Introduction

• Queuing is the study of waiting lines, or queues.

Queuing Models • The objective of queuing analysis is to design
systems that enable organizations to perform
optimally according to some criterion.
• Possible Criteria
– Maximum Profits.

Ref: Applied Management Science, 2nd Ed., Lawrence & Pasternack

– Desired Service Level.
Slides based on textbook slides

1 2

Introduction Elements of the Queuing Process

• Analyzing queuing systems requires a clear • A queuing system consists of three basic
understanding of the appropriate service components:
measurement. – Arrivals: Customers arrive according to some arrival
• Possible service measurements pattern.
– Average time a customer spends in line. – Waiting in a queue: Arriving customers may have to wait
– Average length of the waiting line. in one or more queues for service.
– The probability that an arriving customer must wait
for service. – Service: Customers receive service and leave the system.
3 4

The Arrival Process The Arrival Process

• There are two possible types of arrival • Under three conditions the arrivals can be modeled as a
processes Poisson process
– Orderliness : one customer, at most, will arrive during any
– Deterministic arrival process. time interval.
– Random arrival process. – Stationarity : for a given time frame, the probability of arrivals
within a certain time interval is the same for all time intervals of
• The random process is more common in
equal length.
businesses.
– Independence : the arrival of one customer has no influence
on the arrival of another.
5 6
 The Poisson Arrival Process The Waiting Line Characteristics

P(X = k) = (λt) e • Factors that influence the modeling of queues

k - λt

k!
– Line configuration – Priority
Where
λ = mean arrival rate per time unit. – Jockeying – Tandem Queues
t = the length of the interval. – Balking – Homogeneity
e = 2.7182818 (the base of the natural logarithm).
k! = k (k -1) (k -2) (k -3) … (3) (2) (1).

Note: Arrivals according to Poisson have exponential inter-arrivals.7 8

Waiting Line Configurations Jockeying and Balking

Single Server
Multiple Server
Single Line
Multiple Server
Multiple Lines
Tandem Queue • Jockeying occurs when customers switch lines
once they perceived that another line is moving
faster.
• Balking occurs if customers avoid joining the line
when they perceive the line to be too long.

9 10

Priority Rules Tandem Queues

• These rules select the next customer for service. • These are multi-server systems.
• There are several commonly used rules: • A customer needs to visit several service
– First come first served (FCFS). stations (usually in a distinct order) to complete
– Last come first served (LCFS). the service process.
– Estimated service time. • Examples
– Random selection of customers for service. – Patients in an emergency room.
– Passengers prepare for the next flight.

11 12
 Homogeneity The Service Process

• A homogeneous customer population is one in • In most business situations, service time varies
which customers require essentially the same widely among customers.
type of service. • When service time varies, it is treated as a
• A non-homogeneous customer population is one random variable.
in which customers can be categorized according
to: • The exponential probability distribution is used
– Different arrival patterns sometimes to model customer service time.
– Different service treatments. – Simple model; useful for analysis purposes.

13 14

Schematic illustration of the exponential

The Exponential Service Time Distribution
distribution
f(t) = µe-µt
The probability that service is completed
µ = the average number of customers within t time units
who can be served per time period. P(X ≤ t) = 1 - e-µt
Therefore, 1/µ = the mean service time.

The probability that the service time X is less than some “t.”

P(X ≤ t) = 1 - e-µt

15 16
X=t

The Exponential Distribution - Performance Measures of Queuing

Characteristics System
• The memoryless property.
– No additional information about the time left for the completion of a
• Performance can be measured by focusing on:
service, is gained by recording the time elapsed since the service
started. – Customers in queue.
– For Hank’s, the probability of completing a service within the next 3
minutes is (0.52763) independent of how long the customer has been
– Customers in the system.
served already.
• Performance is measured for a system in steady
• The Exponential and the Poisson distributions are related to state.
one another.
– If customer arrivals follow a Poisson distribution with mean rate λ,
their interarrival times are exponentially distributed with mean time
1/λ. 17 18
 Performance Measures of Queuing Performance Measures of Queuing
System System
• The transient period • The steady state
n n
occurs at the initial period follows the This is a
This is a transient This is the
time of operation. period…like before transient period. transient steady state
period………..
• Initial transient • Meaningful long run period…

performance
behavior is not
measures can be
indicative of long run calculated for the
performance. system when in
Time Time
steady state.

19 20

Steady State Performance Measures

Requirements to Achieve Steady
State Performance P0 = Probability that there are no customers in the system.
Pn = Probability that there are “n” customers in the system.

