SM286D · Introduction to Applied Mathematics with Python · Spring 2020 · Uhan
\n",
"\n",
"
Lesson 19.
\n",
"\n",
"
More data visualization
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## This lesson..."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"In this lesson, we'll walk through the inner workings of an SMO capstone project by MIDN 1/C Arvin, MIDN 1/C Kugel, MIDN 1/C Urrutia, and MIDN 1/C Villacorta (Class of 2020). \n",
"\n",
"The goal of this lesson is three-fold:\n",
"\n",
"- See an example of how to visualize the output of an optimization model in a user-friendly way.\n",
"\n",
"- See more examples of Pyomo and Cartopy in action.\n",
"\n",
"- See an example of a *real* capstone project that tries to solve an actual Navy problem.\n",
"\n",
"As you'll see, all the Python programming for this project is well within your grasp, based on the programming knowledge you gained this semester! You'll learn about some of the more advanced mathematical methods in future courses."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Problem description"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The goal of their project was to help NAVAIR determine the optimal locations for different types of 3D printers for a given budget. They formulated a facility location model that is described below. You'll cover the details of how to model this type of problem in SA405 — Advanced Mathematical Programming. For this lesson, we'll focus on visualizing the output of the optimization model. \n",
"\n",
"First, look over the problem statement and assumptions below. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Given a list of customer locations, which can also serve as locations for 3D printers, and a fixed budget, determine locations for the printers that minimize overall cost. We make the following assumptions:\n",
"\n",
" - There are 23 possible customer locations in the continental US.\n",
" - There are two tiers of 3D printers, tier 1 and tier 2.\n",
" - It is possible to place between 0 and 30 printers at any given location.\n",
" - The demands for customer locations were determined based on data from NAVAIR.\n",
" - There is a penalty to be paid if demand can't be satisfied for a given customer location.\n",
" - Shipping costs were determined based on FedEx shipping rates.\n",
" - There is a uniform one-time charge to purchase tier 1 or tier 2 printers. These were determined based on market research. \n",
" - There is a one-time start-up cost if any 3D printers are introduced at a particular location. \n",
" - 3D printers have a maximum capacity to print parts, and can only print parts corresponding to their respective tier.\n",
" - The distance between possible printer locations are based on the straight line distance between their latitudes and longitudes.\n",
" - The total budget is a fixed dollar amount.\n",
"\n",
"Based on these assumptions, the model formulation is given below. Look it over, and then read the description below."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Indices and sets.__\n",
"\n",
"\\begin{array}{ll}\n",
" i \\in I & \\mbox{customers}, I = \\{1, 2, \\dots, 23\\} \\\\\n",
" j \\in J & \\mbox{possible printer locations}, J = \\{1, 2, ..., 23\\} \\\\\n",
" p \\in P & \\mbox{possible number of printers at a given location}, P = \\{0, 1, 2, ..., 30\\} \\\\\n",
"\\end{array}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Data.__\n",
"\n",
"\\begin{array}{ll}\n",
" \\mathit{demand1}_i & \\mbox{tier 1 customer demand at node $i$} \\\\\n",
" \\mathit{demand2}_i & \\mbox{tier 2 customer demand at node $i$} \\\\\n",
" \\mathit{penalty1} & \\mbox{penalty for not satisfying tier 1 demands} \\\\\n",
" \\mathit{penalty2} & \\mbox{penalty for not satisfying tier 2 demands} \\\\\n",
" \\beta & \\mbox{proportionality constant for shipping costs} \\\\\n",
" \\mathit{f1}_{j} & \\mbox{one-time fixed cost for a tier 1 printer at location $j$} \\\\\n",
" \\mathit{f2}_{j} & \\mbox{one-time fixed cost for a tier 2 printer at location $j$} \\\\\n",
" \\mathit{start\\_cost1}_{j} & \\mbox{start-up cost associated with putting any tier 1 printers at location $j$} \\\\\n",
" \\mathit{start\\_cost2}_{j} & \\mbox{start-up cost associated with putting any tier 2 printers at location $j$} \\\\\n",
" \\mathit{cap1}_j & \\mbox{capacity of a tier 1 printer at location $j$} \\\\\n",
" \\mathit{cap2}_j & \\mbox{capacity of a tier 2 printer at location $j$} \\\\\n",
" \\mathit{dist}_{ij} & \\mbox{length of shortest path between customer $i$ and printer location $j$} \\\\\n",
" \\mathit{budget} & \\mbox{total amount of money available} \\\\\n",
" \\mathit{M} & \\mbox{big enough integer value (set by the length of $P$)} \\\\\n",
"\\end{array}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Decision variables.__\n",
"\n",
"\\begin{array}{ll}\n",
" \\mathit{FRACUN1}_i & \\mbox{fraction of tier 1 customer $i$ demand unfulfilled} \\\\\n",
" \\mathit{FRACUN2}_i & \\mbox{fraction of tier 2 customer $i$ demand unfulfilled} \\\\\n",
" \\mathit{OPEN1}_{jp} & \\mbox{1 if $p$ tier 1 printers are located at $j$, 0 otherwise} \\\\\n",
" \\mathit{OPEN2}_{jp} & \\mbox{1 if $p$ tier 2 printers are located at $j$, 0 otherwise} \\\\\n",
" \\mathit{SAT1}_{ij} & \\mbox{fraction of tier 1 customer $i$ demand satisfied by printer location $j$} \\\\\n",
