 512, 6: ...</td>\n",
       "      <td>{1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 2048}</td>\n",
       "      <td>{1: 896, 2: 320, 3: 320, 4: 320, 5: 512, 6: 1856}</td>\n",
       "      <td>9.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>6715.872143</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Budget  BCR_disabled_RT  BCR_RT/M_RT  BCR_ERT/C_RT    BCR_MAX_RT  \\\n",
       "0  59.666971       10868.5009   10868.5009  25433.083814  10868.500900   \n",
       "1  60.623644       10772.8808   10772.8808  13629.086743  11727.572671   \n",
       "2  61.580317       10772.8808   10772.8808  13629.086743  11631.952571   \n",
       "3  62.536990       10757.7975   10772.8808  13614.003443   6811.492243   \n",
       "4  63.493662       10754.2255   10738.5096  13614.003443   6715.872143   \n",
       "\n",
       "   BCR_disabled_Cost  BCR_RT/M_Cost  BCR_ERT/C_Cost  BCR_MAX_Cost  \\\n",
       "0          59.599543      59.599543       59.666971     59.599543   \n",
       "1          60.328543      60.328543       60.447400     59.980800   \n",
       "2          60.328543      60.328543       60.447400     60.709800   \n",
       "3          61.893543      60.328543       62.012400     62.397943   \n",
       "4          63.458543      63.181114       62.012400     63.126943   \n",
       "\n",
       "                                 BCR_disabled_Config  \\\n",
       "0  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "1  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "2  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "3  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 2...   \n",
       "4  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 2...   \n",
       "\n",
       "                                     BCR_RT/M_Config  \\\n",
       "0  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "1  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "2  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "3  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...   \n",
       "4  {1: 1088, 2: 896, 3: 1088, 4: 128, 5: 512, 6: ...   \n",
       "\n",
       "                                    BCR_ERT/C_Config  \\\n",
       "0   {1: 128, 2: 128, 3: 128, 4: 128, 5: 128, 6: 128}   \n",
       "1  {1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 1856}   \n",
       "2  {1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 1856}   \n",
       "3  {1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 2048}   \n",
       "4  {1: 896, 2: 128, 3: 128, 4: 128, 5: 128, 6: 2048}   \n",
       "\n",
       "                                      BCR_MAX_Config  BCR_disabled_Iterations  \\\n",
       "0  {1: 896, 2: 896, 3: 1088, 4: 128, 5: 512, 6: 1...                      9.0   \n",
       "1  {1: 896, 2: 320, 3: 320, 4: 128, 5: 512, 6: 1472}                      8.0   \n",
       "2  {1: 896, 2: 320, 3: 320, 4: 128, 5: 512, 6: 1856}                      8.0   \n",
       "3  {1: 896, 2: 320, 3: 320, 4: 320, 5: 512, 6: 1472}                      8.0   \n",
       "4  {1: 896, 2: 320, 3: 320, 4: 320, 5: 512, 6: 1856}                      9.0   \n",
       "\n",
       "   BCR_RT/M_Iterations  BCR_ERT/C_Iterations  BCR_MAX_Iterations  \\\n",
       "0                  9.0                   3.0                 9.0   \n",
       "1                  8.0                   5.0                 8.0   \n",
       "2                  8.0                   5.0                 9.0   \n",
       "3                  8.0                   5.0                 9.0   \n",
       "4                  8.0                   5.0                10.0   \n",
       "\n",
       "   BCR_threshold  Best_Answer_RT  \n",
       "0            0.2    10868.500900  \n",
       "1            0.2    10772.880800  \n",
       "2            0.2    10772.880800  \n",
       "3            0.2     6811.492243  \n",
       "4            0.2     6715.872143  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "App6_BPBC_PRCP.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "044fea63-0e19-4bf8-a487-7bb139dde2ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "App6_BCPC_PRCP = App6_BCPC_PRCP.assign(Best_Answer_Cost=lambda App6_BCPC_PRCP: App6_BCPC_PRCP[['BCR_disabled_Cost', 'BCR_M/RT_Cost', 'BCR_C/ERT_Cost', 'BCR_MAX_Cost']].min(1))\n",