L = Average number of customers in the system.

In order to achieve steady state, the
effective arrival rate must be less than Lq = Average number of customers in the queue.
W = Average time a customer spends in the system.
the sum of the effective service rates .
Wq = Average time a customer spends in the queue.
k servers
Pw = Probability that an arriving customer must wait
λ< µ1 +µ2+…+µκ for service.
λ< µ λ< kµ
For k servers Each with
For one server ρ = Utilization rate for each server
with service rates µι service rate of µ (the percentage of time that each server is busy).
21 22

Little’ Classification of Queues

Little’s Formulas (Kendal Notation)
• Little’s Formulas represent important relationships
• Queuing system can be classified by:
between L, Lq, W, and Wq.
– Arrival process.
• These formulas apply to systems that meet the – Service process. Example:
following conditions: – Number of servers.
– Single queue systems, – System size (infinite/finite waiting line). M / M / 6 / 10 / 20
– Customers arrive at a finite arrival rate λ, and – Population size.
– The system operates under a steady state condition. • Notation
– M (Markovian) = Poisson arrivals or exponential service time.
L=λW L q = λ Wq L = Lq + λ/µ – D (Deterministic) = Constant arrival rate or service time.
– G (General) = General probability for arrivals or service time.
For the case of an infinite population 23 24
 M/M/1 Queuing System
M / M /1 Queue - Performance Measures
Assumptions
– Poisson arrival process. P0 = 1 – (λ/µ)
Pn = [1 – (λ/µ)](λ/µ)n
– Exponential service time distribution.
L = λ /(µ – λ)
– A single server. The probability that
Lq = λ2 /[µ(µ – λ)] a customer waits in
– Potentially infinite queue. W = 1 /(µ – λ) the system more than
“t” is P(X>t) = e-(µ - λ)t
– An infinite population. Wq = λ /[µ(µ – λ)]
Pw = λ / µ
ρ =λ/µ
25 26

MARY’
MARY’s SHOES - Solution
MARY’
MARY’s SHOES
– Input
• Customers arrive at Mary’s Shoes every 12 λ = 1/12 customers per minute = 60/12 = 5 per hour.
minutes on the average, according to a Poisson µ = 1/ 8 customers per minute = 60/ 8 = 7.5 per hour.
process.
– Performance Calculations µ –λ = 7.5 – 5 = 2.5 per hr.

• Service time is exponentially distributed with an P0 = 1 - (λ/µ) = 1 - (5/7.5) = 0.3333

Pn = [1 - (λ/µ)](λ/µ)n = (0.3333)(0.6667)n P(X<10min) = 1 – e
-2.5(10/60)
average of 8 minutes per customer. = .565
L = λ/(µ - λ) = 2
Pw = λ/µ = 0.6667
Lq = λ2/[µ(µ - λ)] = 1.3333
• Management is interested in determining the ρ = λ/µ = 0.6667
W = 1/(µ - λ) = 0.4 hours = 24 minutes
performance measures for this service system.
Wq = λ/[µ(µ - λ)] = 0.26667 hours = 16 minutes
27 28

MARY’
MARY’s SHOES
An Arena solution M/M/k Queuing Systems
L Ave. Cust. in System 2 1.96 • Characteristics
Lq Ave. Cust. in Queue 1.33 1.33 – Customers arrive according to a Poisson process at a
mean rate λ.
W Time in System 24 min 25.01 min
– Service times follow an exponential distribution.
Wq Time in Queue 16 min 16.98 min
– There are k servers, each of who works at a rate of µ
Rho Utilization 0.67 0.63 customers (with kµ> λ).

– Infinite population, and possibly infinite line.

29 30
 Tandem Queuing Systems Tandem Queuing Systems
• In a Tandem Queuing System a customer must visit • In a Tandem Queuing System a customer must visit
several different servers before service is completed. several different servers before service is completed.

Meats Beverage Meats Beverage

• Examples • Examples
– All-You-Can-Eat restaurant – All-You-Can-Eat restaurant

31 32

Tandem Queuing Systems

• In a Tandem Queuing System a customer must visit

Questions?
several different servers before service is completed.

Meats Beverage

• Examples
– All-You-Can-Eat restaurant
– A drive-in restaurant, where first you place your order, then
pay and receive it in the next window.
Ref: Applied Management Science, 2nd Ed., Lawrence & Pasternack
– A multiple stage assembly line. Slides based on textbook slides

33 34