QT Theory Merged OCR

BIS 494: Topics in Information Systems
Introduction to Queuing
Theory
1
What is Queuing Theory?
• Mathematical analysis of queues and waiting times in

stochastic systems.
– Used extensively to analyze production and service processes
exhibiting random variability in market demand (arrival times) and
service times.
• Queues arise when the short term demand for service
exceeds the capacity
– Most often caused by random variation in service times and the
times between customer arrivals.
– If long term demand for service > capacity the queue will explode!
2
Why is Queuing Analysis Important?
• Capacity problems are very common in industry and one
of the main drivers of process redesign
– Need to balance the cost of increased capacity against the gains of
increased productivity and service
• Queuing and waiting time analysis is particularly important
in service systems
– Large costs of waiting and of lost sales due to waiting
Prototype Example – ER at County Hospital

• Patients arrive by ambulance or by their own accord
• One doctor is always on duty
• More and more patients seeks help  longer waiting times
 Question: Should another MD position be instated?
3
A Cost/Capacity Tradeoff Model
Total
cost
Cost
Cost of
service
Cost of waiting
Process capacity
4
Examples of Real World Queuing Systems?
• Commercial Queuing Systems

– Commercial organizations serving external customers
– Ex. Dentist, bank, ATM, gas stations, plumber, garage …
• Transportation service systems
– Vehicles are customers or servers
– Ex. Vehicles waiting at toll stations and traffic lights, trucks or ships
waiting to be loaded, taxi cabs, fire engines, elevators, buses …
• Business-internal service systems
– Customers receiving service are internal to the organization providing
the service
– Ex. Inspection stations, conveyor belts, computer support …
• Social service systems
– Ex. Judicial process, the ER at a hospital, waiting lists for organ
transplants or student dorm rooms … 5
Components of a Basic Queuing Process
Input Source The Queuing System
Served
Calling Jobs Service Jobs
Queue Mechanism
Population
leave the
system
Arrival Queue
Process Discipline
Service
Queue
Process
Configuration
6
Components of a Basic Queuing Process (II)
 The calling population

– The population from which customers/jobs originate
– The size can be finite or infinite (the latter is most
common)
– Can be homogeneous (only one type of customers/ jobs)
or heterogeneous (several different kinds of
customers/jobs)
 The Arrival Process
– Determines how, when and where customer/jobs arrive to
the system
– Important characteristic is the customers’/jobs’ inter-
arrival times
– To correctly specify the arrival process requires data
collection of interarrival times and statistical analysis.
7
Components of a Basic Queuing Process (III)
 The queue configuration

– Specifies the number of queues
• Single or multiple lines to a number of service stations
– Their location
– Their effect on customer behavior
• Balking (customers deciding not to join the queue if its too long)
• Reneging (customers leave the queue if they have waited too long
for the service)
• Jockeying (customers switching between the queues if they think
they get served faster by doing so)
– Their maximum size (# of jobs the queue can hold)

• Distinction between infinite and finite capacity
8
Example – Two Queue Configurations
Multiple Queues Single Queue

Servers Servers
9
Multiple v.s. Single Customer Queue
Configuration
Multiple Line Advantages Single Line Advantages
1. The service provided can 1. Guarantees fairness

be differentiated – FIFO applied to all arrivals
– Ex. Supermarket express lanes 2. No customer anxiety
2. Labor specialization possible regarding choice of queue
3. Customer has more flexibility 3. Avoids “cutting in” problems
4. Balking behavior may be 4. The most efficient set up for
deterred minimizing time in the queue
– Several medium-length lines are 5. Jockeying (line switching) is
less intimidating than one very avoided
long line
10
Components of a Basic Queuing Process (IV)
 The Service Mechanism

– Can involve one or several service facilities with one or several
parallel service channels (servers) - Specification is required
– The service provided by a server is characterized by its service time
• Specification is required and typically involves data gathering and
statistical analysis.
• Most analytical queuing models are based on the assumption of
exponentially distributed service times, with some generalizations.
 The queue discipline
– Specifies the order by which jobs in the queue are being served.
– Most commonly used principle is FIFO.
– Other rules are, for example, LIFO, SPT (Smallest Processing Time),
EDD (Earliest Due Date)
– Can entail prioritization based on customer type.
11
Mitigating Effects of Long Queues
1. Concealing the queue from arriving customers
– Ex. Restaurants divert people to the bar or use pagers, amusement parks
require people to buy tickets outside the park, banks broadcast news on
TV at various stations along the queue, casinos snake night club queues
through slot machine areas.
2. Use the customer as a resource
– Ex. Patient filling out medical history form while waiting for physician
3. Making the customer’s wait comfortable and distracting their
attention
– Ex. Complementary drinks at restaurants, computer games, internet
stations, food courts, shops, etc. at airports
4. Explain reason for the wait
5. Provide pessimistic estimates of the remaining wait time
– Wait seems shorter if a time estimate is given.
6. Be fair and open about the queuing disciplines used 12
Kinds (Types) of Queuing Model
1. Single Server
– Ex. Cheque deposition queue in a bank
2. Multiple Server
– Ex. Ticket Reservation Counter
Within these types, we have further classification
1. Finite queue length
- Garage mechanic has 10 slots for cars. The 11th car has to be retuned
back as there are no empty slots.
2. Infinite queue length
- Every person who comes joins the queue
13
A Commonly Seen Queuing Model (I)
The Queuing System
The Service Facility
C S = Server
The Queue C S
Customers (C)
CCC…C •
•
•
Customer =C C S
14
NOTATIONS FOR QUEUING MODELS
• Service times as well as inter-arrival times are assumed

independent and identically distributed
– If not otherwise specified
• Commonly used notation principle: A/B/C
– A = The inter-arrival time distribution
– B = The service time distribution
– C = The number of parallel servers
• Example: M/M/c
– Queuing system with exponentially distributed service and inter-arrival
times and c servers
15
NOTATIONS FOR QUEUING MODELS
16
Description of Queue Models
17
Notations for Various Models of Queues
18
The Exponential Distribution and Queuing
• The most commonly used queuing models are based on the

assumption of exponentially distributed service times and
interarrival times.
Definition: A stochastic (or random) variable Texp( ),

i.e., is exponentially distributed with parameter , if its
frequency function is:
e t when t  0
fT (t)  
0 when t  0
 The Cumulative Distribution Function is: FT ( t )  1  e t

 The mean = E[T] = 1/
 The Variance = Var[T] = 1/ 2 19
The Exponential Distribution
fT(t)

Probability density
t
Mean=
Time between arrivals
E[T]=1/
20
The Poisson Process
 The standard assumption in many queuing

models is that the arrival process is Poisson
Two equivalent definitions of the Poisson

Process
1. The times between arrivals are independent,
identically distributed and exponential
2. X(t) is a Poisson process with arrival rate .
21
Terminology and Notation
 The state of the system = the number of customers in the system

 Queue length = (The state of the system) – (number of
customers being served)
N(t) = Number of customers/jobs in the system at time t

Pn(t) = The probability that at time t, there are n customers/jobs in the system.
n = Average arrival intensity (= # arrivals per unit time) at n customers/jobs
in the system
n = Average service intensity for the system when there are n
customers/jobs in it. (Note, the total service intensity for all occupied
servers)
 = The utilization factor for the service facility. (= The expected fraction of
the time that the service facility is being used). Also called as the
Traffic Intensity. 22
Service Utilization Factor/Traffic Intensity
(ρ)
• Consider an M/M/1 queue with arrival rate =  and service
intensity = 
•  = Expected demand capacity per unit time
•  = Expected available capacity per unit time
Capacity Demand λ
 ρ 
Available Capacity μ
23
Queuing Theory Focus on Steady State
• Steady State condition