" \\mathit{SAT2}_{ij} & \\mbox{fraction of tier 2 customer $i$ demand satisfied by printer location $j$} \\\\\n",
" \\mathit{NOT0P1}_j & \\mbox{1 if any tier 1 printers open at location $j$, 0 otherwise} \\\\\n",
" \\mathit{NOT0P2}_j & \\mbox{1 if any tier 2 printers open at location $j$, 0 otherwise} \\\\\n",
"\\end{array}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Formulation.__\n",
"\n",
"\\begin{alignat}{4}\n",
" \\min \\quad & \\sum_{j \\in J, p \\in P} p \\cdot f1_{j} \\mathit{OPEN1}_{jp} + \\sum_{j \\in J, p \\in P} p \\cdot f2_{j} \\mathit{OPEN2}_{jp} \\\\\n",
" & \\qquad + \\beta \\sum_{i \\in I, j \\in J} \\mathit{demand1}_i dist_{ij} \\mathit{SAT1}_{ij} + \\beta \\sum_{i \\in I, j \\in J} \\mathit{demand2}_i \\mathit{dist}_{ij} \\mathit{SAT2}_{ij} \\\\\n",
" & \\qquad + \\sum_{i \\in I} \\mathit{penalty1} \\mathit{demand1}_i \\mathit{FRACUN1}_i + \\sum_{i \\in I} \\mathit{penalty2} \\mathit{demand2}_i \\mathit{FRACUN2}_i\\\\\n",
" & \\qquad + \\sum_{j \\in J} \\mathit{start\\_cost1}_{j} \\mathit{NOT0P1}_{j} + \\sum_{j \\in J} \\mathit{start\\_cost2}_{j} \\mathit{NOT0P2}_{j} \\\\\n",
" \\mbox{s.t.} \\quad & \\sum_{j \\in J, p \\in P} p \\cdot f1_{j} \\mathit{OPEN1}_{jp} + \\sum_{j \\in J, p \\in P} p \\cdot f2_{j} \\mathit{OPEN2}_{jp} \\\\\n",
" & \\qquad + \\sum_{j \\in J} \\mathit{start\\_cost1}_{j} \\mathit{NOT0P1}_{j} + \\sum_{j \\in J} \\mathit{start\\_cost2}_{j} \\mathit{NOT0P2}_{j} \\le \\mathit{budget} & \\qquad & & \\quad (1) \\\\\n",
" & \\mathit{FRACUN1}_{i} + \\sum_{j\\in J} \\mathit{SAT1}_{ij} = 1 & \\qquad & \\forall i \\in I & \\quad (2) \\\\\n",
" & \\mathit{FRACUN2}_{i} + \\sum_{j\\in J} \\mathit{SAT2}_{ij} = 1 & \\qquad & \\forall i \\in I & \\quad (3) \\\\\n",
" & \\sum_{p \\in P} \\mathit{OPEN1}_{jp} = 1 & & \\forall j \\in J & \\quad (4) \\\\\n",
" & \\sum_{p \\in P} \\mathit{OPEN2}_{jp} = 1 & & \\forall j \\in J & \\quad (5) \\\\\n",
" & \\sum_{p \\in P} p \\cdot \\mathit{OPEN1}_{jp} \\le M \\mathit{NOT0P1}_j & & \\forall j \\in J & \\quad (6) \\\\\n",
" & \\sum_{p \\in P} p \\cdot \\mathit{OPEN2}_{jp} \\le M \\mathit{NOT0P2}_j & & \\forall j \\in J & \\quad (7) \\\\\n",
" & \\sum_{p \\in P} p \\cdot \\mathit{OPEN1}_{jp} \\ge \\mathit{NOT0P1}_j & & \\forall j \\in J & \\quad (8) \\\\\n",
" & \\sum_{p \\in P} p \\cdot \\mathit{OPEN2}_{jp} \\ge \\mathit{NOT0P2}_j & & \\forall j \\in J & \\quad (9) \\\\\n",
" & \\sum_{i \\in I} \\mathit{demand1}_i \\mathit{SAT1}_{ij} \\le \\sum_{p \\in P} p \\cdot \\mathit{cap1}_j \\mathit{OPEN1}_{jp} & & \\forall j \\in J & \\quad (10) \\\\\n",
" & \\sum_{i \\in I} \\mathit{demand2}_i \\mathit{SAT2}_{ij} \\le \\sum_{p \\in P} p \\cdot \\mathit{cap2}_j \\mathit{OPEN2}_{jp} & & \\forall j \\in J & \\quad (11) \\\\\n",
" & 0 \\le \\mathit{FRACUN1}_i \\le 1 & & \\forall i \\in I & \\quad (12) \\\\\n",
" & 0 \\le \\mathit{FRACUN2}_i \\le 1 & & \\forall i \\in I & \\quad (13) \\\\\n",
" & \\mathit{OPEN1}_{jp} \\in \\{0,1\\} & & \\forall j \\in J, p \\in P & \\quad (14) \\\\\n",
" & \\mathit{OPEN2}_{jp} \\in \\{0,1\\} & & \\forall j \\in J, p \\in P & \\quad (15) \\\\\n",
" & 0 \\le \\mathit{SAT1}_{ij} \\le 1 & & \\forall i \\in I, j \\in J & \\quad (16) \\\\\n",
" & 0 \\le \\mathit{SAT2}_{ij} \\le 1 & & \\forall i \\in I, j \\in J & \\quad (17) \\\\\n",
" & \\mathit{NOT0P1}_j \\in \\{0,1\\} & & \\forall j \\in J & \\quad (18) \\\\\n",
" & \\mathit{NOT0P2}_j \\in \\{0,1\\} & & \\forall j \\in J & \\quad (19)\n",
"\\end{alignat}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"- The objective function seeks to minimize the overall cost, which consists of the following components: cost of the tier 1 and tier 2 printers purchased, the shipping costs to transport parts from where they are printed to where they are needed, the penalties for any unsatisfied part demands, and the start-up costs associated with placing printers at different locations. \n",
"\n",
"- Constraint (1) ensures that the cost to buy the printers and pay the start-up costs does not exceed the total budget. \n",
"\n",
"- Constraints (2) and (3) ensure that the sum of satisfied and unsatisfied demand is one at each customer location, for tiers 1 and 2, respectively.\n",
"\n",
"- Constraints (4) and (5) ensure exactly one number (between 0 and 30) of printer is open at each possible location, for tiers 1 and 2, respectively. \n",
"\n",
"- Constraints (6), (7), (8), and (9) are logical constraints that relate the two decision variables $\\mathit{OPEN}$ and $\\mathit{NOT0P}$ for their respective tiers. These logical constraints ensure the value of $\\mathit{NOT0P}$ is set to 1 if there are any 3D printers at a particular location, and 0 if there are no 3D printers at a particular location. \n",
"\n",
"- Constraints (10) and (11) ensure that demand can only be satisfied if there are enough 3D printers open at a location to produce the parts for tiers 1 and 2, respectively. \n",
"\n",
"- Constraints (12) - (19) are variable bounds. "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 1.
\n",
" \n",
"Based on the cardinality of the sets $J$ and $P$ provided above, how many constraints of type (1) will be in the model?\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Write your answer here. Double-click to edit.__\n",
"\n",
"- There will be only one type (1) constraint in the model."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 2.