    "best_cost = [perf_cost_data.query(\"RT<={}\".format(item))['Cost'].min() for item in App6_BCPC_PRCP['Performance_Constraint']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "63370d79-bbaa-4ff9-9980-eca1e51be41c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Performance_Constraint</th>\n",
       "      <th>BCR_disabled_RT</th>\n",
       "      <th>BCR_M/RT_RT</th>\n",
       "      <th>BCR_C/ERT_RT</th>\n",
       "      <th>BCR_MAX_RT</th>\n",
       "      <th>BCR_disabled_Cost</th>\n",
       "      <th>BCR_M/RT_Cost</th>\n",
       "      <th>BCR_C/ERT_Cost</th>\n",
       "      <th>BCR_MAX_Cost</th>\n",
       "      <th>BCR_disabled_Config</th>\n",
       "      <th>BCR_M/RT_Config</th>\n",
       "      <th>BCR_C/ERT_Config</th>\n",
       "      <th>BCR_MAX_Config</th>\n",
       "      <th>BCR_disabled_Iterations</th>\n",
       "      <th>BCR_M/RT_Iterations</th>\n",
       "      <th>BCR_C/ERT_Iterations</th>\n",
       "      <th>BCR_MAX_Iterations</th>\n",
       "      <th>BCR_threshold</th>\n",
       "      <th>Best_Answer_Cost</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2853.318043</td>\n",
       "      <td>2851.234229</td>\n",
       "      <td>2851.234229</td>\n",
       "      <td>2851.234229</td>\n",
       "      <td>2852.520811</td>\n",
       "      <td>87.416771</td>\n",
       "      <td>87.416771</td>\n",
       "      <td>87.416771</td>\n",
       "      <td>91.295400</td>\n",
       "      <td>{1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...</td>\n",
       "      <td>{1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...</td>\n",
       "      <td>{1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...</td>\n",
       "      <td>{1: 1088, 2: 2048, 3: 1664, 4: 2240, 5: 512, 6...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>87.416771</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3081.396485</td>\n",
       "      <td>3070.231189</td>\n",
       "      <td>3070.231189</td>\n",
       "      <td>3070.231189</td>\n",
       "      <td>3061.305594</td>\n",
       "      <td>87.150714</td>\n",
       "      <td>87.150714</td>\n",
       "      <td>87.150714</td>\n",
       "      <td>78.376829</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...</td>\n",
       "      <td>{1: 896, 2: 1472, 3: 1664, 4: 1664, 5: 512, 6:...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>78.376829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3309.474927</td>\n",
       "      <td>3303.131186</td>\n",
       "      <td>3303.131186</td>\n",
       "      <td>3303.131186</td>\n",
       "      <td>3302.656003</td>\n",
       "      <td>76.830914</td>\n",
       "      <td>76.830914</td>\n",
       "      <td>76.830914</td>\n",
       "      <td>79.571714</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...</td>\n",
       "      <td>{1: 704, 2: 1856, 3: 1664, 4: 1664, 5: 704, 6:...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>76.830914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3537.553369</td>\n",
       "      <td>3512.950746</td>\n",
       "      <td>3512.950746</td>\n",
       "      <td>3512.950746</td>\n",
       "      <td>3512.950746</td>\n",
       "      <td>71.181257</td>\n",
       "      <td>71.181257</td>\n",
       "      <td>71.181257</td>\n",
       "      <td>71.181257</td>\n",
       "      <td>{1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...</td>\n",
       "      <td>{1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...</td>\n",
       "      <td>{1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...</td>\n",
       "      <td>{1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>71.181257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3765.631811</td>\n",
       "      <td>3763.767309</td>\n",
       "      <td>3763.767309</td>\n",
       "      <td>3763.767309</td>\n",
       "      <td>3716.532646</td>\n",
       "      <td>71.473400</td>\n",
       "      <td>71.473400</td>\n",
       "      <td>71.473400</td>\n",
       "      <td>70.973257</td>\n",
       "      <td>{1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...</td>\n",
       "      <td>{1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...</td>\n",