– Enough time has passed for the system state to be independent of the
initial state as well as the elapsed time
– The probability distribution of the state of the system remains the same
over time (is stationary).
• Transient condition
– Prevalent when a queuing system has recently begun operations
– The state of the system is greatly affected by the initial state and by the
time elapsed since operations started
– The probability distribution of the state of the system changes with time
With few exceptions Queuing Theory has focused on

analyzing steady state behavior
24
Transient and Steady State Conditions
• Illustration of transient and steady-state conditions
– N(t) = number of customers in the system at time t,
– E[N(t)] = represents the expected number of customers in the system.
30
Transient condition Steady State condition
25
Number of jobs in the system, N(t)
20
15 E[N(t)]
N(t)
10
0
0 5 10 15 20 25 30 35 40 45 50
time, t
25
Notation For Steady State Analysis
Pn = The probability that there are exactly n customers/jobs in the

system (in steady state, i.e., when t)
LS (or L) = Expected number of customers in the system (in
steady state). Also called as length of the system.
= ρ / (1 - ρ) = / ( - )
Lq = Expected number of customers in the queue (in steady state).
Also called as length of the queue.
= (LS - ρ) = 2/  ( - )
WS = Expected time a job (customer) spends in the system, it is
the sum of waiting time and the service time, given by,
= LS /  = 1/ ( - )
Wq= Expected time a job (customer) spends in the queue, it is the
waiting time of the customer in queue, given by,
= Lq /  = /  ( - ) 26
Contd..
P0 = The probability that 0 number of customers are there in
the system, given by,
= 1- ρ
Pn = The probability that n number of customers are there in
the system, given by,
= ρn . P0
27
Problem from A. Sharma
28
29
Problem
Mom-and-Pop’s Grocery Store has a small adjacent parking lot with three parking
spaces reserved for the store’s customers. During store hours, cars enter the lot and
use one of the spaces at a mean rate of 2 per hour. For n = 0, 1, 2, 3, the
probability Pn that exactly n spaces currently are being used is P0 = 0.1, P1 = 0.2,
P2 = 0.4, P3 = 0.3.
(a) Describe how this parking lot can be interpreted as being a queuing system. In
particular, identify the customers and the servers. What is the service being
provided? What constitutes a service time? What is the queue capacity?
(b) Determine the basic measures of performance—L, Lq, W, and Wq—for this
queueing system.
(c) Use the results from part (b) to determine the average length of time that a car
remains in a parking space.
30
Solution
Note
31
Problem
32
Solution
The utilization factor ρ represents the fraction
of time that the server is busy. The server is
busy except when there is nobody in the
system. P0 is the probability of having zero
customers in the system, so ρ = 1 - P0
33
Problem
34
Solution
35
Simulation – What is it?
• Experiment with a model mimicking the

real world system
– Ex. Flight simulation, wind tunnels, …
• In queuing situations computer based
simulation is used for analyzing and
evaluating complex stochastic systems
– Uncertain service and inter-arrival times
36
Simulation – Why use it?
• Cheaper and less risky than experimenting

with the actual system.
• Stimulates creativity since it is easy to test
the effect of new ideas
• A powerful complement to the traditional
symbolical and analytical tools
• Fun tool to work with!
37
Simulation v.s. Symbolic & Analytical Tools
 Strengths
+ Provides a quantitative measure
+ Flexible – can handle any kind of complex system or statistical
interdependencies
+ Capable of finding inefficiencies otherwise not detected until the
system is in operation
 Weaknesses
– Can take a long time to build
 Usually requires a substantial amount of data gathering
– Easy to misrepresent reality and draw faulty conclusions
– Generally not suitable for optimizing system parameters
 A simulation model is primarily descriptive while an

optimization model is by nature prescriptive
38
Modern Simulation Software Packages are
Breaking Compromises
• Graphical interfaces
 Achieves the descriptive benefits of
symbolic tools like flow charts
• Optimization Engines
 Enables efficient automated search
for best parameter values
39
Building a Simulation Model
• General Principles
– The system is broken down into suitable components or entities
– The entities are modeled separately and are then connected to a
model describing the overall system
 A bottom-up approach!
• The basic principles apply to all types of simulation models
– Static or Dynamic
– Deterministic or Stochastic
– Discrete or continuous
• In OM situations computer based Stochastic Discrete Event
Simulation (e.g. in Extend) is the natural choice
– Focuses on events affecting the state of the system and skips all
intervals in between
40
Steps in a BPD Simulation Project
1. Problem formulation Phase 1
Problem Definition
2. Set objectives and overall project plan
4. Data Collection
Phase 2
3. Model conceptualization
Model Building
5. Model Translation
No Phase 3
6. Verified
Yes Experimentation
No No
7. Validated
Yes
8. Experimental Design
Phase 4
Implementation
9. Model runs and analysis
Yes No 11. Documentation, reporting and

10. More runs 41
implementation
Model Verification and Validation
• Verification (efficiency)
– Is the model correctly built/programmed?
– Is it doing what it is intended to do?
• Validation (effectiveness)
– Is the right model built?
– Does the model adequately describe the reality you want to model?
– Does the involved decision makers trust the model?
 Two of the most important and most challenging issues in

performing a simulation study
42
Model Verification Methods
• Find alternative ways of describing/evaluating the system

and compare the results
– Simplification enables testing of special cases with predictable
outcomes
 Removing variability to make the model deterministic
 Removing multiple job types, running the model with one job type at
a time
 Reducing labor pool sizes to one worker
• Build the model in stages/modules and incrementally test
each module
– Uncouple interacting sub-processes and run them separately
– Test the model after each new feature that is added
– Simple animation is often a good first step to see if things are
working as intended
43
Validation - an Iterative Calibration Process
The Real System

Conceptual
validation
Conceptual Model
Calibration and 1. Assumptions on system components
Validation 2. Structural assumptions which define the
interactions between system components
3. Input parameters and data assumptions
Model
verification
Operational Model
(Computerized representation)
44
Example – Simulation of a M/M/1 Queue
• Assume a small branch office of a local bank with only
one teller.
• Empirical data gathering indicates that inter-arrival and
service times are exponentially distributed.
– The average arrival rate =  = 5 customers per hour
– The average service rate =  = 6 customers per hour
• Using our knowledge of queuing theory we obtain
–  = the server utilization = 5/6  0.83
– Lq = the average number of people waiting in line
– Wq = the average time spent waiting in line
Lq = 0.832/(1-0.83)  4.2 Wq = Lq/   4.2/5  0.83

• How do we go about simulating this system?
– How do the simulation results match the analytical ones?
45
QT
(Quantitative Techniques):
An Introduction
1
Introduction
• Operations Research is an Art and Science
• It had its early roots in World War II and is

flourishing in business and industry with the aid
of computer.
• Primary application areas of Operations

Research include forecasting, production
scheduling, inventory control, capital budgeting
and transportation. 2
What is Operations Research?
Operations
The activities carried out in an organization.
Research
The process of observation and testing
characterized by the scientific method. Situation,
problem statement, model construction, validation,
experimentation, solutions.
Operations Research is a quantitative approach to

decision making based on the scientific method of
problem solving. 3
What is Operations Research?
• Operations Research is the scientific
approach to execute decision making, which
consists of:
– The art of mathematical modeling of
complex situations
– The science of the development of solution
techniques used to solve these models
– The ability to effectively communicate the
results to the decision maker
4
What Do We do
1. Provide rational bases for decision making by
seeking to understand and structure complex
situations and to use this understanding to
predict system behavior and improve system
performance.
2. Much of this work is done using analytical and

numerical techniques to develop and manipulate
mathematical and computer models of
organizational systems composed of people,
machines, and procedures.
5
Terminology
• The British/Europeans refer to “Operational Research", the
Americans to “Operations Research" - but both are often
shortened to just "OR".
• Another term used for this field is “Management Science"

("MS"). In U.S. OR and MS are combined together to form
"OR/MS" or "ORMS".
• Yet other terms sometimes used are “Industrial Engineering"

("IE") and “Decision Science" ("DS").
6
Operations Research Models
Deterministic Models Stochastic Models

