Single-Channel, Single-Phase Models

The simplest, most basic of the waiting line structures illustrated in Figure 16.2 is the single-channel,
single-phase--or, simply, single-server--model. There are several variations of the single-server waiting
line system, and in this chapter we present several of the following frequently used variations:

Poisson arrival rate, exponential service times

Poisson arrival rate, general (or unknown) distribution of service times

Poisson arrival rate, constant service times

Poisson arrival rate, exponential service times with a finite queue

Poisson arrival rate, exponential service times with a finite calling population

The Basic Single-Server Model

In the basic single-server model we assume the following:

Poisson arrival rate

Exponential service times

First-come, first-served queue discipline

Infinite queue length

Infinite calling population

The basic operating characteristics of this single-server model are computed using the following
formulas, where l = mean arrival rate, m = mean service rate, and n = the number of customers in the
waiting line system.
The probability that no customers are in the queuing system (either in the queue or being served) is

The probability of exactly n customers in the queuing system is

The average number of customers in the queuing system (i.e., the customers being serviced and in the
waiting line) is

The average number of customers in the waiting line is

The average time a customer spends in the queuing system (i.e., waiting and being served) is

The average time a customer spends waiting in line to be served is

The probability that the server is busy and a customer has to wait, known as the utilization factor is

The probability that the server is idle and a customer can be served is

EXAMPLE
16.1

A Single-Server Model

The Fast Shop Drive-In Market has one checkout counter where one employee operates
the cash register. The combination of the cash register and the operator is the server (or
service facility) in this queuing system; the customers who line up at the counter to pay
for their selections form the waiting line.
Customers arrive at a rate of 24 per hour according to a Poisson distribution (l = 24), and
service times are exponentially distributed with a mean rate of 30 customers per hour
(m = 30). The market manager wants to determine the operating characteristics for this
waiting line system.
SOLUTION:
The operating characteristics are computed using the queuing formulas for the singleserver model as follows:

Remember that these operating characteristics are averages that result over a period of time; they are
not absolutes. In other words, customers who arrive at the Fast Shop Drive-In Market checkout counter
will not find 3.2 customers in line. There could be no customers or 1, 2, 3, or 4 customers. The value
3.2 is simply an average over time, as are the other operating characteristics.

EXAMPLE
16.2

Waiting Line Cost Analysis

At the Fast Shop Drive-In Market in Example 16.1, the arrival rate of 24 customers per hour
means that, on the average, a customer arrives about every 2.5 minutes (i.e., 2-1/2 60
minutes). This indicates the store is busy. Because of the nature of the store, customers
purchase a few items and expect quick service. Customers expect to spend more time in a
supermarket where they make larger purchases, but they shop at a drive-in market
because it is quicker than a supermarket.
Given customers' expectations, the manager believes that it is unacceptable for a
customer to wait 8 minutes and spend a total of 10 minutes in the queuing system (not
including the actual shopping time). The manager wants to test several alternatives for
reducing customer waiting time: (1) another employee to pack up the purchases; and (2)
another checkout counter.
SOLUTION:
Alternative I: Add an Employee
An extra employee will cost the market manager $150 per week. With the help of the
market's national office's marketing research group, the manager has determined that for
each minute that customer waiting time is reduced, the store avoids a loss in sales of $75
per week. (The store loses money when customers leave prior to shopping because of the
long line or when customers do not return.)
If an employee is hired, customers can be served in less time--the service rate, the
number of customers served per time period, will increase. The service rate with one
employee is

The service rate with two employees will be

We assume the arrival rate will remain the same (l = 24 per hour), since the increased
service rate will not increase arrivals but will minimize the loss of customers. (It is not
illogical to assume an increase in service might eventually increase arrivals in the long
run.)
Given the new l and m values, the operating characteristics can be recomputed as follows:

The average waiting time per customer has been reduced from 8 minutes to 2.25 minutes,
a significant amount. The savings (that is, the decrease in lost sales) is computed as

Since the extra employee costs management $150 per week, the total savings will be