       "      <td>{1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...</td>\n",
       "      <td>{1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>70.973257</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Performance_Constraint  BCR_disabled_RT  BCR_M/RT_RT  BCR_C/ERT_RT  \\\n",
       "0             2853.318043      2851.234229  2851.234229   2851.234229   \n",
       "1             3081.396485      3070.231189  3070.231189   3070.231189   \n",
       "2             3309.474927      3303.131186  3303.131186   3303.131186   \n",
       "3             3537.553369      3512.950746  3512.950746   3512.950746   \n",
       "4             3765.631811      3763.767309  3763.767309   3763.767309   \n",
       "\n",
       "    BCR_MAX_RT  BCR_disabled_Cost  BCR_M/RT_Cost  BCR_C/ERT_Cost  \\\n",
       "0  2852.520811          87.416771      87.416771       87.416771   \n",
       "1  3061.305594          87.150714      87.150714       87.150714   \n",
       "2  3302.656003          76.830914      76.830914       76.830914   \n",
       "3  3512.950746          71.181257      71.181257       71.181257   \n",
       "4  3716.532646          71.473400      71.473400       71.473400   \n",
       "\n",
       "   BCR_MAX_Cost                                BCR_disabled_Config  \\\n",
       "0     91.295400  {1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...   \n",
       "1     78.376829  {1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...   \n",
       "2     79.571714  {1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...   \n",
       "3     71.181257  {1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...   \n",
       "4     70.973257  {1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...   \n",
       "\n",
       "                                     BCR_M/RT_Config  \\\n",
       "0  {1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...   \n",
       "1  {1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...   \n",
       "2  {1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...   \n",
       "3  {1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...   \n",
       "4  {1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...   \n",
       "\n",
       "                                    BCR_C/ERT_Config  \\\n",
       "0  {1: 896, 2: 1856, 3: 1664, 4: 2240, 5: 1088, 6...   \n",
       "1  {1: 704, 2: 1856, 3: 1664, 4: 2240, 5: 704, 6:...   \n",
       "2  {1: 704, 2: 1856, 3: 1664, 4: 1280, 5: 1088, 6...   \n",
       "3  {1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...   \n",
       "4  {1: 704, 2: 1664, 3: 1664, 4: 896, 5: 512, 6: ...   \n",
       "\n",
       "                                      BCR_MAX_Config  BCR_disabled_Iterations  \\\n",
       "0  {1: 1088, 2: 2048, 3: 1664, 4: 2240, 5: 512, 6...                      9.0   \n",
       "1  {1: 896, 2: 1472, 3: 1664, 4: 1664, 5: 512, 6:...                      9.0   \n",
       "2  {1: 704, 2: 1856, 3: 1664, 4: 1664, 5: 704, 6:...                      9.0   \n",
       "3  {1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...                      9.0   \n",
       "4  {1: 704, 2: 1280, 3: 1088, 4: 1280, 5: 512, 6:...                      9.0   \n",
       "\n",
       "   BCR_M/RT_Iterations  BCR_C/ERT_Iterations  BCR_MAX_Iterations  \\\n",
       "0                  9.0                   9.0                15.0   \n",
       "1                  9.0                   9.0                19.0   \n",
       "2                  9.0                   9.0                16.0   \n",
       "3                  9.0                   9.0                19.0   \n",
       "4                  9.0                   9.0                20.0   \n",
       "\n",
       "   BCR_threshold  Best_Answer_Cost  \n",
       "0            0.2         87.416771  \n",
       "1            0.2         78.376829  \n",
       "2            0.2         76.830914  \n",
       "3            0.2         71.181257  \n",
       "4            0.2         70.973257  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "App6_BCPC_PRCP.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d48a8e9-70c1-48ed-9fe1-95da029713b6",
   "metadata": {},
   "source": [