• Linear Programming • Discrete-Time Markov Chains
• Network Optimization • Continuous-Time Markov Chains
• Integer Programming • Queuing Theory (waiting lines)
• Nonlinear Programming • Decision Analysis
• Inventory Models Game Theory
Inventory models
Simulation
7
Deterministic vs. Stochastic Models
Deterministic models
assume all data are known with certainty
Stochastic models
explicitly represent uncertain data via
random variables or stochastic processes.
Deterministic models involve optimization
Stochastic models
characterize / estimate system performance.
8
Success Stories of OR
9
Application Areas
• Strategic planning
• Supply chain management
• Pricing and revenue management
• Logistics and site location
• Optimization
• Marketing research
10
Applications Areas (cont.)
• Scheduling
• Portfolio management
• Inventory analysis
• Forecasting
• Sales analysis
• Auctioning
• Risk analysis
11
A Short List of Successful Stories (1)
• Air New Zealand
– Air New Zealand Masters the Art of Crew Scheduling
• AT&T Network
– Delivering Rapid Restoration Capacity for the AT&T Network
• Bank Hapoalim
– Bank Hapoalim Offers Investment Decision Support for Individual Customers
• British Telecommunications
– Dynamic Workforce Scheduling for British Telecommunications
• Canadian Pacific Railway
– Perfecting the Scheduled Railroad at Canadian Pacific Railway
• Continental Airlines
– Faster Crew Recovery at Continental Airlines
• FAA
– Collaborative Decision Making Improves the FAA Ground-Delay Program
12
• Ford Motor Company
– Optimizing Prototype Vehicle Testing at Ford Motor Company
• General Motors
– Creating a New Business Model for OnStar at General Motors
• IBM Microelectronics
– Matching Assets to Supply Chain Demand at IBM Microelectronics
• IBM Personal Systems Group
– Extending Enterprise Supply Chain Management at IBM Personal Systems
Group
• Jan de Wit Company
– Optimizing Production Planning and Trade at Jan de Wit Company
• Jeppesen Sanderson
– Improving Performance and Flexibility at Jeppesen Sanderson
13
• Mars
– Online Procurement Auctions Benefit Mars and Its Suppliers
• Menlo Worldwide Forwarding
– Turning Network Routing into Advantage for Menlo Forwarding
• Merrill Lynch
– Seizing Marketplace Initiative with Merrill Lynch Integrated Choice
• NBC
– Increasing Advertising Revenues and Productivity at NBC
• PSA Peugeot Citroen
– Speeding Car Body Production at PSA Peugeot Citroen
• Rhenania
– Rhenania Optimizes Its Mail-Order Business with Dynamic Multilevel
Modeling
• Samsung
– Samsung Cuts Manufacturing Cycle Time and Inventory to Compete
14
• Spicer
– Spicer Improves Its Lead-Time and Scheduling Performance
• Syngenta
– Managing the Seed-Corn Supply Chain at Syngenta
• Towers Perrin
– Towers Perrin Improves Investment Decision Making
• U.S. Army
– Reinventing U.S. Army Recruiting
• U.S. Department of Energy
– Handling Nuclear Weapons for the U.S. Department of Energy
• UPS
– More Efficient Planning and Delivery at UPS
• Visteon
– Decision Support Wins Visteon More Production for Less
15
The Duckworth–Lewis (D/L)
Method
• It is a mathematical formulation designed to calculate the
target score for the team batting second in a limited overs
cricket match interrupted by weather or other
circumstances.
• It is generally accepted to be the most accurate method of
setting a target score.
• The D/L method was devised by two English
statisticians, Frank Duckworth and Tony Lewis.
• After their retirements Professor Steven Stern became
the custodian of the method.
• In November 2014, it was renamed the Duckworth– 16
Lewis–Stern method (or D/L/S method).

Linear Programming:
An Essential Optimization
Technique
1
Linear Programming
 Has powerful capabilities
 Enables businesses to reduce costs, improve
profitability, use resources effectively, reduce
risks and provide benefits in many other key
dimensions
 Optimization can automate decision processes
 Improves speed of responses
 Allow managers to focus their attention on
critical uncertainties rather than routine matters.
2
Background
 The seeds of linear programming were sown during World
War II when the military supplies and personnel had to be
moved efficiently.
 Although linear programming problems can be very

complicated (being real-world problems), the principles can be
understood by starting with simple algebra.
 Two-dimensional linear programming can be solved graphically
 Problems with more than two variables (as is the case for most
real world problems) can be solved by using a technique called
the simplex method 3
LP Applications
 Scheduling logistics to minimize total distance
traveled
 Scheduling tellers at banks so that needs are met
during each hour of the day while minimizing
the total cost of labor
 Developing a production schedule that will
satisfy future demands for a firm’s product and
at the same time minimize total production and
inventory costs
 Allocating space for a tenant mix in a new
shopping mall so as to maximize revenues to the
leasing company
LP Applications (Contd..)
 Selecting the product mix in a factory to make

best use of machine and labor-hours available
while maximizing the firm’s profit
 Picking blends of raw materials in feed mills to
produce finished feed combinations at minimum
costs
 Determining the distribution system that will
minimize total shipping cost
Additional LP Applications
 Shelf Space Allocation
 Voting on issues
 Assembly line balancing
 Pollution control
 Traffic light control
 Levelling the terrain for a new highway
 Military planning
 Water quality management
 Loading structure
 Allocation of Aircraft to routes
6
Linear Programming
 Linear programming has nothing to do with computer

programming.
 The use of the word “programming” here means

“choosing a course of action.”
 Linear programming involves choosing a course of

action when the mathematical model of the problem
contains only linear functions.
7
Linear Programming (LP)
Problem
 The maximization or minimization of some quantity is
the objective in all linear programming problems.
 All LP problems have constraints that limit the degree
to which the objective can be pursued.
 A feasible solution satisfies all the problem's
constraints.
 An optimal solution is a feasible solution that results in
the largest possible objective function value when
maximizing (or smallest when minimizing).
 A graphical solution method can be used to solve a
linear program with two variables. 8
Linear Programming (LP)
Problem
 If both the objective function and the constraints
are linear, the problem is referred to as a linear
programming problem.
 Linear functions are functions in which each
variable appears in a separate term raised to the
first power and is multiplied by a constant (which
could be 0).
 Linear constraints are linear functions that are
restricted to be "less than or equal to", "equal to",
or "greater than or equal to" a constant. 9
Problem Formulation
 Problem formulation or modeling is the process
of translating a verbal statement of a problem
into a mathematical statement.
 Formulating models is an art that can only be
mastered with practice and experience.
 Every LP problem has some unique features,
but most problems also have common features.
 General guidelines for LP model formulation are
illustrated on the slides that follow.
10
Guidelines for Model
Formulation
 Understand the problem thoroughly.
 Describe the objective.
 Describe each constraint.
 Define the decision variables.
 Write the objective in terms of the decision
variables.
 Write the constraints in terms of the decision
variables.
11
Example 1: A Simple
Maximization Problem
Objective
Max 5x1 + 7x2 Function
s.t. x1 < 6
“Regular”
2x1 + 3x2 < 19
Constraints
x1 + x2 < 8
Non-negativity
x1 > 0 and x2 > 0
Constraints
12
Example 1: Graphical Solution
 First Constraint Graphed
x2
8
7 x1 = 6
6 Shaded region
5 contains all
feasible points
4
for this constraint
3
2
(6, 0)
1
x1
1 2 3 4 5 6 7 8 9 10
13
 Second Constraint Graphed
x2
8 (0, 6 1/3)
7
6
5
2x1 + 3x2 = 19
4
Shaded
3
region contains
2 all feasible points (9 1/2, 0)
1 for this constraint
x1
1 2 3 4 5 6 7 8 9 10
14
 Third Constraint Graphed
x2
(0, 8)
8
7
6 x1 + x2 = 8
5
4
Shaded
3
region contains
2 all feasible points
1 for this constraint (8, 0)
x1
1 2 3 4 5 6 7 8 9 10
15
 Combined-Constraint Graph Showing Feasible
Region x2
8
x1 + x2 = 8
7
6 x1 = 6
5
4
3
Feasible 2x1 + 3x2 = 19
2
Region
1
x1
1 2 3 4 5 6 7 8 9 10
16
 Objective Function Line
x2
8
7
(0, 5)
6 Objective Function
5 5x1 + 7x2 = 35
4
3
2
(7, 0)
1
x1
1 2 3 4 5 6 7 8 9 10
17
Example 1: Graphical
Solution
 Selected Objective Function Lines
x2
8
7
5x1 + 7x2 = 35
6
5 5x1 + 7x2 = 39
4
3 5x1 + 7x2 = 42
2
1
x1
1 2 3 4 5 6 7 8 9 10 18
 Optimal Solution
x2
Maximum
Objective Function Line
8
5x1 + 7x2 = 46
7
6 Optimal Solution
(x1 = 5, x2 = 3)
5
4
3
2
1
x1
1 2 3 4 5 6 7 8 9 10
19
Summary of the Graphical
Solution Procedure for
Maximization Problems
 Prepare a graph of the feasible solutions for each of
the constraints.
 Determine the feasible region that satisfies all the
constraints simultaneously.
 Draw an objective function line.
 Move parallel objective function lines toward larger
objective function values without entirely leaving the
feasible region.
 Any feasible solution on the objective function line
with the largest value is an optimal solution. 20
Extreme Points and the
Optimal Solution
 The corners or vertices of the feasible region are
referred to as the extreme points.
 An optimal solution to an LP problem can be
found at an extreme point of the feasible region.
 When looking for the optimal solution, you do
not have to evaluate all feasible solution points.
 You have to consider only the extreme points of
the feasible region.
21
Example 1: Extreme Points
x2
8
7
5 (0, 6 1/3)
6
5
4
4 (5, 3)
3
Feasible 3 (6, 2)
2 Region
1 1 (0, 0) 2 (6, 0)
x1
1 2 3 4 5 6 7 8 9 10
22
Example 2: Minimization Problem
Min 5x1 + 2x2
s.t. 2x1 + 5x2 > 10