\n",
" \n",
"Based on the cardinality of the sets $I$, $J$, and $P$ provided above, how many constraints of type (10) will be in the model?\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Write your answer here. Double-click to edit.__\n",
"\n",
"- There will be 23 type (10) constraints in the model."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 3
\n",
"\n",
"Is $\\mathit{FRACUN1}_i$ a continuous, integer, or binary variable? How many different $\\mathit{FRACUN1}_i$ variables will there be in the model?\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Write your answer here. Double-click to edit.__\n",
"\n",
"- $\\mathit{FRACUN1}_i$ is a continuous variable that can take on values from 0 to 1.\n",
"- There will be 23 different $\\mathit{FRACUN1}_i$ variables in the model."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 4
\n",
"\n",
"Is $\\mathit{OPEN1}_{jp}$ a continuous, integer, or binary variable? How many different $\\mathit{OPEN1}_{jp}$ variables will there be in the model?\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"__Write your answer here. Double-click to edit.__\n",
"\n",
"- $\\mathit{OPEN1}_{jp}$ is a binary variable that can take on the values of 0 or 1.\n",
"- There will be $23 \\times 31 = 713$ different $\\mathit{OPEN1}_{jp}$ variables in the model."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Reading and generating the input data"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The cell below reads most of the required information for the model from the Excel workbook `3DPrintingFacilityLocation.xlsx` (located in the same folder as this notebook). Run this code cell.\n",
"\n",
" "
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [],
"source": [
"# Import pyomo and xlwings\n",
"import pyomo.environ as pyo\n",
"import xlwings as xw\n",
"\n",
"# Create Book object pointing to Excel workbook\n",
"# NOTE: It is easier to not have spaces in the path or the file name\n",
"wb = xw.Book('3DPrintingFacilityLocation.xlsx')\n",
"\n",
"# Reference customer input information\n",
"information = wb.sheets['Information']\n",
"\n",
"# Reference initials sheet in Excel\n",
"initials = wb.sheets['Initials']\n",
"\n",
"# Customer values from Excel\n",
"customers = information.range('A2').expand('down').value\n",
"\n",
"# We assume all customer locations can also be printer locations\n",
"facilities = customers\n",
"\n",
"# Demand values from Excel\n",
"demand01 = information.range('E2').expand('down').value\n",
"demand02 = information.range('F2').expand('down').value\n",
"\n",
"# Penalty for printers\n",
"penalty1 = initials.range('E3').value\n",
"penalty2 = initials.range('E4').value\n",
"\n",
"# Startup cost\n",
"startup1 = initials.range('F3').value\n",
"startup2 = initials.range('F4').value\n",
"\n",
"# Total budget\n",
"TOTAL_BUDGET = initials.range('B2').value\n",
"\n",
"# Take distance into account\n",
"beta = initials.range('E7').value\n",
"\n",
"# Theoretical number of printers for each type\n",
"max_printers = initials.range('B6').value\n",
"\n",
"# Create list with all possible number of printers\n",
"num_printers = list(range(int(max_printers)))\n",
"\n",
"# Set value of \"Big M\"\n",
"M = len(num_printers) + 2\n",
"\n",
"# Demand dictionary for printer 1\n",
"demand1 = {}\n",
"for i in range(len(customers)):\n",
" demand1[customers[i]] = demand01[i]\n",
"\n",
"# Demand Dictionary for printer 2\n",
"demand2 = {}\n",
"for i in range(len(customers)):\n",
" demand2[customers[i]] = demand02[i]\n",
"\n",
"# Gets the lats and longs for each location\n",
"lats = information.range('B2').expand('down').value\n",
"longs = information.range('C2').expand('down').value"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The cell below contains the `airdist` function we wrote in Lesson 18. Run this code cell."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [],
"source": [
"import numpy as np\n",
"\n",
"def airdist(lat1, long1, lat2, long2):\n",
" \"\"\"\n",
" Compute the distance (as the crow flies) between two points\n",
" whose latitudes and longitudes are given. \n",
" \"\"\"\n",
" # Write your code here\n",
" # Convert to radians\n",
" lat1_rad = lat1 * np.pi / 180\n",
" lat2_rad = lat2 * np.pi / 180\n",
" long1_rad = long1 * np.pi / 180\n",
" long2_rad = long2 * np.pi / 180\n",
" \n",
" # Radius of Earth in meters\n",
" R = 6371000\n",
" \n",
" # Distance in meters\n",
" meters = R * np.arccos(\n",
" np.sin(lat1_rad) * np.sin(lat2_rad) + \n",
" np.cos(lat1_rad) * np.cos(lat2_rad) * np.cos(long2_rad - long1_rad)\n",
" )\n",
" miles = meters * 0.621371 / 1000\n",
" \n",
" return miles"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 5
\n",
"\n",
"In the code cell above where `airdist` was defined, the lists `customers`, `facilities`, `lats`, and `longs` were created by reading the appropriate values from the Excel workbook. Note that `customers` and `facilities` contain the same locations based on the assumptions we made above. \n",
" \n",
"In the cell below, write a `for` loop that prints the values of `customers` with the corresponding `lats` and `longs` values to get a feel for the data structures. Use f-strings so that your output looks like the example below, which only contains the first 4 entries in the lists. Your answer should contain all 23 customer locations and their corresponding latitudes and longitudes.\n",
"\n",
"```\n",
"Customer Location Name latitude longitude\n",
"---------------------- -------- ---------\n",
"New River, NC 34.757 -77.410\n",
"Camp Pendleton, CA 33.318 -117.320\n",
"Lakehurst, NJ 40.015 -74.311\n",
"Yuma, AZ 32.692 -114.628\n",
"```\n",
"
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Customer Location Name Latitude Longitude\n",
"---------------------- -------- ---------\n",
"New River, NC 34.757 -77.410\n",
"Camp Pendleton, CA 33.318 -117.320\n",
"Lakehurst, NJ 40.015 -74.311\n",
"Yuma, AZ 32.692 -114.628\n",
"Cherry Point, NC 34.903 -76.897\n",
"Norfolk, VA 36.851 -76.286\n",
"NAS North Island, CA 32.698 -117.204\n",
"NAS Pax River, MD 38.275 -76.446\n",
"MCAS Miramar, CA 32.870 -117.144\n",
"Macguire AFB, NJ 40.035 -74.588\n",
"NAS Pt. Mugu, CA 34.128 -119.095\n",
"NAS Fallon, NV 39.420 -118.725\n",
"NAS Whidbey Island, WA 48.431 -122.672\n",
"NAWS China Lake, CA 35.655 -117.657\n",
"Tinker AFB, OK 35.428 -97.416\n",
"NAS Pensacola, FL 30.378 -87.288\n",
"NASJRB Fort Worth, TX 32.765 -97.420\n",
"MCAS Beaufort, SC 32.475 -80.715\n",
"NAS Lemoore, CA 36.261 -119.911\n",
"NASJRB New Orleans, LA 29.824 -90.014\n",
"Stewart ANGB, NY 41.502 -74.083\n",
"MCAF Quantico, VA 38.502 -77.306\n",
"NAS Jacksonville, FL 30.223 -81.685\n"
]
}
],
"source": [
"# SOLUTION\n",
"# Print table\n",
"print(\"Customer Location Name Latitude Longitude\")\n",
"print(\"---------------------- -------- ---------\")\n",
"for i in range(len(customers)):\n",
" print(f\"{customers[i]:22s} {lats[i]:8.3f} {longs[i]:9.3f}\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 6
