Motivation Business analytics Mathematical programming History Preview for this course

Operations Research I: Models & Applications

Course Overview

Ling-Chieh Kung

Department of Information Management

National Taiwan University

OR I: Course Overview 1 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Road map

I Motivation.
I Business analytics.
I Mathematical programming.
I History.
I Preview for this course.

OR I: Course Overview 2 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Resource allocation

I What is management?
I Someone said:
I “Management is the attainment of organizational goals in an effective
and efficient manner through planning, organizing, leading, and
controlling organizational resources.”1 (Daft, 2014, p. 7)
I What are resources in an organization?

1 Daft, R. L. (2014). Management.

OR I: Course Overview 3 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Resource allocation

I Setting an objective requires strategic thinking, experiences, vision, etc.

I “Resource allocation” uses Operations Research (OR).
I Names of similar subjects/ideas:
I Management science.
I Decision science.
I Optimization method/algorithm.
I Mathematical programming.
I Let’s see some examples.

OR I: Course Overview 4 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: job allocation

I Two people are going to
hold an event, and they
need to complete some
tasks.
I One task must be assigned ID task processing time (min)
to exactly one person; one 1 boiling boba 20
person can work on one 2 brewing milk tea 30
task at a time. 3 baking cookies 60
I How to assign the tasks so 4 designing poster 15
5 renting handcart 25
that they can complete all
tasks the fastest?
I What are the resources?
What is the objective?

OR I: Course Overview 5 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: project management

I n workers are going to complete

m jobs in a project.
I Some jobs must be processed
with precedence rules.
I Some jobs cannot be done by
certain workers.
I Some jobs can be split and
allocated to several workers.
I Some jobs require different
processing time if allocated to
different workers.
I How many days does it take to
complete this project?

OR I: Course Overview 6 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Industry applications

I Key decisions:
I How to deliver 6.5 millions items to more than 220 countries each day?
I In each region, where to build distribution hubs?
I In each distribution hub, how to classify and sort items?
I In each city, how to choose routes?
I What do you need?
I Well-designed information systems.
I Operations Research!

OR I: Course Overview 7 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Industry applications

I Key decisions:
I How to determine the cities to connect?
I How to schedule more than 2000 flights per day?
I How to assign crews to flights?
I How to reassign crews immediately when there is an emergency?
I What do you need?
I Well-designed information systems.
I Operations Research!

OR I: Course Overview 8 / 34
Motivation Business analytics Mathematical programming History Preview for this course

What is Operations Research?

I Operations Research (OR) is:

I the methodology to “allocate the available resources to the various
activities in a way that is most effective for the organization as a whole.”
I It is “applied to problems that concern how to conduct and coordinate
the operations (i.e., activities) within an organization.”2
I It aims to support decision making.
I Typical tools: intuitions, business senses, and experiences.
I And OR (and other quantitative tools)!
I By doing OR studies, we generate some suggestions to decision makers.

2 Both are quoted from the ninth edition of Introduction to Operations

Research by Hillier and Lieberman.

OR I: Course Overview 9 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Road map

I Motivation.
I Business analytics.
I Mathematical programming.
I History.
I Preview for this course.

OR I: Course Overview 10 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Business analytics

I Today, everybody
talks about business
analytics.
I Master of business
administration
(MBA) becomes
master of business
analytics (or master
of science in business
analytics).

OR I: Course Overview 11 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Operations research/mathematical programming

I A typical process of decision making with business analytics:

I Data analysis: Collect information and understand the problem.
I Operations research: Allocate resources and solve the problem.

OR I: Course Overview 12 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: multi-product inventory

I How to set the inventory levels of multiple products to maximize the

total expected profit?
I Suppose that there is only one product.
I Prevent understocking or overstocking.
I Data analysis: Estimate the random amount of demand during one
order cycle time.
I Operations research: According to the random amount of demand, find
the inventory level to maximize the expected profit.

OR I: Course Overview 13 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: multi-product inventory

I When we have multiple products:

I Demand substitution: “There is no more Coke. How about Pepsi?”
I Data analysis is difficult.
I Estimate the probability of demand substitution between A and B,
which is the probability for one to purchase B when A is sold out (or
purchase A when B is sold out).
I Operations research is also difficult.
I Given the substitution probabilities, find the best inventory levels of all
products.