4x1 - x2 > 12
x1 + x2 > 4
x1, x2 > 0
23
 Graph the Constraints
Constraint 1: When x1 = 0, then x2 = 2; when x2 = 0,
then x1 = 5. Connect (5,0) and (0,2). The ">" side is
above this line.
Constraint 2: When x2 = 0, then x1 = 3. But setting x1 to 0
will yield x2 = -12, which is not on the graph. Thus, to get a
second point on this line, set x1 to any number larger than 3
and solve for x2: when x1 = 5, then x2 = 8. Connect (3,0) and
(5,8). The ">" side is to the right.
Constraint 3: When x1 = 0, then x2 = 4; when x2 = 0,
then x1 = 4. Connect (4,0) and (0,4). The ">" side is
above this line.
24
 Constraints Graphed
x2
6
Feasible Region
5
4x1 - x2 > 12
4
x1 + x2 > 4
3
2 2x1 + 5x2 > 10
x1
1 2 3 4 5 6
25
 Graph the Objective Function
Set the objective function equal to an arbitrary
constant (say 20) and graph it. For 5x1 + 2x2 = 20,
when x1 = 0, then x2 = 10; when x2= 0, then x1 = 4.
Connect (4,0) and (0,10).
 Move the Objective Function Line Toward Optimality

Move it in the direction which lowers its value
(down), since we are minimizing, until it touches the
last point of the feasible region, determined by the last
two constraints. 26
 Objective Function Graphed
x2 Min 5x1 + 2x2
6
5
4x1 - x2 > 12
4
x 1 + x2 > 4
3
2 2x1 + 5x2 > 10
x1
1 2 3 4 5 6 27
 Solve for the Extreme Point at the Intersection of the
Two Binding Constraints
4x1 - x2 = 12
x1+ x2 = 4
Adding these two equations gives:
5x1 = 16 or x1 = 16/5
Substituting this into x1 + x2 = 4 gives: x2 = 4/5
 Solve for the Optimal Value of the Objective Function
5x1 + 2x2 = 5(16/5) + 2(4/5) = 88/5

28
 Optimal Solution
x2
6
4x1 - x2 > 12
5
x1 + x2 > 4
4
Optimal Solution:
3
x1 = 16/5, x2 = 4/5,
2 5x1 + 2x2 = 17.6
1 2x1 + 5x2 > 10
x1
1 2 3 4 5 6 29
Summary of the Graphical
Solution Procedure
for Minimization Problems
 Prepare a graph of the feasible solutions for each of
the constraints.
 Determine the feasible region that satisfies all the
constraints simultaneously.
 Draw an objective function line.
 Move parallel objective function lines toward smaller
objective function values without entirely leaving the
feasible region.
 Any feasible solution on the objective function line
with the smallest value is an optimal solution.
30
Special Cases
 Infeasibility
• No solution to the LP problem satisfies all the
constraints, including the non-negativity
conditions.
• Graphically, this means a feasible region does not
exist.
• Causes include:
 A formulation error has been made.
 Management’s expectations are too high.
 Too many restrictions have been placed on the
problem (i.e. the problem is over-constrained).
31
Example: Infeasible Problem
 Consider the following LP problem.
Max 2x1 + 6x2
s.t. 4x1 + 3x2 < 12

2x1 + x2 > 8
x1, x2 > 0
32
Example: Infeasible Problem
 There are no points that satisfy both constraints,
so there is no feasible region (and no feasible
solution). x2
10
2x1 + x2 > 8
8
6
4x1 + 3x2 < 12
4
x1
2 4 6 8 10
33
Special Cases
 Unbounded
• The solution to a maximization LP problem is
unbounded if the value of the solution may be
made indefinitely large without violating any of
the constraints.
• For real problems, this is the result of improper
formulation. (Quite likely, a constraint has been
inadvertently omitted.)
34
Example: Unbounded Solution
 Consider the following LP problem.
Max 4x1 + 5x2
s.t. x1 + x2 > 5
3x1 + x2 > 8
x1, x2 > 0
35
Example: Unbounded Solution
 The feasible region is unbounded and the objective
function line can be moved outward from the origin
without bound, infinitely increasing the objective
function. x2
10
3x1 + x2 > 8
8
4
x1 + x2 > 5
2
x1
2 4 6 8 10 36
Problem: Production of wooden
tables and chairs
 A company uses wood and labor to produce tables
and chairs. The unit profit for tables is $6 and unit
profit for chairs is $8. There are 300 board feet (bf)
of wood available and 110 hours of labor available.
It takes 30 bf and 5 hours to make a table and 20 bf
and 10 hours to make a chair.
Formulate a LPP for the maximization of the
company’s profit and solve it graphically.
37
Information for the wooden
tables and chairs
38
LPP
39
40
41
42
43
The Transportation Problem
1
Copyright © 2003 Brooks/Cole, a division of Thomson Learning, Inc.
The Transportation Problem
• The problem of finding the minimum-cost

distribution of a given commodity
from a group of supply centers (sources) i=1,…,m
to a group of receiving centers (destinations) j=1,…,n
• Each source has a certain supply (si)
• Each destination has a certain demand (dj)
• The cost of shipping from a source to a destination is
directly proportional to the number of units shipped
2
Simple Network Representation
Sources Destinations
Supply s1 1 Demand d1
1
Supply s2 2
2 Demand d2
…
…
xij
n Demand dn
Supply sm m
Costs cij
3
Description
A transportation problem basically deals with the
problem, which aims to find the best way to fulfill
the demand of n demand points using the capacities
of m supply points. While trying to find the best way,
generally a variable cost of shipping the product
from one supply point to a demand point or a similar
constraint should be taken into consideration.
4
Description (Contd..)
• A transportation problem is a special kind of LPP in
which goods are transported from one set of
sources to one set of destinations subject to the
supplies and demands of the respective sources and
destinations such that the total cost of
transportation is minimized.
5
Formulating Transportation
Problems
Example 1: Powerco has three electric power
plants that supply the electric needs of four cities.
•The associated supply of each plant and demand
of each city is given in the table 1.
•The cost of sending 1 million kwh of electricity
from a plant to a city depends on the distance the
electricity must travel.
6
Transportation tableau
A transportation problem is specified by the

supply, the demand, and the shipping costs. So the
relevant data can be summarized in a
transportation tableau. The transportation tableau
implicitly expresses the supply and demand
constraints and the shipping cost between each
demand and supply point.
7
Table 1. Shipping costs, Supply, and Demand
for Powerco Example
From To
City 1 City 2 City 3 City 4 Supply
(Million kwh)
Plant 1 $8 $6 $10 $9 35
Plant 2 $9 $12 $13 $7 50
Plant 3 $14 $9 $16 $5 40
Demand 45 20 30 30
(Million kwh)
Transportation Tableau
8
Solution
1. Decision Variable:
Since we have to determine how much
electricity is sent from each plant to each city;
Xij = Amount of electricity produced at plant i

and sent to city j
X14 = Amount of electricity produced at plant 1

and sent to city 4
9
2. Objective function
Since we want to minimize the total cost of shipping

from plants to cities;
Minimize Z = 8X11+6X12+10X13+9X14
+9X21+12X22+13X23+7X24
+14X31+9X32+16X33+5X34
10
3. Supply Constraints
Since each supply point has a limited production
capacity;
X11+X12+X13+X14 <= 35
X21+X22+X23+X24 <= 50
X31+X32+X33+X34 <= 40
11
4. Demand Constraints
Since each supply point has a limited production
capacity;
X11+X21+X31 >= 45
X12+X22+X32 >= 20
X13+X23+X33 >= 30
X14+X24+X34 >= 30
12
5. Sign Constraints
Since a negative amount of electricity can not be
shipped all Xij’s must be non negative;
Xij >= 0 (i= 1,2,3; j= 1,2,3,4)
13
LP Formulation of Powerco’s Problem
Min Z = 8X11+6X12+10X13+9X14+9X21+12X22+13X23+7X24
+14X31+9X32+16X33+5X34
S.T.: X11+X12+X13+X14 <= 35 (Supply Constraints)