    "### BPBC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "d9ff0437-bbb9-46c8-a8c7-990b694b6493",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(6,4))\n",
    "ax=plt.subplot(111)\n",
    "ax.grid()\n",
    "\n",
    "results = optimization_results.query(\"problem=='BPBC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "constraint_list = results['constraint'].to_list()\n",
    "result_list = results['duration_avg'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "result_list = results['duration_percentile90'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "result_list = results['duration_percentile10'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "ax.plot(constraint_list, App6_BPBC_PRCP['Best_Answer_RT'], marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "# ax.plot(opt_curve_data_BPBC['Budget'], opt_curve_data_BPBC['BCR_MAX_RT'], marker='o', label='BCR MAX', markersize=2, linewidth=1) \n",
    "# ax.plot(opt_curve_data_BPBC['Budget'], opt_curve_data_BPBC['BCR_ERT/C_RT'], marker='o', label='BCR ERT/C', markersize=2, linewidth=1)\n",
    "# ax.plot(opt_curve_data_BPBC['Budget'], opt_curve_data_BPBC['BCR_RT/M_RT'], marker='o', label='BCR RT/M', markersize=2, linewidth=1)  \n",
    "# ax.plot(opt_curve_data_BPBC['Budget'], opt_curve_data_BPBC['Best_Answer_RT'] , marker='o', label='Best Answer', markersize=2, linewidth=1) \n",
    "# ax.plot(opt_curve_data_BPBC['Budget'], best_rt, marker='o', label='Ideal BPBC Solution', markersize=2, linewidth=1) \n",
    "# ax.legend(loc='upper right', bbox_to_anchor=(1,0.86))\n",
    "# ax.set_xlabel('Budget Constraint in USD (per 1 Million Executions)')\n",
    "# ax.set_ylabel('End-to-end Response time in ms')\n",
    "# ax.set_ylim(2500,6500)\n",
    "# ax2 = ax.twinx()\n",
    "# ax2.plot(opt_curve_data_BPBC['Budget'], BPBC_accuracy, marker = 'x', label='Accuracy', color='magenta')\n",
    "# ax2.legend(loc='upper right', bbox_to_anchor=(1,0.96))\n",
    "# ax2.set_ylim(0,100)\n",
    "# ax2.set_ylabel('Accuracy in Percentage')\n",
    "# fig.savefig('Optimization_Curve_BPBC', dpi=300)\n",
    "fig.savefig('Optimization_Curve_BPBC.pdf')\n",
    "# print('Average Accuracy: ',BPBC_accuracy.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9377cfd6-f0a8-4825-be76-b4200bd5320d",
   "metadata": {},
   "outputs": [],
   "source": [
    "results.to_csv(\"BPBC_avg.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75f5e7d1-ba4d-4561-8254-8b60927aaa9b",
   "metadata": {},
   "source": [
    "### BCPC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "a05cad8d-3e5e-4fad-802a-ad7e3b6c7210",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(6,4))\n",
    "ax=plt.subplot(111)\n",
    "ax.grid()\n",
    "\n",
    "results = optimization_results.query(\"problem=='BCPC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "constraint_list = results['constraint'].to_list()\n",
    "result_list = results['cost_avg'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "result_list = results['cost_percentile90'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "result_list = results['cost_percentile10'].to_list()\n",
    "ax.plot(constraint_list, result_list, marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "ax.plot(constraint_list, App6_BCPC_PRCP['Best_Answer_Cost'], marker='o', label='Without BCR', markersize=2, linewidth=1)\n",
    "fig.savefig('Optimization_Curve_BCPC.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "681c6f66-b2d0-4516-a903-4cc726f84e46",
   "metadata": {},
   "outputs": [],
   "source": [
    "results.to_csv(\"BCPC_avg.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6e5c34c-4734-41a6-8208-e9de30053f0f",
   "metadata": {},
   "source": [
    "### Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "aa17341f-d5a3-499f-9dd3-70292984b333",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99.0886138652194"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = optimization_results.query(\"problem=='BPBC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "result_list = results['duration_avg']\n",
    "accuracy = (np.array(result_list) - np.array(best_rt))/np.array(best_rt).tolist()\n",
    "accuracy = [num * 100 if num > 0 else 0 for num in accuracy]\n",
    "accuracy = [0 if num < 1 else num for num in accuracy]\n",