\n",
"\n",
"In the cell below, use a nested `for` loop, appropriate `if-else` logic, and the function `airdist` to create the dictionary, `dist`, of distances from each customer to each facility. The keys of `dist` should be tuples with 2 elements: first, the customer, and second, the facility.\n",
" \n",
"For example, `dist[(customers[2], facilities[4])]` should be the distance between customer 2 and facility 4.\n",
" \n",
"Make sure to set the value of `dist` to 0 whenever the customer and facility are the same location.\n",
"
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [],
"source": [
"# SOLUTION\n",
"dist = {}\n",
"for i in range(len(customers)):\n",
" for j in range(len(facilities)):\n",
" if i == j:\n",
" dist[(customers[i], facilities[j])] = 0\n",
" else:\n",
" dist[(customers[i], facilities[j])] = airdist(lats[i], longs[i], lats[j], longs[j])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Creating the Pyomo version of the optimization model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The code in the cell below creates a Pyomo model of the formulation above, using the information for the printer costs and capacities from the Excel workbook. Run this code."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [],
"source": [
"# Create concrete model\n",
"model = pyo.ConcreteModel()\n",
"\n",
"# Model sets\n",
"model.I = pyo.Set(initialize=customers, doc='Set of customers')\n",
"model.J = pyo.Set(initialize=facilities, doc='Set of possible printer locations')\n",
"model.P = pyo.Set(initialize=num_printers, doc='Set of possible number of printers')\n",
"\n",
"# Model parameters\n",
"# Fixed cost dictionary\n",
"fixedcost1 = {}\n",
"rngfixedcost1 = information.range('K2').expand('down')\n",
"row_index1 = 0\n",
"for j in model.J:\n",
" fixedcost1[j] = rngfixedcost1[row_index1].value\n",
" row_index1 += 1\n",
"\n",
"fixedcost2 = {}\n",
"rngfixedcost2 = information.range('L2').expand('down')\n",
"row_index1 = 0\n",
"for j in model.J:\n",
" fixedcost2[j] = rngfixedcost2[row_index1].value\n",
" row_index1 += 1\n",
"\n",
"# Capacity dictionaries\n",
"capacity1 = {}\n",
"rngcap1 = information.range('H2').expand('down')\n",
"row_index1 = 0\n",
"for j in model.J:\n",
" capacity1[j] = rngcap1[row_index1].value\n",
" row_index1 += 1\n",
"\n",
"capacity2 = {}\n",
"rngcap2 = information.range('I2').expand('down')\n",
"row_index1 = 0\n",
"for j in model.J:\n",
" capacity2[j] = rngcap2[row_index1].value\n",
" row_index1 += 1\n",
"\n",
"model.dist = pyo.Param(model.I, model.J, initialize=dist,\n",
" doc='Length of shortest path between nodes i and j')\n",
"model.beta = pyo.Param(initialize=beta, doc='Takes Distance into account')\n",
"model.budget = pyo.Param(initialize=TOTAL_BUDGET, doc ='Total Budget')\n",
"\n",
"model.penalty1 = pyo.Param(initialize=penalty1, doc='Penalty value for tier 1')\n",
"model.penalty2 = pyo.Param(initialize=penalty2, doc='Penalty value for Tter 2')\n",
"\n",
"model.demand1 = pyo.Param(model.I, initialize=demand1,\n",
" doc='Demand value for customer i for tier 1 parts')\n",
"model.demand2 = pyo.Param(model.I, initialize=demand2,\n",
" doc='Demand value for customer i for tier 2 parts')\n",
"\n",
"model.f1 = pyo.Param(model.J, initialize=fixedcost1,\n",
" doc='Fixed cost for a tier 1 printer at location j')\n",
"model.f2 = pyo.Param(model.J, initialize=fixedcost2,\n",
" doc='Fixed cost for a tier 2 printer at location j')\n",
"\n",
"model.cap1 = pyo.Param(model.J, initialize=capacity1,\n",
" doc='Capacity of facility j for a tier 1 Printer')\n",
"model.cap2 = pyo.Param(model.J, initialize=capacity2,\n",
" doc='Capacity of facility j for a tier 2 Printer')\n",
"\n",
"model.start_cost1 = pyo.Param(model.J, initialize=startup1,\n",
" doc=\"Fee you pay if you open ANY tier 1 at location j\")\n",
"model.start_cost2 = pyo.Param(model.J, initialize=startup2,\n",
" doc=\"Fee you pay if you open ANY tier 2 at location j\")\n",
"\n",
"# Model decision variables\n",
"model.FRACUN1 = pyo.Var(model.I, bounds=(0.0, 1.0),\n",
" doc='Fraction of demand unfulfilled for tier 1')\n",
"model.FRACUN2 = pyo.Var(model.I, bounds=(0.0, 1.0),\n",
" doc='Fraction of demand unfulfilled for tier 2')\n",
"\n",
"model.OPEN1 = pyo.Var(model.J, model.P, within=pyo.Binary,\n",
" doc='1 if j is a facility, 0 otherwise, for tier 1')\n",
"model.OPEN2 = pyo.Var(model.J, model.P, within=pyo.Binary,\n",
" doc='1 if j is a facility, 0 otherwise, for tier 2')\n",
"\n",
"model.NOT0P1 = pyo.Var(model.J, within=pyo.Binary,\n",
" doc=('1 if there are more than 0 tier 1 printers at location j, '\n",
" '0 if there no printers at j'))\n",
"model.NOT0P2 = pyo.Var(model.J, within=pyo.Binary,\n",
" doc=('1 if there are more than 0 tier 2 printers at location j, '\n",
" '0 if there no printers at j'))\n",
"\n",
"model.SAT1 = pyo.Var(model.I, model.J, bounds=(0.0, 1.0),\n",
" doc='Fraction of node i demand satisfied by facility j, for tier 1')\n",
"model.SAT2 = pyo.Var(model.I, model.J, bounds=(0.0, 1.0),\n",
" doc='Fraction of node i demand satisfied by facility j, for tier 2')\n",
"\n",
"# Model Objective\n",
"def cost_rule(model):\n",
" return (\n",
" sum(p * model.f1[j] * model.OPEN1[j,p] for j in model.J for p in model.P) +\n",
" sum(p * model.f2[j] * model.OPEN2[j,p] for j in model.J for p in model.P) +\n",
" sum(beta * model.demand1[i] * model.dist[i,j] * model.SAT1[i,j]\n",
" for i in model.I for j in model.J) +\n",
" sum(beta * model.demand2[i] * model.dist[i,j] * model.SAT2[i,j]\n",
" for i in model.I for j in model.J) +\n",
" sum(penalty1 * model.demand1[i] * model.FRACUN1[i] for i in model.I) +\n",
" sum(penalty2 * model.demand2[i] * model.FRACUN2[i] for i in model.I) +\n",
" sum(model.start_cost1[j] * model.OPEN1[j,p] for j in model.J for p in model.P) +\n",
" sum(model.start_cost2[j] * model.OPEN2[j,p] for j in model.J for p in model.P)\n",
" )\n",
"model.cost = pyo.Objective(rule=cost_rule, sense=pyo.minimize)\n",
"\n",
"# Model Constraints\n",
"def total_budget_rule(model):\n",
" return (\n",
" sum(p * model.f1[j] * model.OPEN1[j,p] for j in model.J for p in model.P) +\n",
" sum(model.start_cost1[j] * model.NOT0P1[j] for j in model.J) +\n",
" sum(p * model.f2[j] * model.OPEN2[j,p] for j in model.J for p in model.P) +\n",
" sum(model.start_cost2[j] * model.NOT0P2[j] for j in model.J) <= model.budget\n",
" )\n",
"model.total_budget_const = pyo.Constraint(rule=total_budget_rule,\n",
" doc='Cannot exceed total budget')\n",
"\n",
"def sat_plus_unsat_is_one_rule(model, i):\n",
" return sum(model.SAT1[i,j] for j in model.J) + model.FRACUN1[i] == 1\n",
"model.sat_plus_unsat_is_one_rule = pyo.Constraint(\n",
" model.I, rule=sat_plus_unsat_is_one_rule,\n",
" doc='Total demand met by sum of all facilities plus slack for tier 1'\n",
")\n",
"\n",
"def sat_plus_unsat_is_one_rule2(model, i):\n",
" return sum(model.SAT2[i,j] for j in model.J) + model.FRACUN2[i] == 1\n",
