Waiting Line

Management
Saurabh Chandra

Where the Time Goes

In a life time, the average
American will spend--

SIX MONTHS
Waiting at stop lights
EIGHT MONTHS
Opening junk mail
ONE YEAR
Looking for misplaced objects

TWO YEARS

Unsuccessfully returning phone calls

FOUR YEARS Doing housework

FIVE YEARS Waiting in line

SIX YEARS Eating

Waiting Realities & Perceptions

Inevitability of Waiting: Waiting results from variations in
arrival rates and service rates
Economics of Waiting: High utilization purchased at the
price of customer waiting. Make waiting productive (salad
bar) or profitable (drinking bar).
Skinners Law:
The other line always moves faster.
Jenkins Corollary:
However, when you switch to another other line, the line you
left moves faster.
13-3

Waiting Line Analysis

Waiting occurs in production (?) and service processes.
Time is a valuable resource and hence reduction of
waiting time desirable.
Waiting line exists as
Customers (people or things) arrive faster than
they can be served.
Customers do not arrive at constant rate

Service time is also not constant

5-4

Traditional Cost Relationships

as service improves, cost
increases

5-5

Elements of
Waiting Line Analysis
Waiting line system

consists of arrivals, servers, and waiting line structure

Queue

a single waiting line; finite or infinite

Infinite queue
Finite queue
can be of any length; length of a finite queue is limited

Calling population

source of customers; infinite or finite

5-6

5-7

Essential Features of Queuing Systems

Renege

Calling
population

Arrival
process

Balk

Queue
configuration

Queue
discipline

Service
process

Departure

No future
need for
service

Arrival Process
Arrival
process

Static

Dynamic

Random
arrivals with
constant rate

Random arrival
rate varying
with time

Facilitycontrolled

Accept/Reject

Price

Appointments

Customerexercised
control

Reneging

Balking

Distribution of Patient Inter-arrival Times at a Doctors

clinic

Poisson and Exponential Equivalence

Poisson distribution for number of arrivals per hour (top view)

One-hour
1

Arrival

Arrivals

0
Arrivals

interval

Arrival

62 min.
40 min.
123 min.
Exponential distribution of time between arrivals in minutes (bottom view)

Queue Configurations
Multiple Queue

Single queue

Take a Number
Enter
3

4
8

10
12

11

Are these two

systems same?

 What is meant by Queue Discipline?

Whats the most common one?
Examples of systems with different QDs?

Queue Discipline

Queue
discipline

Static
(FCFS rule)

Dynamic
Selection based
on individual
customer
attributes

selection
based on status
of queue

Number of
customers
waiting

Round robin

Priority

Preemptive

Processing time
of customers
(SPT rule)

Outpatient Service Process Distributions

Elements of
Waiting Line Analysis
Queue discipline

order in which customers are served

First come first served
Last in first out
Random
Others

Elements of Waiting Line Analysis

Basic Waiting Line Structures
Single Channel Single phase (single
server)
Multiple Channel Single phase
(Multiple server)
Single Channel Multiple phase
Multiple Channel Multiple phase
5-16

Configurations
Queue
Departures
After Service

Service
Facility

Arrivals

Single-Server, Single-Phase System

Queue
Arrivals

Phase 1
Service
Facility

Single-Server, Multiphase System

Phase 2
Service
Facility

Departures
after Service

Configurations
Queue
Arrivals

Service
Facility
1

Departures

Service
Facility
2

after

Service
Facility
3

Service

Multi-Server, Single-Phase System

Configurations
Queue
Arrivals

Type 1
Service
Facility
1

Type 2
Service
Facility
1

Type 1
Service
Facility
2

Type 2
Service
Facility
2

Multi-Server, Multiphase System

Departures
after service

Elements of
Waiting Line Analysis
Channels
number of
parallel
servers for
servicing
customers

Phases
number of
servers in
sequence a
customer
must go
through

Arrival Characteristics
Size of the arrival population
Infinite or finite

Arrival distribution
Arrival rate
Average arrival time
Poisson distribution

Behavior
Patient,
Balking: leaving on seeing a line
Reneging: leaving the line before service

Poisson Distribution

where
X
P(X)