X21+X22+X23+X24 <= 50
X31+X32+X33+X34 <= 40
X11+X21+X31 >= 45 (Demand Constraints)
X12+X22+X32 >= 20
X13+X23+X33 >= 30
X14+X24+X34 >= 30
Xij >= 0 (i= 1,2,3; j= 1,2,3,4)
14
General Description of a Transportation
Problem
1. A set of m supply points from which a good is
shipped. Supply point i can supply at most si
units.
2. A set of n demand points to which the good is
shipped. Demand point j must receive at least di
units of the shipped good.
3. Each unit produced at supply point i and
shipped to demand point j incurs a variable cost
of cij.
15
Xij = number of units shipped from supply point i to
demand point j
i m j n
min  cijXij
i 1 j 1
j n
s.t. Xij  si (i  1,2,..., m)
j 1
i m
X
i 1
ij  dj ( j  1,2,..., n)
Xij  0(i  1,2,..., m; j  1,2,..., n)
16
Balanced Transportation Problem
If Total supply equals to total demand, the
problem is said to be a balanced
transportation problem:
i m j n
s  d
i 1
i
j 1
j
17
Balancing a TP if total supply exceeds total
demand
If total supply exceeds total demand, we

can balance the problem by adding
dummy demand point. Since shipments to
the dummy demand point are not real,
they are assigned a cost of zero.
18
Balancing a Transportation Problem if Total
Supply is Less than Total Demand
If a transportation problem has a total supply that

is strictly less than total demand the problem has
no feasible solution. There is no doubt that in such
a case one or more of the demand will be left
unmet. Generally in such situations a penalty cost is
often associated with unmet demand and as one
can guess this time the total penalty cost is desired
to be minimum
19
Solution of a Transportation Problem
It is always solved in two phases:
1. Finding the initial basic feasible solution
2. Optimization of the initial basic feasible solution which
is obtained in phase I
Methods to find the Basic Feasible Solution for a balanced

TP
There are three basic methods:
1. Northwest Corner Method
2. Minimum Cost Method
3. Vogel’s Approximation Method
20
1. Northwest Corner Method
To find the Basic Feasible Solution by the NWC

method:
Begin in the upper left (northwest) corner of the

transportation tableau and set x11 as large as possible
(here the limitations for setting x11 to a larger number,
will be the demand of demand point 1 and the supply
of supply point 1. Your x11 value can not be greater
than minimum of this 2 values).
21
According to the explanations in the previous slide
we can set x11=3 (meaning demand of demand
point 1 is satisfied by supply point 1).
3 5 2 3
3 2
X 5 2 3
22
After we check the east and south cells, we saw that we
can go east (meaning supply point 1 still has capacity to
fulfill some demand).
3 2 X
X 3 2 3
3 2 X
3 3
X X 2 3
23
After applying the same procedure, we saw that we can
go south this time (meaning demand point 2 needs more
supply by supply point 2).
3 2 X
3 2 1
X X X 3
3 2 X
3 2 1 X
24 X X X 2
Finally, we will have the following bfs, which is:
x11=3, x12=2, x22=3, x23=2, x24=1, x34=2
3 2 X
3 2 1 X
2 X
X X X X
25
2. Minimum Cost Method
The Northwest Corner Method dos not utilize
shipping costs. It can yield an initial bfs easily but the
total shipping cost may be very high. The minimum
cost method uses shipping costs in order come up
with a bfs that has a lower cost. To begin the
minimum cost method, first we find the decision
variable with the smallest shipping cost (Xij). Then
assign Xij its largest possible value, which is the
minimum of si and dj
26
After that, as in the Northwest Corner Method we
should cross out row i and column j and reduce the
supply or demand of the noncrossed-out row or
column by the value of Xij. Then we will choose the
cell with the minimum cost of shipping from the cells
that do not lie in a crossed-out row or column and we
will repeat the procedure.
27
An example for Minimum Cost Method
Step 1: Select the cell with minimum cost.
2 3 5 6
5
2 1 3 5
10
3 8 4 6
15
12 8 4 6
28
Step 2: Cross-out column 2
2 3 5 6
5
2 1 3 5
2
8
3 8 4 6
15
12 X 4 6
29
Step 3: Find the new cell with minimum shipping cost
and cross-out row 2
2 3 5 6
5
2 1 3 5
X
2 8
3 8 4 6
15
10 X 4 6
30
and cross-out row 1
2 3 5 6
X
5
2 1 3 5
X
2 8
3 8 4 6
15
5 X 4 6
31
and cross-out column 1
2 3 5 6
X
5
2 1 3 5
X
2 8
3 8 4 6
10
5
X X 4 6
32
and cross-out column 3
2 3 5 6
X
5
2 1 3 5
X
2 8
3 8 4 6
6
5 4
X X X 6
33
Step 7: Finally assign 6 to last cell. The bfs is found as:
X11=5, X21=2, X22=8, X31=5, X33=4 and X34=6
2 3 5 6
X
5
2 1 3 5
X
2 8
3 8 4 6
X
5 4 6
X X X X
34
3. Vogel’s Approximation Method
Begin with computing each row and column a
penalty. The penalty will be equal to the difference
between the two smallest shipping costs in the row or
column. Identify the row or column with the largest
penalty. Find the first basic variable which has the
smallest shipping cost in that row or column. Then
assign the highest possible value to that variable, and
cross-out the row or column as in the previous
methods. Compute new penalties and use the same
procedure.
35
An example for Vogel’s Method
Step 1: Compute the penalties.
Supply Row Penalty
6 7 8
10 7-6=1
15 80 78
15 78-15=63
Demand 15 5 5
Column Penalty 15-6=9 80-7=73 78-8=70
36
Step 2: Identify the largest penalty and assign the
highest possible value to the variable.
Supply Row Penalty
6 7 8
5 8-6=2
5
15 80 78
15 78-15=63
Demand 15 X 5
Column Penalty 15-6=9 _ 78-8=70
37
Supply Row Penalty
6 7 8
0 _
5 5
15 80 78
15 _
Demand 15 X X
Column Penalty 15-6=9 _ _
38
Supply Row Penalty
6 7 8
X _
0 5 5
15 80 78
15 _
Demand 15 X X
Column Penalty _ _ _
39
Step 5: Finally the bfs is found as X11=0, X12=5, X13=5,
and X21=15
Supply Row Penalty
6 7 8
X _
0 5 5
15 80 78
X _
15
Demand X X X
Column Penalty _ _ _
40
hil61217_ch03_case.qxd 5/12/04 17:11 Page 1
ADDITIONAL CASES
■ CASE 3.2 CUTTING CAFETERIA COSTS

A cafeteria at All-State University has one special dish it These two ingredients are responsible for the greatest costs,
serves like clockwork every Thursday at noon. This suppos- nutritional content, and taste of the dish.
edly tasty dish is a casserole that contains sautéed onions, Maria buys the potatoes and green beans from a whole-
boiled sliced potatoes, green beans, and cream of mushroom saler each week. Potatoes cost $0.40 per pound, and green
soup. Unfortunately, students fail to see the special quality of beans cost $1.00 per pound.
this dish, and they loathingly refer to it as the Killer Casse- All-State University has established nutritional require-
role. The students reluctantly eat the casserole, however, be- ments that each main dish of the cafeteria must meet. Specif-
cause the cafeteria provides only a limited selection of dishes ically, the total amount of the dish prepared for all the stu-
for Thursday’s lunch (namely, the casserole). dents for one meal must contain 180 grams (g) of protein, 80
Maria Gonzalez, the cafeteria manager, is looking to cut milligrams (mg) of iron, and 1,050 mg of vitamin C. (There
costs for the coming year, and she believes that one sure way are 453.6 g in 1 lb and 1,000 mg in 1 g.) For simplicity when
to cut costs is to buy less expensive and perhaps lower-quality planning, Maria assumes that only the potatoes and green
ingredients. Because the casserole is a weekly staple of the beans contribute to the nutritional content of the casserole.
cafeteria menu, she concludes that if she can cut costs on the Because Maria works at a cutting-edge technological
ingredients purchased for the casserole, she can significantly university, she has been exposed to the numerous resources
reduce overall cafeteria operating costs. She therefore decides on the Internet. She decides to conduct a search to find
to invest time in determining how to minimize the costs of the the nutritional content of potatoes and green beans. Her re-
casserole while maintaining nutritional and taste requirements. search yields the following nutritional information about the
Maria focuses on reducing the costs of the two main in- two ingredients:
gredients in the casserole, the potatoes and green beans.
Potatoes Green Beans
Protein 1.5 g per 100 g 5.67 g per 10 ounces