The market manager would probably welcome this savings and prefer two employees.
Alternative II: Add a Checkout Counter
The total cost of a new checkout counter would be $6,000, plus an extra $200 per week
for an additional cashier.
The new checkout counter would be opposite the present counter (so that the servers
would have their backs to each other in an enclosed counter area). There would be several
display cases and racks between the two lines, so that customers waiting in line would not
move back and forth between lines. (Such movement, called jockeying, would invalidate
the queuing formulas for this model.) We will assume that the customers would divide
themselves equally between both lines, so the arrival rate for each line would be half of
the arrival rate for a single checkout counter, or

The service rate remains the same for each counter:

Substituting this new arrival rate and the service rate into our queuing formulas results in
the following operating characteristics:

Using the same sales savings of $75 per week for each minute's reduction in waiting time,
we find that the store would save

Next we subtract the $200 per week cost for the new cashier from this amount saved:

Since the capital outlay of this project is $6,000, it would take about 20 weeks
($6,000/$300 = 20 weeks) to recoup the initial cost (ignoring the cost of interest on the
$6,000). Once the cost has been recovered, the store would save $18.75 ($300.00-281.25)
more per week by adding a new checkout counter rather than simply hiring an extra
employee. However, we must not disregard the fact that during the 20-week cost recovery
period, the $281.25 savings incurred by simply hiring a new employee would be lost.
For the market manager both of these alternatives seem preferable to the original
conditions, which resulted in a waiting time of 8 minutes per customer. However, the
manager might have a difficult time selecting between the two alternatives. It might be
appropriate to consider other factors besides waiting time. For example, the portion of
time the employee is idle is 40 percent with the first alternative and 60 percent with the
second, a significant difference. An additional factor is the loss of space resulting from a
new checkout counter.
The final decision must be based on the manager's own experience and perceived needs.
As we said, the results of queuing analysis simply provide information for decision making.
These two alternatives illustrate the cost trade-offs associated with improved service. As
the level of service increases, the corresponding cost of this service also increases. For
example, when we added an extra employee in alternative I, the service was improved,
but the cost of providing service also increased. But when the level of service was
increased, the costs associated with customer waiting decreased.

Solution of the Single-Server Model with POM for Windows

The "waiting lines" module in POM for Windows can be used to solve all of the queuing models in this
chapter. The solution screen for the single-server model for the Fast Shop Drive-In Market in Example
16.1 is shown in Exhibit 16.1.
Constant Service Times
The single-server model with Poisson arrivals and constant service times is a queuing variation that is
of particular interest in operations management, since the most frequent occurrence of constant
service times is with automated equipment and machinery. This type of queuing model has direct
application for many manufacturing operations.
The constant service time model is actually a special case of a more general variation of the singleserver model in which service times cannot be assumed to be exponentially distributed. Service times
are said to be general, orundefined. The basic queuing formulas for the operating characteristics of the
undefined service time model are as follows:

The key formula for undefined service times is for Lq, the number of customers in the waiting line. In
this formula m and s are the mean and standard deviation, respectively, for any general probability
distribution with independent service times. If we let s = m in the formula for Lq for undefined service
times, it becomes the same as our basic formula with exponential service times. In fact all the queuing
formulas become the same as the basic single-server model.
In the case of constant service times, there is no variability in service times (i.e., service time is the
same constant value for each customer); thus, s = 0. Substituting s = 0 into the undefined service
time formula for Lq results in the following formula for constant service times:

Notice that this new formula for Lq for constant service times is simply the basic single-server formula
for Lq divided by 2. The remaining formulas for L, Wq, and W are the same as the single server formulas
using this formula for Lq.
EXAMPLE
16.3

A Single-Server Model with Constant Service Times

The Petrolco Service Station has an automatic car wash, and cars purchasing gas at the
station receive a discounted car wash, depending on the number of gallons of gas they
buy. The car wash can accommodate one car at a time, and it requires a constant time of
4.5 minutes for a wash. Cars arrive at the car wash at an average rate of 10 per hour
(Poisson distributed). The service station manager wants to determine the average length
of the waiting line and the average waiting time at the car wash.
SOLUTION:
First determine l and m such that they are expressed as rates:

Substituting l and m into the queuing formulas for constant service time gives

The Excel screen for Example 16.3 is shown in Exhibit 16.2.