"model.sat_plus_unsat_is_one_rule2 = pyo.Constraint(\n",
" model.I, rule=sat_plus_unsat_is_one_rule2,\n",
" doc='Total demand met by sum of all facilities plus slack for tier 2'\n",
")\n",
"\n",
"def one_loc(model, j):\n",
" return sum(model.OPEN1[j,p] for p in model.P) == 1\n",
"model.one_loc = pyo.Constraint(\n",
" model.J, rule=one_loc,\n",
" doc='Exactly one number of printers at each location'\n",
")\n",
"\n",
"def one_loc2(model, j):\n",
" return sum(model.OPEN2[j,p] for p in model.P) == 1\n",
"model.one_loc2 = pyo.Constraint(\n",
" model.J, rule=one_loc2,\n",
" doc = 'Exactly one number of printers at each location'\n",
")\n",
"\n",
"def tier1_logic_one_rule(model, j):\n",
" return sum(p * model.OPEN1[j,p] for p in model.P) <= M * model.NOT0P1[j]\n",
"model.tier1_logic_one_rule = pyo.Constraint(\n",
" model.J, rule=tier1_logic_one_rule, doc='Force 1 for NOT0P1 when p > 0'\n",
")\n",
"\n",
"def tier2_logic_one_rule(model, j):\n",
" return sum(p * model.OPEN2[j,p] for p in model.P) <= M * model.NOT0P2[j]\n",
"model.tier2_logic_one_rule = pyo.Constraint(\n",
" model.J, rule=tier2_logic_one_rule, doc='Force 1 for NOT0P2 when p > 0'\n",
")\n",
"\n",
"def tier1_logic_two_rule(model, j):\n",
" return sum(p * model.OPEN1[j,p] for p in model.P) >= model.NOT0P1[j]\n",
"model.tier1_logic_two_rule = pyo.Constraint(\n",
" model.J, rule=tier1_logic_two_rule, doc='Force 0 for NOT0P1 when p = 0'\n",
")\n",
"\n",
"def tier2_logic_two_rule(model, j):\n",
" return sum(p * model.OPEN2[j,p] for p in model.P) >= model.NOT0P2[j]\n",
"model.tier2_logic_two_rule = pyo.Constraint(\n",
" model.J, rule=tier2_logic_two_rule, doc='Force 0 for NOT0P2 when p = 0'\n",
")\n",
"\n",
"def demand_met_only_if_open_rule(model, j):\n",
" return (\n",
" sum(model.demand1[i] * model.SAT1[i,j] for i in model.I) <=\n",
" sum(p * model.cap1[j] * model.OPEN1[j,p] for p in model.P)\n",
" )\n",
"model.demand_met_only_if_open_rule = pyo.Constraint(\n",
" model.J, rule=demand_met_only_if_open_rule,\n",
" doc='Facility can only satisfy demand if open'\n",
")\n",
"\n",
"def demand_met_only_if_open_rule2(model,j):\n",
" return (\n",
" sum(model.demand2[i] * model.SAT2[i,j] for i in model.I) <=\n",
" sum(p*model.cap2[j]*model.OPEN2[j,p] for p in model.P)\n",
" )\n",
"model.demand_met_only_if_open_rule2 = pyo.Constraint(\n",
" model.J, rule=demand_met_only_if_open_rule2,\n",
" doc='Facility can only satisfy demand if open'\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Solving the optimization model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The code in the cell below solves the model using the GLPK solver. The code\n",
"\n",
"```Python\n",
"opt.options[\"mipgap\"] = 0.10\n",
"```\n",
"\n",
"stops the solver when it is within 10% of the optimal solution. You'll learn more about what this means in SA405. We use this option here because this model can take a while to solve to optimality, especially when using the GLPK solver. \n",
"\n",
"Run this code, and wait until it says \"RELATIVE MIP GAP TOLERANCE REACHED; SEARCH TERMINATED\" before running any of the code below."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"GLPSOL: GLPK LP/MIP Solver, v4.65\n",
"Parameter(s) specified in the command line:\n",
" --mipgap 0.1 --write /var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmp18pnwm0k.glpk.raw\n",
" --wglp /var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpd77nvqzq.glpk.glp\n",
" --cpxlp /var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpsjxzoaq3.pyomo.lp\n",
"Reading problem data from '/var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpsjxzoaq3.pyomo.lp'...\n",
"/var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpsjxzoaq3.pyomo.lp:14690: warning: lower bound of variable 'x47' redefined\n",
"/var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpsjxzoaq3.pyomo.lp:14690: warning: upper bound of variable 'x47' redefined\n",
"232 rows, 2531 columns, 9017 non-zeros\n",
"1426 integer variables, all of which are binary\n",
"16116 lines were read\n",
"Writing problem data to '/var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmpd77nvqzq.glpk.glp'...\n",
"15558 lines were written\n",
"GLPK Integer Optimizer, v4.65\n",
"232 rows, 2531 columns, 9017 non-zeros\n",
"1426 integer variables, all of which are binary\n",
"Preprocessing...\n",
"46 hidden covering inequaliti(es) were detected\n",
"231 rows, 2392 columns, 8510 non-zeros\n",
"1334 integer variables, all of which are binary\n",
"Scaling...\n",
" A: min|aij| = 1.000e+00 max|aij| = 5.000e+06 ratio = 5.000e+06\n",
"GM: min|aij| = 2.098e-01 max|aij| = 4.766e+00 ratio = 2.271e+01\n",
"EQ: min|aij| = 4.681e-02 max|aij| = 1.000e+00 ratio = 2.136e+01\n",
"2N: min|aij| = 2.289e-02 max|aij| = 1.465e+00 ratio = 6.400e+01\n",
"Constructing initial basis...\n",
"Size of triangular part is 231\n",
"Solving LP relaxation...\n",
"GLPK Simplex Optimizer, v4.65\n",
"231 rows, 2392 columns, 8510 non-zeros\n",
" 0: obj = 6.583543000e+09 inf = 2.623e+02 (47)\n",
" 47: obj = 6.469143000e+09 inf = 2.220e-16 (0)\n",
"* 319: obj = 7.132248214e+08 inf = 6.559e-15 (0) 1\n",
"OPTIMAL LP SOLUTION FOUND\n",
"Integer optimization begins...\n",
"Long-step dual simplex will be used\n",
"+ 319: mip = not found yet >= -inf (1; 0)\n",
"+ 538: >>>>> 7.925315445e+08 >= 7.137752770e+08 9.9% (104; 0)\n",
"+ 538: mip = 7.925315445e+08 >= 7.137752770e+08 9.9% (100; 4)\n",
"RELATIVE MIP GAP TOLERANCE REACHED; SEARCH TERMINATED\n",
"Time used: 0.3 secs\n",
"Memory used: 2.9 Mb (3006952 bytes)\n",
"Writing MIP solution to '/var/folders/5k/rxb0jk152pb3hcrczw45mrm40000gn/T/tmp18pnwm0k.glpk.raw'...\n",
"2772 lines were written\n"
]
}
],
"source": [
"# Solve the model\n",
"from pyomo.opt import SolverFactory\n",
"opt = SolverFactory(\"glpk\")\n",
"\n",
"# Change MIP gap option in case it takes too long to solve\n",
"opt.options[\"mipgap\"] = 0.10\n",
"\n",
"results = opt.solve(model, tee = True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Displaying the optimization results as text"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Tier 1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The code in the cell below writes the tier 1 results from the optimization model to the screen and the Excel workbook on the **Results** worksheet:\n",
"\n",
"- The code produces a table of all the customer locations and how many printers are at that location in the optimal solution. \n",
"\n",
"- The code also produces a table that shows each customer location that has its demand filled, and the printer location that satisfied the demand along with the fraction that it supplied.\n",
"\n",
"- Finally, the code produces a table of locations that did not have all of their demand fulfilled along with the fraction of demand that was unfulfilled at that location. \n",
"\n",
"Look through this code, reading the comments to understand what different portions of the code are doing. Run this code."