OR I: Course Overview 14 / 34
Motivation Business analytics Mathematical programming History Preview for this course

The process of conducting an OR study

OR I: Course Overview 15 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Analytics is not everything

I Advantages:
I It describes a problem in a precise and concise way (among people and
between a person and a computer).
I It facilitates the use of computers to solve a problem whose model has
been formulated.
I Disadvantages:
I Some of the problems are hard to be formulated into mathematical
models.
I Some critical information is missing.
I The aim of this course: Through various examples, we want to let
people get an concrete idea about:
I What may be solved by operations research.
I What cannot be solved by operations research.

OR I: Course Overview 16 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Road map

I Motivation.
I Business analytics.
I Mathematical programming.
I History.
I Preview for this course.

OR I: Course Overview 17 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: knapsack problem

I You preparing for hiking. There

are some useful items, but your ID Item Weight Value
backpack (knapsack) can only 1 compass 0.5 6
carry 5 kilograms. 2 hatchet 1.5 5
I An item cannot be split: Each 3 matches 0.4 4
item should be either chosen or 4 tarpaulin 1 4
discarded. 5 telescope 1.1 3
6 cylinder 1.6 4
I Which items should you bring to
7 Rilakkuma 0.8 1
maximize the total value?

OR I: Course Overview 18 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Example: knapsack problem

I Choosing items
according to their
weight-value ratios
results in items 1, 2, 3, ID Item Weight Value Value / Weight
4, and 5. The total
1 compass 0.5 6 12
value is 22. 2 hatchet 1.5 5 3.33
I The optimal solution is 3 matches 0.4 4 10
to bring items 1, 2, 3, 4 tarpaulin 1 4 4
4, and 6. The total 5 telescope 1.1 3 2.73
6 cylinder 1.6 4 2.5
value is 23. 7 Rilakkuma 0.8 1 1.25
I How may we model or
formulate this
problem?

OR I: Course Overview 19 / 34
Motivation Business analytics Mathematical programming History Preview for this course

A mathematical programming model

I Decision variables: What may we determine?

1 if item i is chosen,
xi =
0 otherwise.

I Objective function: What do we want?

max 6x1 + 5x2 + 4x3 + 4x4 + 3x5 + 4x6 + x7 .

I Constraints: What are the limitations?

0.5x1 + 1.5x2 + 0.4x3 + x4 + 1.1x5 + 1.6x6 + 0.8x7 ≤ 5.

and
xi ∈ {0, 1} ∀i = 1, ..., 7.

OR I: Course Overview 20 / 34
Motivation Business analytics Mathematical programming History Preview for this course

A mathematical programming model

I Collectively, a complete formulation (model) is:

max 6x1 + 5x2 + 4x3 + 4x4 + 3x5 + 4x6 + x7

s.t. 0.5x1 + 1.5x2 + 0.4x3 + x4 + 1.1x5 + 1.6x6 + 0.8x7 ≤ 5
xi ∈ {0, 1} ∀i = 1, ..., 7.

OR I: Course Overview 21 / 34
Motivation Business analytics Mathematical programming History Preview for this course

A compact formulation (advanced)

I Let wi and vi be the weight and value of item i.

I Let n be the number of items and B be the maximum allowable weight.
I A compact (and more abstract) formulation is
n
X
max vi x i
i=1
Xn
s.t. wi xi ≤ B
i=1
xi ∈ {0, 1} ∀i = 1, ..., n.

OR I: Course Overview 22 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Application: order selection

I Let wi and vi be the processing time and gross profit of order i.

I Let n be the number of orders and B be the factory capacity (the
maximum total processing time that a factory has).
I The formulation is (still)
n
X
max vi x i
i=1
Xn
s.t. wi xi ≤ B
i=1
xi ∈ {0, 1} ∀i = 1, ..., n.

OR I: Course Overview 23 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Application: portfolio optimization

I Let wi and vi be the maximum possible investment amount and
expected return of stock i.
I Let n be the number of assets and B be the total budget.
I The formulation is
n
X
max vi x i
i=1
Xn
s.t. w i xi ≤ B
i=1
0 ≤ xi ≤ 1 ∀i = 1, ..., n.

I Note that investment decision is not all-or-nothing.

