MSBA7003 Quantitative Analysis Methods

HU, Xing
Associate Professor
HKU Business School

05 Linear Programming II

1
 2

Agenda

• Review of LP programming (1)

• LP applications for practical problems
• Advertising Model
• Robot Scheduling Model
• Portfolio Selection Model
• Transportation Model
• Production Planning Model
 3

Review of Linear Programming (1)

 4

Formulate a Linear Programming

• Define decision variables properly

• Properly define your objective function (linear objective function)

• Identify all constraints and correctly add them to the LP formulation

(linear constraints)
T = number of tables to be produced per week
C = number of chairs to be produced per week

Maximize profit = $70𝑇 + $50𝐶

s.t. 4𝑇 + 3𝐶 ≤ 240 ( carpentry hours )
2𝑇 + 1𝐶 ≤ 100 (painting and varnishing hours )
𝑇, 𝐶 ≥ 0
 5

How to Find LP Solutions

• Graphical solution method

• Isoprofit line solution method
• Corner point solution method

• Computer Software (e.g., Excel Solver, CPLEX solver)

 6

Graphic Solution Method

Maximize profit = $70𝑇 + $50𝐶

s.t. 4𝑇 + 3𝐶 ≤ 240 ( carpentry hours )
2𝑇 + 1𝐶 ≤ 100 (painting and varnishing hours )
𝑇, 𝐶 ≥ 0

Feasible region Isoprofit line

 7

Graphical Solution Method

Profit maximization = go up as much as possible.
At the same time, the isoprofit lines should contain points in
C the feasible region; otherwise, the profit is not achievable.

100 –
–
80 – Maximum Profit Line
Number of Chairs

–
60 – Optimal Solution Point
– (T = 30, C = 40)
40 –
– $4,100 = $70T + $50C
20 –
–
| | | | | | | | | | | |

0 20 40 60 80 100 T
Number of Tables
 8

Corner Point Solution Method

C
NUMBER OF NUMBER OF PROFIT
100 – TABLES (T) CHAIRS (C) = $70T + $50C
– 0 0 $0
(0, 80)
80 – 50 0 $3,500
Number of Chairs

– 0 80 $4,000
60 –
30 40 $4,100
– (30,40)
40 –
–
20 –
– (50, 0)
| | | | | | | | | | | |
(0, 0) 0 20 40 60 80 100 T
Number of Tables
 9

Slack and Surplus

Maximize profit = $70𝑇 + $50𝐶
s.t. 4𝑇 + 3𝐶 ≤ 240 ( carpentry hours )
2𝑇 + 1𝐶 ≤ 100 (painting and varnishing hours )
𝑇, 𝐶 ≥ 0

Optimal solution: T = 30, C = 40

Slack is the amount of a resource that is not used.

Surplus indicates the amount by which a greater-than-or-equal-to constraint

is exceeded.

If no slack or surplus, then a constraint is binding.

Both the carpentry and P&V constraints are binding (no slack). If we
require 𝐶 ≥ 20, then we have surplus for this constraint
 Sensitivity Analysis
10

Max $70𝑇 + $50𝐶

s.t. 4𝑇 + 3𝐶 ≤ 240 ( carpentry hours )
2𝑇 + 1𝐶 ≤ 100 (painting and varnishing hours )
𝑇, 𝐶 ≥ 0
 11

Excel Solver

• Step 2: Write the formulas for the objective function and the
constraints.
 12

Excel Solver: Sensitivity Report

Coefficient may change by these amounts and

Units Produced the optimal solution will not change.

Resource may
change by these
amounts and the
shadow price will
not change.

If we add one more unit to the resource, the

Hours Used
marginal impact on the optimal profit
 13

In-Class Exercise

• True or False?
• A new technology that saves one carpentry hour per chair will lead to the
optimal solution of producing fewer chairs.
• If the number of carpentry hours is increased, it is optimal to produce more
tables.
• If the profit of a chair is increased to $60, it is optimal to produce more
chairs.
 14

Agenda

• LP applications for practical problems

• Advertising Model
• Robot Scheduling Model
• Portfolio Selection Model
• Transportation Model
• Production Planning Model
 Advertising Channels for
15

Casinos
Car radio ads

Ads magazine

TV ads
 16

Media Selection

• The Win Big Gambling Club promotes gambling junkets to casinos

in the Bahamas.
• It has $8,000 per week to spend on advertising, which is to be
allocated among four promotional media (shown on the next
slide).
• Its goal is to reach the largest possible high-potential audience
(i.e., gamblers).
• It needs to place at least five radio spots per week.
• No more than $1,800 can be spent on radio advertising each
week.
 17

Media Selection

AUDIENCE NUMBER
REACHED COST PER MAX ADS TAKEN PER
MEDIUM PER AD AD ($) PER WEEK WEEK
TV spot (1 minute) 5,000 800 12 𝑋!