Iron 0.3 mg per 100 g 3.402 mg per 10 ounces
Vitamin C 12 mg per 100 g 28.35 mg per 10 ounces
(There are 28.35 g in 1 ounce.)
Edson Branner, the cafeteria cook who is surprisingly can be served for many days thereafter or can be used cre-
concerned about taste, informs Maria that an edible casse- atively in preparing other dishes.
role must contain at least a six to five ratio in the weight of
(a) Determine the amount of potatoes and green beans Maria
potatoes to green beans. should purchase each week for the casserole to minimize the
Given the number of students who eat in the cafeteria, ingredient costs while meeting nutritional, taste, and demand
Maria knows that she must purchase enough potatoes and requirements.
green beans to prepare a minimum of 10 kilograms (kg) of
casserole each week. (There are 1,000 g in 1 kg.) Again for Before she makes her final decision, Maria plans to explore
simplicity in planning, she assumes that only the potatoes the following questions independently except where other-
and green beans determine the amount of casserole that can wise indicated.
be prepared. Maria does not establish an upper limit on the (b) Maria is not very concerned about the taste of the casserole;
amount of casserole to prepare, since she knows all leftovers she is only concerned about meeting nutritional requirements
1
2 ADDITIONAL CASES
and cutting costs. She therefore forces Edson to change the can purchase lima beans for $0.60 per lb from the wholesaler.
recipe to allow for only at least a one to two ratio in the weight She also knows that lima beans contain 22.68 g of protein per
of potatoes to green beans. Given the new recipe, determine 10 ounces of lima beans, 6.804 mg of iron per 10 ounces of
the amount of potatoes and green beans Maria should purchase lima beans, and no vitamin C. Using the new cost and nutri-
each week. tional content of lima beans, determine the amount of potatoes
(c) Maria decides to lower the iron requirement to 65 mg since she and lima beans Maria should purchase each week to minimize
determines that the other ingredients, such as the onions and the ingredient costs while meeting nutritional, taste, and demand
cream of mushroom soup, also provide iron. Determine the requirements. The nutritional requirements include the reduced
amount of potatoes and green beans Maria should purchase each iron requirement from part (c).
week given this new iron requirement. (f) Will Edson be happy with the solution in part (e)? Why or why
(d) Maria learns that the wholesaler has a surplus of green beans not?
and is therefore selling the green beans for a lower price of (g) An All-State student task force meets during Body Awareness
$0.50 per lb. Using the same iron requirement from part (c) and Week and determines that All-State University’s nutritional re-
the new price of green beans, determine the amount of pota- quirements for iron are too lax and that those for vitamin C
toes and green beans Maria should purchase each week. are too stringent. The task force urges the university to adopt
(e) Maria decides that she wants to purchase lima beans instead of a policy that requires each serving of an entrée to contain at
green beans since lima beans are less expensive and provide a least 120 mg of iron and at least 500 mg of vitamin C. Using
greater amount of protein and iron than green beans. Maria again potatoes and lima beans as the ingredients for the dish and us-
wields her absolute power and forces Edson to change the recipe ing the new nutritional requirements, determine the amount of
to include lima beans instead of green beans. Maria knows she potatoes and lima beans Maria should purchase each week.
■ CASE 3.3 STAFFING A CALL CENTER1

California Children’s Hospital has been receiving numerous exclusively to appointments and registration. The hospital is
customer complaints because of its confusing, decentralized currently in the middle of the planning stages for the call
appointment and registration process. When customers want center. Lenny Davis, the hospital manager, plans to operate
to make appointments or register child patients, they must the call center from 7 A.M. to 9 P.M. during the weekdays.
contact the clinic or department they plan to visit. Several Several months ago, the hospital hired an ambitious
problems exist with this current strategy. Parents do not al- management consulting firm, Creative Chaos Consultants,
ways know the most appropriate clinic or department they to forecast the number of calls the call center would receive
must visit to address their children’s ailments. They there- each hour of the day. Since all appointment and registration-
fore spend a significant amount of time on the phone being related calls would be received by the call center, the con-
transferred from clinic to clinic until they reach the most ap- sultants decided that they could forecast the calls at the call
propriate clinic for their needs. The hospital also does not center by totaling the number of appointment and registra-
publish the phone numbers of all clinic and departments, tion-related calls received by all clinics and departments.
and parents must therefore invest a large amount of time in The team members visited all the clinics and departments,
detective work to track down the correct phone number. Fi- where they diligently recorded every call relating to ap-
nally, the various clinics and departments do not communi- pointments and registration. They then totaled these calls
cate with each other. For example, when a doctor schedules and altered the totals to account for calls missed during data
a referral with a colleague located in another department or collection. They also altered totals to account for repeat calls
clinic, that department or clinic almost never receives word that occurred when the same parent called the hospital many
of the referral. The parent must contact the correct depart- times because of the confusion surrounding the decentral-
ment or clinic and provide the needed referral information. ized process. Creative Chaos Consultants determined the av-
In efforts to reengineer and improve its appointment and erage number of calls the call center should expect during
registration process, the children’s hospital has decided to each hour of a weekday. The following table provides the
centralize the process by establishing one call center devoted forecasts.
1
This case is based on an actual project completed by a team of master’s students in the Department of Engineering-Economic Systems and Operations
Research at Stanford University. (Following a later merger, this department now is the Department of Management Science and Engineering.)
CASE 3.4 PROMOTING A BREAKFAST CEREAL 3
Work Shift Average Number of Calls
7 A.M.–9 A.M. 40 calls per hour

9 A.M.–11 A.M. 85 calls per hour
11 A.M.–1 P.M. 70 calls per hour
1 P.M.–3 P.M. 95 calls per hour
After the consultants submitted these forecasts, Lenny be- For the following analysis consider only the labor cost
came interested in the percentage of calls from Spanish for the time employees spend answering phones. The cost
speakers since the hospital services many Spanish patients. for paperwork time is charged to other cost centers.
Lenny knows that he has to hire some operators who speak (a) How many Spanish-speaking operators and how many English-
Spanish to handle these calls. The consultants performed fur- speaking operators does the hospital need to staff the call cen-
ther data collection and determined that on average, 20 per- ter during each 2-hour shift of the day in order to answer all
cent of the calls were from Spanish speakers. calls? Please provide an integer number since half a human op-
Given these call forecasts, Lenny must now decide how erator makes no sense.
to staff the call center during each 2 hour shift of a weekday. (b) Lenny needs to determine how many full-time employees who
During the forecasting project, Creative Chaos Consultants speak Spanish, full-time employees who speak English, and part-
closely observed the operators working at the individual clin- time employees he should hire to begin on each shift. Creative
ics and departments and determined the number of calls op- Chaos Consultants advise him that linear programming can be
erators process per hour. The consultants informed Lenny that used to do this in such a way as to minimize operating costs
while answering all calls. Formulate a linear programming
an operator is able to process an average of six calls per hour.
model of this problem.
Lenny also knows that he has both full-time and part-time (c) Obtain an optimal solution for the linear programming model
workers available to staff the call center. A full-time employee formulated in part (b) to guide Lenny’s decision.
works 8 hours per day, but because of paperwork that must (d) Because many full-time workers do not want to work late into the
also be completed, the employee spends only 4 hours per day evening, Lenny can find only one qualified English-speaking op-
on the phone. To balance the schedule, the employee alter- erator willing to begin work at 1 P.M. Given this new constraint,
nates the 2-hour shifts between answering phones and com- how many full-time English-speaking operators, full-time Spanish-
pleting paperwork. Full-time employees can start their day ei- speaking operators, and part-time operators should Lenny hire for
ther by answering phones or by completing paperwork on the each shift to minimize operating costs while answering all calls?
first shift. The full-time employees speak either Spanish or (e) Lenny now has decided to investigate the option of hiring bilin-
English, but none of them are bilingual. Both Spanish-speak- gual operators instead of monolingual operators. If all the op-
erators are bilingual, how many operators should be working
ing and English-speaking employees are paid $10 per hour
during each 2-hour shift to answer all phone calls? As in part
for work before 5 P.M. and $12 per hour for work after 5 P.M. (a), please provide an integer answer.
The full-time employees can begin work at the beginning of (f) If all employees are bilingual, how many full-time and part-time
the 7 A.M. to 9 A.M. shift, 9 A.M. to 11 A.M. shift, 11 A.M. to employees should Lenny hire to begin on each shift to minimize
1 P.M. shift, or 1 P.M. to 3 P.M. shift. The part-time employees operating costs while answering all calls? As in part (b), for-
work for 4 hours, only answer calls, and only speak English. mulate a linear programming model to guide Lenny’s decision.
They can start work at the beginning of the 3 P.M. to 5 P.M. (g) What is the maximum percentage increase in the hourly wage
shift or the 5 P.M. to 7 P.M. shift, and like the full-time em- rate that Lenny can pay bilingual employees over monolingual
ployees, they are paid $10 per hour for work before 5 P.M. employees without increasing the total operating costs?
and $12 per hour for work after 5 P.M. (h) What other features of the call center should Lenny explore to
improve service or minimize operating costs?
■ CASE 3.4 PROMOTING A BREAKFAST CEREAL