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Results\n",
"\n",
"Minimum weighted distance is 792531544.4600626\n",
"\n",
"----------------------------------------------------------------\n",
"Tier 1\n",
"\n",
"Printer Location Number of Printers\n",
"---------------------- ------------------\n",
"New River, NC 0\n",
"Camp Pendleton, CA 27\n",
"Lakehurst, NJ 0\n",
"Yuma, AZ 0\n",
"Cherry Point, NC 0\n",
"Norfolk, VA 0\n",
"NAS North Island, CA 11\n",
"NAS Pax River, MD 0\n",
"MCAS Miramar, CA 0\n",
"Macguire AFB, NJ 0\n",
"NAS Pt. Mugu, CA 0\n",
"NAS Fallon, NV 0\n",
"NAS Whidbey Island, WA 0\n",
"NAWS China Lake, CA 0\n",
"Tinker AFB, OK 0\n",
"NAS Pensacola, FL 0\n",
"NASJRB Fort Worth, TX 0\n",
"MCAS Beaufort, SC 0\n",
"NAS Lemoore, CA 0\n",
"NASJRB New Orleans, LA 0\n",
"Stewart ANGB, NY 0\n",
"MCAF Quantico, VA 0\n",
"NAS Jacksonville, FL 0\n",
"\n",
"\n",
"Customer Location Printer Location Percent of Service\n",
"---------------------- ---------------------- ------------------\n",
"New River, NC Camp Pendleton, CA 1.000\n",
"Camp Pendleton, CA Camp Pendleton, CA 1.000\n",
"Lakehurst, NJ Camp Pendleton, CA 1.000\n",
"Yuma, AZ NAS North Island, CA 1.000\n",
"Cherry Point, NC Camp Pendleton, CA 1.000\n",
"Norfolk, VA Camp Pendleton, CA 1.000\n",
"NAS North Island, CA NAS North Island, CA 1.000\n",
"NAS Pax River, MD Camp Pendleton, CA 1.000\n",
"MCAS Miramar, CA NAS North Island, CA 1.000\n",
"Macguire AFB, NJ Camp Pendleton, CA 1.000\n",
"NAS Pt. Mugu, CA Camp Pendleton, CA 1.000\n",
"NAS Fallon, NV Camp Pendleton, CA 1.000\n",
"NAS Whidbey Island, WA Camp Pendleton, CA 1.000\n",
"NAWS China Lake, CA Camp Pendleton, CA 1.000\n",
"Tinker AFB, OK Camp Pendleton, CA 1.000\n",
"NAS Pensacola, FL NAS North Island, CA 1.000\n",
"NASJRB Fort Worth, TX NAS North Island, CA 1.000\n",
"MCAS Beaufort, SC Camp Pendleton, CA 0.709\n",
"MCAS Beaufort, SC NAS North Island, CA 0.291\n",
"NAS Lemoore, CA Camp Pendleton, CA 1.000\n",
"NASJRB New Orleans, LA NAS North Island, CA 1.000\n",
"Stewart ANGB, NY Camp Pendleton, CA 1.000\n",
"MCAF Quantico, VA Camp Pendleton, CA 1.000\n",
"NAS Jacksonville, FL NAS North Island, CA 1.000\n",
"\n",
"\n",
"Customer Location Fraction Unfulfilled\n",
"---------------------- -------------------\n"
]
}
],
"source": [
"# Worksheet for results\n",
"sresults = wb.sheets['Results']\n",
"\n",
"# Clear any existing information in the spreadsheet on the Results worksheet\n",
"sresults.clear_contents()\n",
"\n",
"# Write the objective value to the screen and generate the column headings in Excel\n",
"print(\"Results\\n\")\n",
"print(\"Minimum weighted distance is \" + str(pyo.value(model.cost)) +\"\\n\")\n",
"print(\"----------------------------------------------------------------\")\n",
"print(\"Tier 1\\n\")\n",
"sresults.range('A1').value = 'Tier 1 Results'\n",
"sresults.range('A2').value = 'Printer Location'\n",
"sresults.range('B2').value = 'Number of Printers'\n",
"\n",
"# Use this to help keep track of row number in Excel\n",
"row_num = 3\n",
"\n",
"# Create blank lists to store the Tier 1 facility locations\n",
"# and how many printers are at those locations in the optimal solution\n",
"sol_facilities_t1 = []\n",
"number_of_t1 = []\n",
"\n",
"# Write the locations and number of printers to screen\n",
"# and the Excel Results worksheet\n",
"print(\"Printer Location Number of Printers\")\n",
"print(\"---------------------- ------------------\")\n",
"for j in model.J:\n",
" for p in model.P:\n",
"\n",
" # Technically want != 0 on the line below,\n",
" # but we get \"noise\" in output with very small approximately\n",
" # 0 values. Use > 0.000001 instead to eliminate this issue.\n",
" if model.OPEN1[j,p].value > 0.000001:\n",
" print(f\"{j:22s} {p:0.0f}\")\n",
" sresults.range(f'A{row_num}').value = j\n",
" sresults.range(f\"B{row_num}\").value = p\n",
" row_num += 1\n",
"\n",
" # If there are more than 0 printers at a given location,\n",
" # append that location to sol_facilities_t1 and append\n",
" # the number of printers to number_of_t1.\n",
" if p > 0:\n",
" sol_facilities_t1.append(j)\n",
" number_of_t1.append(p)\n",
"\n",
"# Create a blank index list to store the indices of the locations\n",
"# that have more than 0 printers in the optimal solution\n",
"sol_fac_t1_idx = []\n",
"\n",
"# Loop through all the solution facilities that had more than\n",
"# 0 printers in the optimal solution\n",
"for i in sol_facilities_t1:\n",
" # Append the index of the customer location from the customers\n",
" # list to the sol_fac_t1_idx list\n",
" sol_fac_t1_idx.append(customers.index(i))\n",
"\n",
"# Create a blank index list to store the indices of the number of printers\n",
"# for the locations that have more than 0 printers in the optimal solution\n",
"number_of_t1_idx = []\n",
"\n",
"# Loop through all the solution printer numbers that had more than\n",
"# 0 printers in the optimal solution\n",
"for p in number_of_t1:\n",
" # Append the index of the printer number from the NumOfPrinters\n",
" # list to the number_of_t1_idx list\n",
" number_of_t1_idx.append(num_printers.index(p))\n",
"\n",
"# Create blank list to store the longitudes of the locations that have\n",
"# more than 0 printers in the optimal solution\n",
"sol_longs_t1 = []\n",
"\n",
"# Create blank list to store the latitudes of the locations that have\n",
"# more than 0 printers in the optimal solution\n",
"sol_lats_t1 = []\n",
"\n",