I xi ∈ {0, 1} becomes xi ∈ [0, 1] or 0 ≤ xi ≤ 1.

OR I: Course Overview 24 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Linear programming vs. integer programming

I This is a linear program: I This is an integer program:

n
X n
X
max vi x i max vi x i
i=1 i=1
Xn Xn
s.t. wi xi ≤ B s.t. wi xi ≤ B
i=1 i=1
0 ≤ xi ≤ 1 ∀i = 1, ..., n. xi ∈ {0, 1} ∀i = 1, ..., n.

I Various types of mathematical programs will be introduced and

compared in this course.

OR I: Course Overview 25 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Road map

I Motivation.
I Business analytics.
I Mathematical programming.
I History.
I Preview for this course.

OR I: Course Overview 26 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Origin of linear programming

I George Dantzig (1914-2005).
I Ph.D. in Statistics, University of California, Berkeley (1946).
I He solved two well-known difficult problems which were mistakenly
considered as homework assignment.
I During World War II, he served in the United States Air Force doing
route planning for military aircraft.
I Each plan is called a “program.”
I He found:
I Many problems are linear optimization problems.
I No one is able to systematically solve large-scale optimization problems.
I He invented the simplex method.
I The “first” effective solution for linear programming in the world.
I Still an important part in commercial software nowadays.

OR I: Course Overview 27 / 34
Motivation Business analytics Mathematical programming History Preview for this course

I Numbers of personnel required at an airport vary a lot in a day.

0–6 6–8 8–10 10–12 12–14 14–16 16–18 18–20 22–24
6 10 15 20 16 24 28 20 10
I How many people will you hire?
I Each person works for eight hours continuously.
I They may start their shifts at different time.
I Linear programming is used by United Airlines to reduce the number of
flight delays by 50% and save more than $5 million per year in 1992.3

3 Rakshit A., N. Krishnamurthy, and G. Yu. (1996) “System Operations

Advisor: A Real-Time Decision Support System for Managing Airline Operations

at United Airlines.” INFORMS Journal on Applied Analytics 26(2) 50-58.
OR I: Course Overview 28 / 34
Motivation Business analytics Mathematical programming History Preview for this course

I If you produce foods, what are important in getting an order from

restaurants and retailers?
I You have more than 8000 customers sharing your capacity and inventory.
I Customers ask “When may I get them?” and “How much may I get?”
I Once you promise a customer, you need to immediately update the
availability information that are needed elsewhere.
I Updating requires a lot of planning and calculations.
I Swift & Company increased its annual earning by more than 12 million
U.S. dollars in 2006 with a linear programming solution.4

4 Bixby A., B. Downs, and M. Self, (2006) “A Scheduling and

Capable-to-Promise Application for Swift & Company.” INFORMS Journal of

Applied Analytics 36(1) 69-86.
OR I: Course Overview 29 / 34
Motivation Business analytics Mathematical programming History Preview for this course

History of linear programming

I Linear programming has been widely used for government, business,

and organizations.
I Two linear programming scholars, Leonid Kantorovich and Tjalling
Koopmans, are awarded the 1975 Nobel Memorial Prize in Economic
Sciences.
I “For their contributions to the theory of optimum allocation of
resources.”
I How about Dantzig?
I At least they three took a photo together.5

5 Gass, S. I. and Assad, A. A. (2005). An Annotated Timeline of Operations

Research: An Informal History. Boston, MA: Springer.

OR I: Course Overview 30 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Road map

I Motivation.
I Business analytics.
I Mathematical programming.
I History.
I Preview for this course.

OR I: Course Overview 31 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Course structure

Week Topic Computer

1 Course overview
2 Linear programming V
3 Integer programming V
4 Nonlinear programming V
5 Case studies
6 Course summary and future directions

OR I: Course Overview 32 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Course contents

I In this course, we focus on models and applications.

I We talk about the basic principles.
I We formulate linear, integer, and nonlinear programs.
I We experience solving mathematical programs using MS Excel solver.
I We go though cases, applications, and business implications.
I We (hopefully) let you know when to use (and when not to use)
operations research.
I Prerequisites:
I High-school mathematics (mainly algebra).
I High-school English.

OR I: Course Overview 33 / 34
Motivation Business analytics Mathematical programming History Preview for this course

Let’s learn operations research together!

OR I: Course Overview 34 / 34