= number of arrivals per unit of time (e.g.,

hour)
= probability exactly X arrivals
= average arrival rate (i.e., average number of
arrival per unit of time)
= 2.7183 (known as the exponential constant)

Poisson Distribution

Probability

0.25
0.20
0.15
0.10
0.05

= 2 Distribution

9 10 11

= 4 Distribution
Figure 9.2

Queue Characteristics
Length

Finite (limited) or infinite (unlimited)

Discipline

FIFO common
Other ways to prioritize arrivals

Service Characteristics
Configuration
Servers (channels) and phases (service
stops)
Single-server, multiple-server
Single phase, multiphase system

Service Distribution
Constant or random
Exponential distribution
Service rate, service time

Exponential Distribution

where
t

P(t)

= probability that service time will be greater

than t
= average service rate (i.e., average number of
customers served per unit of time)
= 2.7183 (known as the exponential constant)

service time

Exponential Distribution
Probability That Service Time t

1.0
0.9
0.8
0.7
0.6
0.5

Probability That Service Time t = et for t 0

Average service Rate = 3 Customers per Hour
Average Service Time = 20 Minutes (or 1/3 Hours)
per Customer

= Average Service Rate

0.4
Average Service Rate =
1 customer per hour

0.3
0.2
0.1

0.0|
|
|
|
|
|
|
|
|
|
|
|
|
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00
Time t in Hours

Measuring Queue Performance

= utilization factor of the system

(i.e., probability all servers are busy)

Lq

= average length (i.e., the number of

customers) of the queue

= average number of customers in the

system (i.e., the number in the queue
plus the number being served)

Wq

= average time that each customer

spends in the queue

Measuring Queue Performance

W

= average time that each customer

spends in the system (i.e., the time
spent waiting plus the time spent being
served)

P0

= probability that there are no customers

in the system (i.e., the probability that
the service facility will be idle)

Pn

= probability that there are exactly n

customers in the system

Kendalls Notation
A / B / s / Lmax / POPsize
where
A

= the arrival probability distribution. Typically

choices are M (Markovian) for a Poisson distribution,
D for a constant or deterministic distribution, or G for
a general distribution with a known mean variance.
= the service time probability distribution. Typical
choices are M for an exponential distribution, D for a
constant or deterministic distribution, or G for a
general distribution with a known mean and variance.
= number of servers.

Queuing Models Studied

NAME
(KENDALL
NOTATION)

EXAMPLE

Table 9.2

# OF
SERVERS

TIME
PATTERN

POPLN.
SIZE

Simple system
(M/M/1)

Information
counter at
department store

Single

Exponential

Unlimited

Multiple-server
(M/M/s)

Airline
ticket
counter

Multiple

Exponential

Unlimited

Constant service
(M/D/1)

Automated
car wash

Single

Constant

Unlimited

General service
(M/G/1

Auto repair
shop

Single

General

Unlimited

Shop with
exactly ten
machines that
might break

Multiple

Exponential

Limited

Limited
population
(M/M/s//N)

All models are single phase with a Poisson arrival pattern and a FIFO queue discipline

Queuing Models Studied

1. Arrivals follow the Poisson probability
distribution
2. FIFO queue discipline
3. A single-phase service facility
4. Infinite, or unlimited, queue length. That is,
the fourth symbol in Kendalls notation is
5. Service systems that operate under steady,
ongoing conditions. This means that both
arrival rates and service rates remain
stable during the analysis.

M/M/1 Model

Assumptions
1.
2.
3.
4.
5.
6.
7.

Arrivals are served on a FIFO basis.

Every arrival waits to be served, regardless of the length
of the line; no balking or reneging.
Arrivals are independent, the average number of arrivals
(the arrival rate) is constant.
Arrivals are described by a Poisson probability
distribution, infinite or very large population.
Service times vary from one customer to the next, are
independent of each other, with a known average rate.
Service times occur according to the exponential
probability distribution.
The average service rate is greater than the average
arrival rate; that is, > .