"# Loop through all the indices of facilities that have more than\n",
"# 0 printers in the optimal solution\n",
"for i in sol_fac_t1_idx:\n",
" # Append the longs and lats for those facilities to the lists\n",
" # sol_longs_t1 and sol_lats_t1\n",
" sol_longs_t1.append(longs[i])\n",
" sol_lats_t1.append(lats[i])\n",
"\n",
"# Add an extra blank row for Excel output\n",
"row_num += 1\n",
"\n",
"# Create column headers for the Excel spreadsheet Results worksheet\n",
"sresults.range(f'A{row_num}').value = 'Customer Location'\n",
"sresults.range(f'B{row_num}').value = 'Printer Location'\n",
"sresults.range(f'C{row_num}').value = 'Percent of Service'\n",
"\n",
"# Add an extra blank row for Excel output\n",
"row_num += 1\n",
"\n",
"# Determine which printer locations service which customer locations\n",
"# and write those pairings to the screen and Excel spreadsheet\n",
"print(\"\\n\")\n",
"print(\"Customer Location Printer Location Percent of Service\")\n",
"print(\"---------------------- ---------------------- ------------------\")\n",
"for i in model.I:\n",
" for j in model.J:\n",
" if model.SAT1[i,j].value > 0.000001:\n",
" print(f\"{i:22s} {j:22s} {model.SAT1[i,j].value:.3f}\")\n",
" sresults.range(f'A{row_num}').value = i\n",
" sresults.range(f\"B{row_num}\").value = j\n",
" sresults.range(f'C{row_num}').value = model.SAT1[i,j].value\n",
" row_num += 1\n",
"\n",
"# Add an extra blank row for Excel output\n",
"row_num += 1\n",
"\n",
"# Create column headers for the Excel spreadsheet Results worksheet\n",
"sresults.range(f'A{row_num}').value = 'Customer Location'\n",
"sresults.range(f'B{row_num}').value = 'Fraction Unfulfilled'\n",
"\n",
"# Add an extra blank row for Excel output\n",
"row_num += 1\n",
"\n",
"# Determine which customer locations have unfulfilled demand and\n",
"# what fraction is unfulfilled. Write these values to the screen\n",
"# and the Excel spreadsheet.\n",
"print(\"\\n\")\n",
"print(\"Customer Location Fraction Unfulfilled\")\n",
"print(\"---------------------- -------------------\")\n",
"\n",
"for i in model.I:\n",
" if model.FRACUN1[i].value > 0.000001:\n",
" print(f\"{i:22s} {model.FRACUN1[i].value:.3f}\")\n",
" sresults.range(f'A{row_num}').value = i\n",
" sresults.range(f'B{row_num}').value = model.FRACUN1[i].value\n",
" row_num += 1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Tier 2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The code in the cell below writes the same results as above, but for tier 2. The code is essentially the same, but with fewer comments. Run this code."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Tier 2\n",
"\n",
"Printer Location Number of Printers\n",
"---------------------- ------------------\n",
"New River, NC 19\n",
"Camp Pendleton, CA 4\n",
"Lakehurst, NJ 0\n",
"Yuma, AZ 0\n",
"Cherry Point, NC 0\n",
"Norfolk, VA 0\n",
"NAS North Island, CA 0\n",
"NAS Pax River, MD 0\n",
"MCAS Miramar, CA 0\n",
"Macguire AFB, NJ 0\n",
"NAS Pt. Mugu, CA 0\n",
"NAS Fallon, NV 0\n",
"NAS Whidbey Island, WA 0\n",
"NAWS China Lake, CA 0\n",
"Tinker AFB, OK 0\n",
"NAS Pensacola, FL 0\n",
"NASJRB Fort Worth, TX 0\n",
"MCAS Beaufort, SC 0\n",
"NAS Lemoore, CA 0\n",
"NASJRB New Orleans, LA 0\n",
"Stewart ANGB, NY 0\n",
"MCAF Quantico, VA 0\n",
"NAS Jacksonville, FL 0\n",
"\n",
"\n",
"Customer Location Printer Location Percent of Service\n",
"---------------------- ---------------------- ------------------\n",
"New River, NC New River, NC 1.000\n",
"Camp Pendleton, CA Camp Pendleton, CA 1.000\n",
"Lakehurst, NJ New River, NC 1.000\n",
"Cherry Point, NC New River, NC 1.000\n",
"Norfolk, VA New River, NC 1.000\n",
"NAS Pax River, MD New River, NC 1.000\n",
"MCAS Miramar, CA Camp Pendleton, CA 0.797\n",
"Macguire AFB, NJ New River, NC 1.000\n",
"NAS Pensacola, FL New River, NC 0.456\n",
"MCAS Beaufort, SC New River, NC 1.000\n",
"Stewart ANGB, NY New River, NC 1.000\n",
"MCAF Quantico, VA New River, NC 1.000\n",
"NAS Jacksonville, FL New River, NC 1.000\n",
"\n",
"\n",
"Customer Location Fraction Unfulfilled\n",
"---------------------- -------------------\n",
"Yuma, AZ 1.000\n",
"NAS North Island, CA 1.000\n",
"MCAS Miramar, CA 0.203\n",
"NAS Pt. Mugu, CA 1.000\n",
"NAS Fallon, NV 1.000\n",
"NAS Whidbey Island, WA 1.000\n",
"NAWS China Lake, CA 1.000\n",
"Tinker AFB, OK 1.000\n",
"NAS Pensacola, FL 0.544\n",
"NASJRB Fort Worth, TX 1.000\n",
"NAS Lemoore, CA 1.000\n",
"NASJRB New Orleans, LA 1.000\n"
]
}
],
"source": [
"print(\"Tier 2\\n\")\n",
"\n",
"sresults.range('E1').value = 'Tier 2 Results'\n",
"sresults.range('E2').value = 'Printer Location'\n",
"sresults.range('F2').value = 'Number of Printers'\n",
"\n",
"# Use this to help keep track of row number in Excel\n",
"row_num = 3\n",
"\n",
"sol_facilities_t2 = []\n",
"number_of_t2 = []\n",
"print(\"Printer Location Number of Printers\")\n",
"print(\"---------------------- ------------------\")\n",
"for j in model.J:\n",
" for p in model.P:\n",
" if model.OPEN2[j,p].value > 0.000001:\n",
" print(f\"{j:22s} {p:0.0f}\")\n",
" sresults.range(f'E{row_num}').value = j\n",
" sresults.range(f\"F{row_num}\").value = p\n",
" row_num += 1\n",
" if p > 0:\n",
" sol_facilities_t2.append(j)\n",
" number_of_t2.append(p)\n",
"\n",
"sol_fac_t2_idx = []\n",
"for i in sol_facilities_t2:\n",
" sol_fac_t2_idx.append(customers.index(i))\n",
"\n",
"number_of_t2_idx = []\n",
"for p in number_of_t2:\n",
" number_of_t2_idx.append(num_printers.index(p))\n",