Solution of the Constant Service Time Model with Excel

Several of the queuing models in this chapter can be solved with Excel including the single-server
model with constant service times. The Excel screen for Example 16.3 is shown in Exhibit 16.2. Notice
that the queuing formula for Lqin cell D6 is also shown on the formula bar at the top of the screen.
Finite Queue Length
For some waiting line systems, the length of the queue may be limited by the physical area in which
the queue forms; space may permit only a limited number of customers to enter the queue. Such a
waiting line is referred to as a finite queue; it results in another variation of the single-phase, singlechannel queuing model.
The basic single-server model must be modified to consider the finite queuing system. For this case
the service rate does not have to exceed the arrival rate (m > l) to obtain steady-state conditions. The
resultant operating characteristics, where M is the maximum number in the system, are

Since Pn is the probability of n units in the system, if we define M as the maximum number allowed in
the system, then PM (the value of Pn for n = M) is the probability that a customer will not join the
system. The remaining equations are

EXAMPLE
16.4

A Single-Server Model with Finite Queue

Slick's Quick Lube is a one-bay service facility next to a busy highway. The facility has
space for only one vehicle in service and three vehicles lined up to wait for service. There
is no space for cars to line up on the busy adjacent highway, so if the waiting line is full (3
cars), prospective customers must drive on.
The mean time between arrivals for customers seeking lube service is 3 minutes. The
mean time required to perform the lube operation is 2 minutes. Both the interarrival times
and the service times are exponentially distributed. The maximum number of vehicles in
the system is four. Determine the average waiting time, the average queue length, and
the probability that a customer will have to drive on.
SOLUTION:

First, we compute the probability that the system is full and the customer must drive
on, PM. However, this first requires the determination of P0, as follows:

Next, to compute the average queue length, Lq, the average number of cars in the

system, L, must be computed as follows:

To compute the average waiting time Wq, the average time in the system, W, must be
computed first.

Finite Calling Populations

The single-server model with a Poisson arrival and exponential service times and a finite calling
population has the following set of formulas for determining operating characteristics:

In this model l is the arrival rate of each member of the population. The formulas for P0 and Pn are both
relatively complex and can be cumbersome to compute manually. POM for Windows has the capability
of solving the finite calling population model.
EXAMPLE
16.5

A Single-Server Model with Finite Calling Population

The Wheelco Manufacturing Company operates a job shop that has 20 machines. Due to
the type of work performed in the shop, there is a lot of wear and tear on the machines,
and they require frequent repair. When a machine breaks down, it is tagged for repair,
with the date of breakdown noted. The company has one senior repair person, with an
assistant, who repairs the machines based on an oldest-date-of-breakdown rule (i.e., a
FIFO queue discipline). Each machine operates an average of 200 hours before breaking
down, and the mean repair time is 3.6 hours. The breakdown rate is Poisson distributed
and the service times are exponentially distributed. The company would like an analysis
performed of machine idle time due to breakdowns in order to determine if the present
repair staff is sufficient.
SOLUTION:

These results show that the repairperson and assistant are busy 35 percent of the time
repairing machines. Of the 20 machines, an average of 0.52 or 2.6 percent, are broken
down waiting for repair or under repair. Each broken-down machine is idle (broken down
waiting for repair or under repair) an average of 5.33 hours. The system seems adequate.
Computer Solution of the Finite Population Model with Excel OM
The finite population queuing model is especially tedious to solve using an Excel spreadsheet because
of the complexity of entering the formula for P in the spreadsheet. Excel OM that we have used in

several other chapters includes a spreadsheet macro for waiting line (i.e., queuing) analysis. Excel OM
includes all the different queuing models that are presented in this chapter. The queuing macro in
Excel OM is one of the more useful ones in the text since some of the queuing formulas are complex
and often difficult to set up manually in spreadsheet cells.
After Excel OM has been activated, the "Waiting Line Analysis" menu is accessed from the "OM" menu
on the menu bar at the top of the spreadsheet. The finite population model is selected from the
resulting menu of different waiting line models. After "Spreadsheet Initialization," the resulting
spreadsheet for Example 16.5 for the Wheelco Manufacturing Company is shown in Exhibit 16.3. As in
our other Excel OM examples, the spreadsheet initially includes example data values, thus the first
step is to enter the parameters for our own example in cells B7:B9 as shown in Exhibit 16.3.