Daily newspaper (full-page ad) 8,500 925 5 𝑋"

Radio spot (30 sec., prime time) 2,400 290 25 𝑋#

Radio spot (1 minute, afternoon) 2,800 380 20 𝑋$

 18

Media Selection
 19

Media Selection
 20

Agenda

• LP applications for practical problems

• Advertising Model
• Robot Scheduling Model
• Portfolio Selection Model
• Transportation Model
• Production Planning Model
 21

Robot Scheduling

A hotel equipped with multi-task

robots requires different numbers of
robots on different days of the week.

Each robots must work five

consecutive days and take two days
off for maintenance and repair.
 22

Robot Scheduling
Day # Robots Required • Formulate LP to minimize
Monday 17
the number of robots that
must be used. Robots are
Tuesday 13 very expensive!
Wednesday 15

Thursday 19

Friday 14

Saturday 16

Sunday 11
 23

What is Wrong with this Formulation?

𝑋( : the number of robots working on day i
Minimize: 𝑋! + 𝑋" + 𝑋# + 𝑋$ + 𝑋% + 𝑋& + 𝑋'
• Constraints:
• Monday constraint: 𝑋! ≥ 17
• Tuesday constraint: 𝑋" ≥ 13
• Wednesday constraint: 𝑋# ≥ 15
• Thursday constraint: 𝑋$ ≥ 19
• Friday constraint: 𝑋% ≥ 14
• Saturday constraint: 𝑋& ≥ 16
• Sunday constraint: 𝑋' ≥ 11
• Non-negativity: 𝑋! , 𝑋" , 𝑋# , 𝑋$ , 𝑋% , 𝑋& , 𝑋' ≥ 0
 24

Robot Scheduling

• Decision variables:

• Objective:
 25

Robot Scheduling
𝑋( : the number of robots whose first working day is Day i

Req

17
13
15
19
14
16
11
 26

Robot Scheduling
 27

Robot Scheduling

Worker Scheduling Model

Decision Variables Monday Tuesday Wednesday Thursday Friday Saturday Sunday Total No. of FT Employees
No. of FT Employees: 6.333333333 3.333333333 2 7.333333333 0 3.333333333 0 22.33333

Constraints RHS
Monday 1 1 1 1 1 17 >= 17
Tuesday 1 1 1 1 1 13 >= 13
Wednesday 1 1 1 1 1 15 >= 15
Thursday 1 1 1 1 1 19 >= 19
Friday 1 1 1 1 1 19 >= 14
Saturday 1 1 1 1 1 16 >= 16
Sunday 1 1 1 1 1 12.66667 >= 11

• To get integer solutions

• Add integer constraints
• Round up the solutions to the nearest integers
 28

Agenda

• LP applications for practical problems

• Advertising Model
• Robot Scheduling Model
• Portfolio Selection Model
• Transportation Model
• Production Planning Model
 29

Portfolio Optimization
 30

Exercise: Portfolio Selection

• The Heinlein and Krampf Brokerage firm is instructed by a client to

invest $250,000, with guidelines:
• Municipal bonds should constitute at least 20% of the investment
• At least 40% of the investment should be placed in a combination of
electronic firms, aerospace firms, and drug manufacturers
• No more than 50% of the amount invested in municipal bonds should be
placed in a high-risk, high-yield nursing home stock
• The goal is to maximize the projected return.
Los Angeles Thompson United Happy Days
municipal Electronics, Aerospace Nursing
Investment bonds Inc. Corp. Palmer Drugs Homes
Projected
Return (%) 5.3 6.8 4.9 8.4 11.8
 31

Portfolio Selection

Question: How to set cells I9 and I10?

 32

Portfolio Selection: Sensitivity Report

Variable Cells
Final Reduced Objective Allowable Allowable
Cell Name Value Cost Coefficient Increase Decrease
$B$3 Amount LA Municipal Bonds 50000 0 0.053 0.014 0.406
$C$3 Amount Thompson Electronics, Inc. 0 -0.016 0.068 0.016 1E+30
$D$3 Amount United Aerospace Corp. 0 -0.035 0.049 0.035 1E+30
$E$3 Amount Palmer Drugs 175000 0 0.084 0.034 0.009333333
$F$3 Amount Happy Days Nursing Homes 25000 0 0.118 0.028 0.034

Constraints
Final Shadow Constraint Allowable Allowable
Cell Name Value Price R.H. Side Increase Decrease
$G$8 Total Amount Constraint 250000 0.0812 250000 1E+30 250000
$G$9 Municipal Bond Constraint 50000 -0.014 0 50000 50000
$G$10 E.A.D. Constraint 175000 0 0 75000 1E+30
$G$11 Nursing Home Constraint 25000 0.034 0 75000 25000

Question: What is the marginal return rate?