Claire Syverson, vice president for marketing of the Super company’s new breakfast cereal—Crunchy Start—to suc-
Grain Corporation, is facing a daunting challenge. She needs cessfully enter a crowded breakfast cereal market. Fortu-
to develop a promotional campaign that will enable the nately, Crunchy Start has a lot going for it. Great taste.
4 ADDITIONAL CASES
Nutritious. Crunchy from start to finish. She can recite the Giacomi & Jackowitz has identified the three most ef-
litany in her sleep. It has the makings of a winning promo- fective advertising media for this product:
tional campaign.
Medium 1: Television commercials on Saturday morn-
However, Claire knows that she has to avoid the mis-
ing programs for children.
takes she made in her last campaign for a breakfast cereal
Medium 2: Advertisements in food and family-oriented
(her first big assignment since she won this promotion). She
magazines.
thought she had developed a really good campaign, but
Medium 3: Advertisements in Sunday supplements of
somehow it had failed to connect with the most crucial seg-
major newspapers.
ments of the market—young children and parents of young
children. She also has concluded that it was a mistake not The problem now is to determine which levels should be
to include cents-off coupons (coupons that provide rebates) chosen for these advertising activities to obtain the most ef-
in the magazine and newspaper advertising. fective advertising mix.
She had better get it right this time, especially after the To determine the best mix of activity levels for this par-
big stumble last time. The company’s president, David Sloan, ticular advertising problem, it is necessary (as always) to iden-
already has impressed on her how important the success of tify the overall measure of performance for the problem and
Crunchy Start is to the future of the company. She remem- then the contribution of each activity toward this measure. An
bers exactly how David concluded the conversation. “The ultimate goal for Super Grain is to maximize its profits, but
company’s shareholders are not happy. We need to get those it is difficult to make a direct connection between advertising
earnings headed in the right direction again.” exposure and profits. Therefore, as a surrogate for profit,
Claire already has employed a leading advertising firm, Claire decides to use expected number of exposures as the
Giacomi & Jackowitz, to help design a nationwide promo- overall measure of performance, where each viewing of an
tional campaign that will achieve the largest possible expo- advertisement by some individual counts as one exposure.
sure for Crunchy Start. Super Grain will pay this firm a fee Giacomi & Jackowitz has made preliminary plans for ad-
based on services performed (not to exceed $1 million), vertisements in the three media. The firm also has estimated
and has allocated an additional $4 million for advertising the expected number of exposures for each advertisement in
expenses. each medium, as given in the bottom row of Table 1.
■ TABLE 1 Cost and exposure data
Costs
Each TV Each Each

Cost Category Commercial Magazine Ad Sunday Ad
Advertising costs $300,000 $150,000 $100,000

Planning costs $ 90,000 $ 30,000 $ 40,000
Expected number 1,300,000 600,000 500,000

of exposures
The number of advertisements that can be run in the Consequently, the three limited resources for this prob-
different media are restricted by both the advertising bud- lem are
get (a limit of $4 million) and the planning budget (a limit
Resource 1: Advertising budget ($4 million),
of $1 million for the fee to Giacomi & Jackowitz). An-
Resource 2: Planning budget ($1 million),
other restriction is that there are only five commercial spots
Resource 3: Commercial spots available (5).
available for running different commercials on children’s
television programs on Saturday morning (medium 1). Table 1 shows how much of the advertising budget and the
The other two media have an ample number of spots planning budget would be used by each advertisement in the
available. respective media.
CASE 3.4 PROMOTING A BREAKFAST CEREAL 5
• The first row gives the cost per advertisement in each Requirement 2: The advertising of one type or another
medium. (The cost of using only a fraction of an adver- should be seen by at least 5 million
tising spot is assumed to be that fraction of the cost given parents of young children.
in the table.) In effect, these two requirements are minimum acceptable
• The second row shows Giacomi & Jackowitz’s estimates levels for two special benefits to be achieved by the adver-
of its total cost (including overhead and profit) for de- tising activities.
signing and developing each advertisement for the re-
spective media.1 (This cost represents the billable fee Benefit 1: Promoting the new breakfast cereal to young
from Super Grain.) children.
• The last row then gives the expected number of expo- Benefit 2: Promoting the new breakfast cereal to par-
sures per advertisement. ents of young children.
Since the promotional campaign is for a breakfast cereal that Because of the way the requirements have been articulated,
should have special appeal to young children, Claire feels the level of each of these benefits is measured by the num-
that two audiences should be especially targeted—young ber of people in the specified category that are reached by
children and parents of young children. (This is why one of the advertising.
the three advertising media recommended by Giacomi & To enable the construction of the corresponding bene-
Jackowitz is commercials on children’s television programs fit constraints, Claire asks Giacomi & Jackowitz to estimate
Saturday morning.) Consequently, Claire has established how much each advertisement in each of the media will con-
two requirements for the campaign. tribute to each benefit, as measured by the number of peo-
ple reached in the specified category. These estimates are
Requirement 1: The advertising of one type or another given in Table 2.
should be seen by at least 5 million
young children.
■ TABLE 2 Benefit data
Number Reached in Target Category
Minimum
Each TV Each Each Acceptable
Target Category Commercial Magazine Ad Sunday Ad Level
Young children 1.2 million 0.1 million 0 5 million

Parents of young children 0.5 million 0.2 million 0.2 million 5 million
Claire has one more consideration she wants to incor- ment for coupon redemptions. Because of the importance of
porate into the model. She is a strong believer in the pro- Crunchy Start to the company, she has decided to use this
motional value of cents-off coupons (coupons that shoppers entire remaining allotment in the campaign promoting this
can clip from printed advertisements to obtain a refund of a cereal. Both medium 2 (advertisements in food and family-
designated amount when purchasing the advertised item). oriented magazines) and medium 3 (advertisements in Sun-
Consequently, she always earmarks a major portion of her day supplements of major newspapers) will feature cents-off
annual marketing budget for the redemption of these coupons. The estimates of the amount of coupon redemption
coupons. She still has $1,490,000 left from this year’s allot- per advertisement in each of these media is given in Table 3.
1
When presenting its estimates in this form, the firm is making two simplifying assumptions. One is that its cost for designing and developing each ad-
ditional advertisement in a medium is roughly the same as for the first advertisement in that medium. The second is that its cost when working with one
medium is unaffected by how much work it is doing (if any) with the other media.
6 ADDITIONAL CASES
■ TABLE 3 Coupon redemption data
Contribution Toward Required Amount
Each TV Each Each Required

Requirement Commercial Magazine Ad Sunday Ad Amount
Coupon Redemption 0 $40,000 $120,000 $1,490,000
(a) You now are in Claire’s shoes. Formulate and solve a linear pro- (c) In light of your conclusions in part (b), do you feel that the lin-
gramming model to determine the number of advertisements to ear programming model used in part (a) adequately captures the
run in each of the media in order to maximize the expected num- complexities of this problem for Claire’s purposes? Explain.
ber of exposures while satisfying all the constraints.
(b) For each of the four assumptions of linear programming pre- Note: This case will be continued in Case 13.3, so we sug-
sented in Sec. 3.3, discuss how well you feel it is satisfied for gest that you save your results.
this problem.
Linear Programming
Sensitivity Analysis
What if there is uncertainly about one or
more values in the LP model?
Sensitivity analysis allows us to determine

how “sensitive” the optimal solution is to
changes in data values.
This includes analyzing changes in:

1. An Objective Function Coefficient (OFC)
2. A Right Hand Side (RHS) value of a
constraint
Graphical Sensitivity Analysis
We can use the graph of an LP to see what

happens when:
1. An OFC changes, or
2. A RHS changes
Recall the Furniture problem

Furniture Problem
Max 7T + 5C (profit)
Subject to the constraints:
3T + 4C < 2400 (carpentry hrs)
2T + 1C < 1000 (painting hrs)
C < 450 (max # chairs)
T > 100 (min # tables)
T, C > 0 (non-negativity)
Objective Function
Coefficient (OFC) Changes
What if the profit contribution for tables
changed from $7 to $8 per table?
8
Max X7 T + 5 C (profit)
Clearly profit goes up, but would we want to
make more tables and less chairs?
(i.e. Does the optimal solution change?)
Characteristics of OFC Changes
• There is no effect on the feasible region
• The slope of the level profit line changes
• If the slope changes enough, a different

corner point will become optimal
C
Original
Objective Function Optimal Corner
7T + 5 C = $4040 500 (T=320, C=360)
Still optimal
Revised
400
Objective Function
8T + 5 C = $4360
300
200
Feasible
100 Region
0 100 200 300 400 500 T

C
1000
What if the OFC
became higher? Both have new
Or lower? optimal corner
points
11T + 5C = $5500 600
Optimal Solution 450
(T=500, C=0)
3T + 5C = $2850
Optimal Solution Feasible
(T=200, C=450) Region
0
0 100 500 800 T
• There is a range for each OFC where the
current optimal corner point remains
optimal.
• If the OFC changes beyond that range a

new corner point becomes optimal.
• Excel’s Solver will calculate the OFC

range.
Right Hand Side (RHS) Changes
What if painting hours available changed
from 1000 to 1300?
1300
X
2T + 1C < 1000 (painting hrs)
This increase in resources could allow us to

increase production and profit.
Characteristics of RHS Changes
• The constraint line shifts, which could
change the feasible region
• Slope of constraint line does not change
• Corner point locations can change
• The optimal solution can change