    "BPBC_accuracy = [100 - num for num in accuracy]\n",
    "np.mean(BPBC_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "1fe4abd5-7164-4964-8aeb-b6149dc2193f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99.74424843610724"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = optimization_results.query(\"problem=='BCPC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "result_list = results['cost_avg']\n",
    "accuracy = (np.array(result_list) - np.array(best_cost))/np.array(best_cost).tolist()\n",
    "accuracy = [num * 100 if num > 0 else 0 for num in accuracy]\n",
    "accuracy = [0 if num < 1 else num for num in accuracy]\n",
    "BCPC_accuracy = [100 - num for num in accuracy]\n",
    "np.mean(BCPC_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "119fc7d2-7e65-424f-94ed-830ae02b7998",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x331.2 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(nrows=2, ncols=1, figsize=(6, 4.6))\n",
    "ax1.grid()\n",
    "\n",
    "results = optimization_results.query(\"problem=='BPBC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "constraint_list = results['constraint'].to_list()\n",
    "result_list = results['duration_avg'].to_list()\n",
    "ax1.plot(constraint_list, result_list, marker='o', label='Mean (DFBA)', markersize=2, linewidth=1)\n",
    "result_list = results['duration_percentile90'].to_list()\n",
    "ax1.plot(constraint_list, result_list, marker='o', label='90th %tile (DFBA)', markersize=2, linewidth=1)\n",
    "result_list = results['duration_percentile10'].to_list()\n",
    "ax1.plot(constraint_list, result_list, marker='o', label='10th %tile (DFBA)', markersize=2, linewidth=1)\n",
    "ax1.plot(constraint_list, App6_BPBC_PRCP['Best_Answer_RT'], marker='o', label='PRCP Best Solution', markersize=2, linewidth=1, alpha=0.4)\n",
    "ax1.plot(constraint_list, best_rt, marker='o', label='Ideal Solution', markersize=2, linewidth=1, alpha=0.5)\n",
    "ax1_twin = ax1.twinx()\n",
    "ax1_twin.plot(constraint_list, BPBC_accuracy, marker = 'x', label='DFBA Accuracy', color='grey', alpha=0.6)\n",
    "ax1_twin.set_ylim(40, 100)\n",
    "ax1.legend()\n",
    "# ax1_twin.legend()\n",
    "# ax1_twin.legend(loc='upper right', bbox_to_anchor=(0.945, 0.37))\n",
    "ax1.set_xlabel(\"Budget Constraint in USD (per 1 Million Executions)\")\n",
    "ax1.set_ylabel(\"Response Time in ms\")\n",
    "ax2.grid()\n",
    "\n",
    "results = optimization_results.query(\"problem=='BCPC' and percentile.isnull()\").sort_values(by='constraint')\n",
    "constraint_list = results['constraint'].to_list()\n",
    "result_list = results['cost_avg'].to_list()\n",
    "ax2.plot(constraint_list, result_list, marker='o', label='Mean', markersize=2, linewidth=1)\n",
    "result_list = results['cost_percentile90'].to_list()\n",
    "ax2.plot(constraint_list, result_list, marker='o', label='90-th Percentile', markersize=2, linewidth=1)\n",
    "result_list = results['cost_percentile10'].to_list()\n",
    "ax2.plot(constraint_list, result_list, marker='o', label='10-th Percentile', markersize=2, linewidth=1)\n",
    "ax2.plot(constraint_list, App6_BCPC_PRCP['Best_Answer_Cost'], marker='o', label='PRCP Best Solution', markersize=2, linewidth=1, alpha=0.4)\n",
    "ax2.plot(constraint_list, best_cost, marker='o', label='Ideal Solution', markersize=2, linewidth=1, alpha=0.5)\n",
    "ax2_twin = ax2.twinx()\n",
    "ax2_twin.plot(constraint_list, BCPC_accuracy, marker = 'x', label='DFBA Accuracy', color='grey', alpha=0.6)\n",
    "ax2_twin.legend(loc='upper right', bbox_to_anchor=(0.945, 0.96))\n",
    "ax2_twin.set_ylim(40, 100)\n",
    "ax2.set_xlabel(\"Performance Constraint (Response Time) in ms\")\n",
    "ax2.set_ylabel(\"Cost per 1 Million Executions\")\n",
    "\n",
    "fig.text(0.98, 0.5, 'Accuracy of the DFBA Algorithm in Percentage', va='center', ha='center', rotation='vertical')\n",
    "\n",
    "fig.tight_layout(rect=[0, 0, 0.98, 1])\n",
    "plt.show()\n",
    "fig.savefig(\"DFBA-PRCP.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ff32234b-dc71-4157-85e7-6d0f5dd16813",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9",
   "language": "python",
   "name": "python3.9"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}