 33

Agenda

• LP applications for practical problems

• Advertising Model
• Robot Scheduling Model
• Portfolio Selection Model
• Transportation Model
• Production Planning Model
 34

An Order Placed at JD app

Order completed

Order delivered

Order out for delivery

Order arrived at HK
facility

Order shipped from JD

facility in Dongguan

Order confirmed

Order submitted

Transportation and logistics

 35

JD’s Self-Operated B2C Model

• After the supplier's goods are produced, they are often stored in
43 "Asia No. 1" large-scale intelligent logistics parks and about
1,400 warehouses operated by JD Logistics nationwide.
 36

JD’s Self-Operated B2C Model

• After consumers place an order, these goods are picked and then sent to
210 sorting centers, distributed to 7,200 terminal distribution stations
across the country by 18,000 trucks, and finally delivered to the hands of
hundreds of millions of consumers by more than 200,000 self-operated
delivery personnel.
 37

A Simple Example of Transportation

Problem
Minimum cost
Supply locations Demand locations
 38

Transportation Problem
• The Executive Furniture Corporation is faced with the transportation
problem and is trying to minimize the daily transportation cost. How
to optimize the shipping plan, while the demand must be satisfied?

Albuquerque Boston Cleveland Supply

(Market 1) (Market 2) (Market 3)

Des Moines 5 4 3 100

(Plant 1)

Evansville 8 4 3 300
(Plant 2)

Fort Lauderdale 9 7 5 300

(Plant 3)

Demand 300 200 200

Transportation Matrix
 39

Transportation Problem
• The Executive Furniture Corporation is faced with the transportation
problem and is trying to minimize the daily transportation cost. How
to optimize the shipping plan, while the demand must be satisfied?

Transportation network
Daily Supply Origin Destination Daily Demand

Des Moines $5 Albuquerque

100 300
(Plant 1)
$3 $4 (Market 1)

Evansville $8 Boston
300 $4 200
(Plant 2) $3 (Market 2)

Fort $9 $7 Cleveland
300 Lauderdale 200
$5 (Market 3)
(Plant 3)
Per-unit
shipping costs
 40

LP Formulation
 41

Transportation Problem
 42

Production Planning
 43

Production Planning

• During the next four months the SureStep Company must meet the
following demand (pairs of shoes) on time: 3000 in month 1; 5000 in
month 2; 2000 in month 3; and 1000 in month 4.
• At the beginning of month 1, 1500 pairs of shoes are on hand and the
company has 75 workers. Each worker can work up to 160 hours a
month before he or she receives overtime. A worker can work up to
20 hours of overtime a month and is paid $13 per hour for overtime.
It takes 4 hours of labor and $15 of raw material to produce a pair of
shoes.
• At the end of each month, each pair of unsold shoes incurs a holding
cost of $3.
• SureStep wants to find the minimum-cost solution that meets
forecasted/contracted demand on time.
 44

Production Planning

• There are two types of decisions to make:

• Monthly overtime
• Monthly production quantity

• Constraints:
• Monthly production capacity
• Monthly overtime limit
• Monthly demand satisfaction
• Monthly Production Capacity = (Number of workers available * 160 + Monthly overtime) / 4

• Monthly overtime <= Number of workers available * 20

 45

Production Planning

• The decisions in different months will affect each other through the
following inventory balance equation:
Ending inventory of last month
+ Production quantity of this month
– Demand of this month
= Ending inventory of this month

• Demand satisfaction means that: Ending inventory >= 0.

 46

Production Planning
 47

Production Planning(2)
Lost Sales
• Suppose demand satisfaction is not a hard constraint, but there will
be a penalty of $50 for each unit of unsatisfied demand. How to
minimize the total cost?

• Sales = min(Demand, Available Inventory)

• Lost Sales = Demand – Sales
• Ending Inventory = Beginning Inv. + Production – Sales

• However, min() function is not allow for linear programming!

• Introduce two constraints:
• Sales <= Demand
• Sales <= Available Inventory
 48

Exercise:
Production Planning(3)
• During the next four months the SureStep Company must meet the
following demand for pair of shoes on time: 3000 in month 1; 5000 in
month 2; 2000 in month 3; and 3800 in month 4.
• At the beginning of month 1, 1500 pairs of shoes are on hand and the
company has 100 workers.
• A worker is paid $1200 per month. Each worker can work up to 160
hours a month before he or she has to work overtime. A worker can
work up to 20 hours of overtime a month and is paid $12 per hour for
overtime. It takes four hours of labor and $15 of raw material to produce
a pair of shoes.
• At the beginning of each month, workers can be hired or fired. Each
hiring costs $1600 and each firing costs $2000.
• At the end of each month, each pair of unsold shoes incurs a holding
cost of $3.
 49

Exercise

• Reconsider the Transportation Problem.

• Suppose each source can supply more than its capacity by
outsourcing at price 𝑝( = 3, and demand can be unsatisfied with a
revenue loss of 𝑟) = 10 per unit. What is the optimal outsourcing and
transportation plan?