Old optimal
corner point
C (T=320,C=360)
500 Profit=$4040
Feasible region
400 becomes larger
300 New optimal

corner point
(T=560,C=180)
200
Profit=$4820
Original
100
Feasible
Region
0
0 100 200 300 400 500 600 T

Effect on Objective Function Value
New profit = $4,820
Old profit = $4,040
Profit increase = $780 from 300 additional
painting hours
$2.60 in profit per hour of painting
• Each additional hour will increase profit by $2.60

• Each hour lost will decrease profit by $2.60
Shadow Price
The change in the objective function
value per one-unit increase in the
RHS of the constraint.
Will painting hours be worth $2.60 per

hour regardless of how many hours
are available ?
Range of Shadow Price Validity
Beyond some RHS range the value of each
painting hour will change.
While the RHS stays within this range, the

shadow price does not change.
Excel will calculate this range as well as the

shadow price.
Solver’s Sensitivity Report
When Excel Solver is used to find an optimal
solution, the option of generating the
“Sensitivity Report” is available.
Constraint RHS Changes
If the change in the RHS value is within the
allowable range, then:
• The shadow price does not change
• The change in objective function value =
(shadow price) x (RHS change)
If the RHS change goes beyond the

allowable range, then the shadow price
will change.
Objective Function
Coefficient (OFC) Changes
If the change in OFC is within the allowable
range, then:
• The optimal solution does not change
• The new objective function value can be
calculated
Anderson Electronics Example
Decision: How many of each of 4 products
to make?
Objective: Maximize profit
Decision Variables:
V = number of VCR’s
S = number of stereos
T = number of TV’s
D = number of DVD players
Max. 29V + 32S + 72T + 54D (in $ of profit)
Subject to the constraints:
3V + 4S + 4T + 3D < 4700 (elec. components)

2V + 2S + 4T + 3D < 4500 (non-elec. components)
V + S + 3T + 2D < 2500 (assembly hours)

V, S, T, D > 0 (non-negativity)
RHS Change Questions
• What if the supply of nonelectrical
components changes?
• What happens if the supply of electrical

components
– increased by 400 (to 5100)?
– increased by 4000 (to 8700)?
• What if we could buy an additional 400
elec. components for $1 more than usual?
Would we want to buy them?
• What if would could get an additional 250

hours of assembly time by paying $5 per
hour more than usual? Would this be
profitable?
Decision Variables That Equal 0
We are not currently making any VCR’s
(V=0) because they are not profitable
enough.
How much would profit need to increase

before we would want to begin making
VCR’s?
Reduced Cost
of a Decision Variable
(marginal contribution to the obj. func. value)

- (marginal value of resources used)
= Reduced Cost
marginal profit of a VCR = $29

- marginal value of resources = ?
Reduced Cost of a VCR = - $1.0
Reduced Cost is:
• The minimum amount by which the OFC

of a variable should change to cause that
variable to become non-zero.
• The amount by which the objective

function value would change if the variable
were forced to change from 0 to 1.
OFC Change Questions
• For what range of profit contributions for
DVD players will the current solution
remain optimal?
• What happens to profit if this value drops

to $50 per DVD player?
Alternate Optimal Solutions
May be present when there are 0’s in

the Allowable Increase or Allowable
Decrease values for OFC values.
Simultaneous Changes
All changes discussed up to this point have
involved only 1 change at a time.
What if several OFC’s change?

Or
What if several RHS’s change?
Note: they cannot be mixed
When more than one variable changes we turn to

the 100% rule.
The 100% Rule
1. For each increase (decrease) in an objective function
coefficient, calculate (and express as a percentage) the
ratio of the change in the coefficient to the maximum
possible increase (decrease) as determined by the limits
of the range of optimality.
2. Sum all these percent changes. If the total is less than

100 percent, the optimal solution will not change. If this
total is greater than or equal to 100%, the optimal
solution may change
The 100% Rule
∑ (change / allowable change) < 1
RHS Example
• Electrical components decrease 500
500 / 950 = 0.5263
• Assembly hours increase 200
200 / 466.67 = 0.4285
0.9548
The sensitivity report can still be used
Pricing New Variables
Suppose they are considering selling a new
product, Home Theater Systems (HTS)
Need to determine whether making HTS’s

would be sufficiently profitable
Producing HTS’s would take limited

resources away from other products
• To produce one HTS requires:
5 electrical components
4 nonelectrical components
4 hours of assembly time
• Can shadow prices be used to calculate

reduction in profit from other products?
(check 100% rule)
5/950 + 4/560 + 4/1325 = 0.015 < 1
Required Profit Contribution per HTS
elec cpnts 5 x $ 2 = $10
nonelec cpnts 4 x $ 0 = $ 0
assembly hrs 4 x $24 = $96
$106
Shadow
Prices
Making 1 HTS will reduce profit (from other

products) by $106
• Need (HTS profit contribution) > $106
• Cost to produce each HTS:
elec cpnts 5 x $ 7 = $35
nonelec cpnts 4 x $ 5 = $20
assembly hrs 4 x $10 = $40
$95
(HTS profit contribution) = (selling price) - $95
So selling price must be at least $201

Is HTS Sufficiently Profitable?
• Marketing estimates that selling price
should not exceed $175
• Producing one HTS will cause profit to fall

by $26 ($201 - $175)
Sensitivity Analysis for
a Minimization Problem
Burn-Off makes a “miracle” diet drink
Decision: How much of each of 4

ingredients to use?
Objective: Minimize cost of ingredients

Data
Units of Chemical per Ounce of Ingredient
Ingredient
Chemical A B C D Requirement
X 3 4 8 10 > 280 units
Y 5 3 6 6 > 200 units
Z 10 25 20 40 < 1050 units
$ per ounce of ingredient

$0.40 $0.20 $0.60 $0.30
Min 0.40A + 0.20B + 0.60C + 0.30D ($ of cost)
Subject to the constraints
A+B+C+D > 36 (min daily ounces)

3A + 4B + 8C + 10D > 280 (chem x min)
5A + 3B + 6C + 6D > 200 (chem y min)
10A + 25B + 20C + 40D < 280 (chem z max)
A, B, C, > 0
Go to file 4-5.xls

QT Theory Merged OCR

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

QT Theory Merged OCR

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

QT Theory Merged OCR

Uploaded by

Copyright:

Available Formats

BIS 494: Topics in Information Systems

• Mathematical analysis of queues and waiting times in

Prototype Example – ER at County Hospital

• Commercial Queuing Systems

Input Source The Queuing System

 The calling population

 The queue configuration

– Their maximum size (# of jobs the queue can hold)

Multiple Queues Single Queue

1. The service provided can 1. Guarantees fairness

 The Service Mechanism

The Queuing System

The Service Facility

• Service times as well as inter-arrival times are assumed

• The most commonly used queuing models are based on the

Definition: A stochastic (or random) variable Texp( ),

 The Cumulative Distribution Function is: FT ( t )  1  e t

 The standard assumption in many queuing

Two equivalent definitions of the Poisson

 The state of the system = the number of customers in the system

N(t) = Number of customers/jobs in the system at time t

• Steady State condition

With few exceptions Queuing Theory has focused on

Pn = The probability that there are exactly n customers/jobs in the

• Experiment with a model mimicking the

• Cheaper and less risky than experimenting

 A simulation model is primarily descriptive while an

Yes No 11. Documentation, reporting and

 Two of the most important and most challenging issues in

• Find alternative ways of describing/evaluating the system

The Real System

Lq = 0.832/(1-0.83)  4.2 Wq = Lq/   4.2/5  0.83

• It had its early roots in World War II and is

• Primary application areas of Operations

Operations Research is a quantitative approach to

2. Much of this work is done using analytical and

• Another term used for this field is “Management Science"

• Yet other terms sometimes used are “Industrial Engineering"

Deterministic Models Stochastic Models

Deterministic models involve optimization

Lewis–Stern method (or D/L/S method).

 Although linear programming problems can be very

 Two-dimensional linear programming can be solved graphically

 Selecting the product mix in a factory to make

 Linear programming has nothing to do with computer

 The use of the word “programming” here means

 Linear programming involves choosing a course of

Min 5x1 + 2x2

s.t. 2x1 + 5x2 > 10

2 2x1 + 5x2 > 10

 Move the Objective Function Line Toward Optimality

2 2x1 + 5x2 > 10

5x1 + 2x2 = 5(16/5) + 2(4/5) = 88/5

1 2x1 + 5x2 > 10

Max 2x1 + 6x2

s.t. 4x1 + 3x2 < 12

Max 4x1 + 5x2

• The problem of finding the minimum-cost