Queuing Models

Quantitative queuing models have been

developed to understand the behavior of
waiting lines and enable the operation
managers to make better decision regarding
the operation of these lines.
Queuing models may be used in order to
understand:
 Input into the system.

These are the units, customers, equipment,

and so on arriving at the queue and waiting
for service.
(1) The population arriving at the queue may
be of an infinite size, or may be finite and have
a limited size.
In a moderate – size manufacturing facility,
with a limited amount of equipment, the
equipment that will fail and form a waiting
line for service would be considered a limited
or a finite population.
 Input into the system contd.
(2) The arriving pattern of the units is another
characteristic of the system inputs.
In most queuing disciplines, arrivals follow the
Poisson probability distribution and the number
of arrival per unit of time can be estimated as a
number of units per hour or day.
For planning a maintenance strategy, the rate of
equipment failure constitutes the arrival pattern.
Therefore, the rate at which the equipment fails,
or the failure rate, is the rate at which the
equipment arrives at the waiting line for the
repair.
 Input into the system contd.
(3) The third characteristic of the arrivals is
whether they remain in the line for service or
get impatient and leave or jump to a different
waiting lines.
Fortunately, in the maintenance queue, the
arriving unit do not have the luxury of
becoming impatient or jumping to a different
line, and therefore they are unable to
complicate the model unreasonably.
 Waiting line characteristics.
There are two major characteristics of waiting lines:
(1) the length of the line and
(2) the line disciple.
A line is allowed to grow without bounds.
Such a waiting line is said to be unlimited.
In the maintenance department, however, the
waiting line or queue by necessity would be a limited
or a restricted line.
If such were to grow without bounds, it would mean
that the maintenance department was getting
farther and farther behind and all equipment would
eventually be out of service and waiting for repair.
 The queue discipline refers to the
order in which the unit in the lines are
serviced.
Most lines follow the first – in, first –
out (FIFO) procedure; that is, the
customers or the unit are serviced in
the order of arrival.
In some situations, however, it may
become necessary for some priority
arrangements to override the FIFO
discipline.
 Service characteristics.
The pattern of the service time and the
configuration of the service facility are the
major characteristics of the service system.
For most activities, service time is random and
is best described by a negative exponential
probability distribution.
Once the average service time has been
established based on historical data or various
work measurement methods, we can
determine the probability that a unit in the
system will require beyond a certain amount
of time:
Where
P(Service time

u = 1 / Average service time

Let us assume that the average service time is
equal to 30 minutes. What is the probability that
an arriving unit will require more then one hour of
service?
 Solution:

There is 13.5% likelihood that an arriving unit

will require more than an hour of service. Stated
differently, the probability is 1 – 0.135 = 0.865 or
86.5% that arriving units will require less than
one hour of service
Therefore
 Figure below shows the probability of service for
average service time of 30 and 60 minutes.
Notice that for a given average service time, the
probability that an arriving unit will require a very
long service time is low.
 Contd.
Given the average service time of 30 minutes,
the probability that a unit will require more
than 180 minutes, of service is nearly zero (0 .
2%).
In most machine – operated service
environments, such as an auto – matic car
wash, service time is fixed and follows a
constant pattern.
 Service Configuration.
The number of phases or interaction point between
the arriving units and the service system and the
number of servers determines the type of service
configuration:
 Single – channel, single – phase.
This is the simplest queue configuration. In this setup
the customers or arriving unit wait in a line or queue
with only one server or service station. The units
receive service from the server and than exit the
system. The drive – in window at a bank (with only
one drive – in facility) or a grocery checkout line with
only one check – out clerk on duty are examples of
this queue configuration, shown in figure 2a.
Figure 2 – various queue configuration
Single – channel, multiple phase.
In this case, the customers interact with more
than one server, as shown in figure 2b. The
bureau of motor vehicles, where various
agents perform different tasks, or a fast food
restaurant, where one clerk takes your order
and receives payment and another serves
your food, are common examples of single –
channel, multiple – phase waiting line
disciplines.
Multiple – channel, single phase.
A bank with service tellers on duty is an
example of this system.
A customer interacts only with the first
available teller to receive complete service
and then exits the system as shown in figure
2c.
 Multiple – channel, multiple – phase.
 Several steps or service phases are required for
complete service, but multiple stations for each
step of the service are available.
 Although the service steps are usually performed in
a particular sequence, all stations are available to
all customers and to further complicate the system,
each phase may have its own waiting line.
 Figure 2d shows multiple – channel, multiple phase
system in which arriving units may move or “jump”
in a random manner to various service facilities in
order to complete different stages of service or
transactions.
 Contd.
 The maintenance department of an
organization constitutes a waiting line.
 Machines awaiting repairs or equipment in
need of preventive maintenance form a
waiting line or a queue.
 Maintenance mangers with the help of
quantitative models, can understand and
evaluate the performance of the maintenance
department.
 Queuing models can help us understand the
following operating characteristics:
 Contd.
• The average number of the units or customers
waiting in the queue for service
• The average number of units in the system (waiting
and being served)
• The average time each unit spends waiting in the
queue
• The average time each unit spends in the system
(wait and service time )
• The probability that a service facility is idle
• The probability that an arriving unit will have to
wait
• The probability of a specific number of units in the
system
Basic Queuing Model Calculations
Since maintenance departments and their
operations usually resemble a single-channel,
single-phase system, we will limit our
mathematical model to this particular
configuration.
 (1) Average number of units waiting to be serviced
or repair.
This can be calculated as follows

(2) Calculating the total number of units that is

either being repaired or waiting to be repaired
 (3) Average wait time in line can be
calculated thus

(4) Average total time in the repair shop can

be calculated as follows
(5) The probability that the maintenance
department is idle can be calculated

(6) The probability that a certain number of

units are in the system can be calculated
as follows