Operating Characteristics
= average number of arrivals per time period (e.g., per hour )
= average number of people or items served per time period
1. Average server utilization in the system:
2. Average number of customers or units waiting in line for service:

3. Average number of customers or units in the system:

Operating Characteristics

4. Average time a customer or unit spends waiting in line for service:

5. Average time a customer or unit spends in the system:

6. Probability that there are zero customers or units in the system:

7. Probability that there are n customers or units in the system:

Advanced Single-Server Models

Constant service times
occur most often when automated equipment or machinery
performs service

Finite queue lengths

occur when there is a physical limitation to length of waiting
line

Finite calling population

number of customers that can arrive is limited

5-36

Advanced Single-Server
Models (cont.)

5-37

M/M/S Model

Same assumptions apply

More than 1 server

= average number of arrivals per time period (e.g., per hour )

= average number of customers served per time per server
s = number of servers

single waiting line and service facility with several

independent servers in parallel
same assumptions as single-server model
s >
s = number of servers
servers must be able to serve customers faster than they arrive

Operating Characteristics
1. Average server utilization in the system:

2. Probability that there are zero customers or units in the system:

3. Average number of customers or units waiting in line for service:

Operating Characteristics

4. Average number of customers or units in the system:

5. Average time a customer or unit spends waiting in line for service:

Wq

Lq

6. Average time a customer or unit spends in the system:

7. Probability that there are n customers or units in the system:

M/D/1 Model
Service rate is constant
Waiting times and number of customers/units
always less than
M/M/s system

= average number of arrivals per time period (e.g., per hour )

= constant number of people or items served per time period

Operating Characteristics
1. Average server utilization in the system:

2. Average number of customers or units waiting in line for service:

3. Average number of customers or units in the system:

Operating Characteristics
4. Average time a customer or unit spends waiting in line for service:

5. Average time a customer or unit spends in the system:

6. Probability that there are zero customers or units in the system:

M/G/1 Model
Service time follows a general distribution
= average number of arrivals per time period (e.g., per hour )
= average number of people or items served per time period
= standard deviation of service time

Operating Characteristics
1. Average server utilization in the system:

2. Average number of customers or units waiting in line for service:

3. Average number of customers or units in the system:

Operating Characteristics
4. Average time a customer or unit spends waiting in line for service:

Wq

Lq

5. Average time a customer or unit spends in the system:

6. Probability that there are zero customers or units in the system:

M/M/S//N Model

Dependent relationship between queue length and

arrival rate
Assumptions
1.
2.
3.
4.
5.
6.

There are s servers with identical service time

distributions.
The population of units seeking service is finite, of size
N.
The arrival distribution of each customer in the
population follows a Poisson distribution, with an average
rate of .
Service times are exponentially distributed, with an
average rate of .
Both and are specified for the same time period.
Customers are served on a first-come, first-served basis.

Operating Characteristics

s
N

= average number of arrivals per time period (e.g., per hour )

= average number of people or items served per time period
= number of servers
= size of population

1. Probability that there are zero customers or units in the system:

Operating Characteristics
2. Probability that there are exactly n customers in the system:

3. Average number of customers or units in line, waiting for service:

Operating Characteristics
4. Average number of customers or units in the system:

5. Average time a customer or unit spends in the queue waiting:

6. Average time a customer or unit spends in the system:

More Complex Systems

Variations may be present
More complex models have been developed
May require computer simulation

Psychology of Waiting
That Old Empty Feeling: Unoccupied time goes
slowly
A Foot in the Door: Pre-service waits seem
longer that in-service waits
The Light at the End of the Tunnel: Reduce
anxiety with attention
Excuse Me, But I Was First: Social justice with
FCFS queue discipline
They Also Serve, Who Sit and Wait: Avoids
idle service capacity

13-52

Psychology of Waiting
Waiting rooms
magazines and newspapers
Televisions

Use of
Mirrors

Supermarkets
magazines
impulse purchases
5-53

Disney

costumed
characters
mobile vendors
accurate wait times
special passes

Psychology of Waiting (cont.)

Preferential treatment

Grocery stores: express lanes for customers with few

purchases
Airlines/Car rental agencies: special cards available to
frequent-users or for an additional fee
Phone retailers: route calls to more or less experienced
salespeople based on customers sales history

Critical service providers

services of police department, fire department, etc.

waiting is unacceptable; cost is not important

5-54