"\n",
"sol_longs_t2 = []\n",
"sol_lats_t2 = []\n",
"for i in sol_fac_t2_idx:\n",
" sol_longs_t2.append(longs[i])\n",
" sol_lats_t2.append(lats[i])\n",
"\n",
"\n",
"# Add an extra blank row for Excel output\n",
"row_num += 1\n",
"\n",
"sresults.range(f'E{row_num}').value = 'Customer Location'\n",
"sresults.range(f'F{row_num}').value = 'Printer Location'\n",
"sresults.range(f'G{row_num}').value = 'Percent of Service'\n",
"\n",
"row_num += 1\n",
"\n",
"# Determine which printers service which locations\n",
"print(\"\\n\")\n",
"print(\"Customer Location Printer Location Percent of Service\")\n",
"print(\"---------------------- ---------------------- ------------------\")\n",
"for i in model.I:\n",
" for j in model.J:\n",
" if model.SAT2[i,j].value > 0.000001:\n",
" print(f\"{i:22s} {j:22s} {model.SAT2[i,j].value:.3f}\")\n",
" sresults.range(f'E{row_num}').value = i\n",
" sresults.range(f\"F{row_num}\").value = j\n",
" sresults.range(f'G{row_num}').value = model.SAT2[i,j].value\n",
" row_num += 1\n",
"\n",
"row_num += 1\n",
"\n",
"sresults.range(f'E{row_num}').value = 'Customer Location'\n",
"sresults.range(f'F{row_num}').value = 'Fraction Unfulfilled'\n",
"\n",
"row_num += 1\n",
"\n",
"# Determine locations with unfulfilled demand\n",
"print(\"\\n\")\n",
"print(\"Customer Location Fraction Unfulfilled\")\n",
"print(\"---------------------- -------------------\")\n",
"\n",
"for i in model.I:\n",
" if model.FRACUN2[i].value > 0.000001:\n",
" print(f\"{i:22s} {model.FRACUN2[i].value:.3f}\")\n",
" sresults.range(f'E{row_num}').value = i\n",
" sresults.range(f'F{row_num}').value = model.FRACUN2[i].value\n",
" row_num += 1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Displaying the results of the optimization graphically"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"While it is important to have the detailed information provided by the tabular results we generated above, it is often useful to display this same information graphically as well. This section describes how the information can be displayed graphically both to the screen and the Excel workbook."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We begin by constructing a dictionary to store the information we will need to plot any locations that have more than zero printers in the optimal solution. Look at the code below, including the in-line comments to make sure you understand what it is doing. Note how the data for the two different tiers are separated in the dictionary. Run this code."
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [],
"source": [
"# Initialize the blank dictionary\n",
"overall_coords = {}\n",
"\n",
"# Add the key, value pairs for tier 1 longitudes where > 0 printers will be located\n",
"overall_coords[('tier1', 'longs')] = sol_longs_t1\n",
"\n",
"# Add the key, value pairs for tier 1 latitudes where > 0 printers will be located\n",
"overall_coords[('tier1', 'lats')] = sol_lats_t1\n",
"\n",
"# Add the key, value pairs for tier 1 where number of printers > 0 \n",
"overall_coords[('tier1', 'NumberOfPrinters')] = number_of_t1_idx\n",
"\n",
"# Add the key, value pairs for tier 2 longitudes where > 0 printers will be located\n",
"overall_coords[('tier2', 'longs')] = sol_longs_t2\n",
"\n",
"# Add the key, value pairs for tier 2 latitudes where > 0 printers will be located\n",
"overall_coords[('tier2', 'lats')] = sol_lats_t2\n",
"\n",
"# Add the key, value pairs for tier 2 where number of printers > 0\n",
"overall_coords[('tier2', 'NumberOfPrinters')] = number_of_t2_idx"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
"
Problem 7
\n",
"\n",
"In the cell below, write code to print the contents of `overall_coords` to the output terminal. Look through the printed output to make sure you understand what information is stored in the dictionary `overall_coords`.\n",
"
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"run_control": {
"marked": true
},
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{('tier1', 'longs'): [-117.3205, -117.2044], ('tier1', 'lats'): [33.3178, 32.6978], ('tier1', 'NumberOfPrinters'): [27, 11], ('tier2', 'longs'): [-77.4097, -117.3205], ('tier2', 'lats'): [34.7574, 33.3178], ('tier2', 'NumberOfPrinters'): [19, 4]}\n"
]
}
],
"source": [
"# SOLUTION\n",
"print(overall_coords)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Graphical representation of tier 1 printers"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The code below produces a graphical representation of the optimal solution for tier 1 printers. Key features of the graphical output include the following:\n",
"\n",
" - A map of the continential U.S. with state borders, rivers, and lakes\n",
" - Customer locations plotted with red circles\n",
" - Printer locations plotted with blue pentagons that increase in size as the number of printers at the location increase\n",
" - A legend to indicate the meaning of the different color markers\n",
" - Lines with different thickness levels between cutomers and printers to indicate the level of satisfaction provided \n",
"\n",
"Look through the code below and read the in-line comments to help you understand what different portions of the code are doing. Run the code and look at the resulting output in Jupyter and the Map1 worksheet in Excel."
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {
"run_control": {
"marked": true
}
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"