Lecture 4: Waiting line

Instructor: Nooshin Salari

Office: RJC 225
Email: [email protected]
Waiting in lines

• Waiting in lines is a part of everyday life.

• Some estimates state that Americans spend 37 billion hours per year waiting in
lines.
• Whether it is waiting in line at a grocery store to buy deli items (by taking a
number) or checking out at the cash registers (finding the quickest line), waiting in
line at the bank for a teller, or waiting at an amusement park to go on the newest
ride, we spend a lot of time waiting.

1
Surveys

• Waiting in lines is a part of everyday life. Some estimates state that Americans
spend 37 billion hours per year waiting in lines.
• According to a survey1, 44% of customers are unsatisfied with 5-15 minutes of
wait time. Additionally, 56% of them said that long wait times are the most
frustrating forms of poor customer service.
• In a waiting line system, managers must decide what level of service to offer.

1. The Replicant contact center survey

2
 Studying waiting lines in operations management optimizes resource use, reduces wait times, enhances efficiency, manages peak demand,
minimizes costs, prevents bottlenecks, improves customer satisfaction, and supports better decision-making across industries.

Introduction: Why do we need to study waiting lines?

• Waiting lines, also known as queues, are a fundamental aspect of operations management.
• They occur whenever demand for a service temporarily exceeds the system's capacity to
provide that service.
• Companies are able to reduce waiting time and provide faster service by increasing their
service capacity, which usually means adding more servers—that is, more tellers, more
mechanics, or more checkout clerks.
• However, increasing service capacity has a monetary cost, and therein lies the basis of
waiting line, analysis; the trade-off between the cost of improved service and the cost of
making customers wait.
• Waiting lines are analyzed with a set of mathematical formulas which comprise a field of
study called queuing theory.
• Different queuing models and mathematical formulas exist to deal with different types of
waiting line systems.

3
Motivating Example

• Consider a call center with just one employee from 7:00 AM to 8:00 AM.
• Based on prior observations, call center management estimates that, on average, a
call takes 4 minutes to complete.
• On average, 12 calls arrive during a 60-minute period ( equal to 1 call every 5
minutes).
• We need to determine the average waiting time for a customer before speaking to a
customer service representative.

4
Motivating Example

• At first glance, it might seem there should be no waiting time at all. The call center
can handle 15 calls per hour (60 minutes / 4 minutes per call), while calls arrive at
a rate of 12 per hour.

• If a call arrives exactly every 5 minutes and takes exactly 4 minutes to be served,
the call center would indeed have no waiting time, as each call would be perfectly
spaced out and served immediately.

5
Motivating Example

6
Motivating Example

• However, real operations are more variable. In practice, calls do not arrive in a
perfectly systematic manner, and service times vary. This variability can lead to
waiting times even when average capacity exceeds demand.
• For instance, imagine the following scenario:
• The first call arrives at 7:00 am and takes 5 minutes to complete. The call center is
idle for the next 2 minutes.
• The second call arrives at 7:07 am and takes 6 minutes. The call center is busy
until 7:13 am.
• The third call arrives at 7:09 am and must wait until 7:13 am to be served, creating
a waiting time.

7
8
Motivating Example

• This example illustrates how variability in arrival and service times can create
waiting periods, despite the average capacity being sufficient.
• Customers may experience long waits even when the call center has idle periods,
illustrating the need for more sophisticated process analysis tools to manage
variability effectively.
• Managers need to account for this variability to achieve consistent service quality
and minimize customer waiting times.

9
Why would a waiting line ever develop?

• A waiting line forms because of a temporary imbalance between the demand for
service and the capacity of the system to provide the service
• Waiting lines result because customers do not arrive at a constant, evenly paced
rate, nor are they all served in an equal amount of time
• A waiting line is one or more “customers” waiting for service. The customers can
be people or objects, such as machines requiring maintenance, sales orders waiting
for shipping, or inventory items waiting to be used.

10
Why would a waiting line ever develop?

• In a similar fashion, waiting lines can develop even if the time to process a
customer is constant. For example, a subway train is computer controlled to arrive
at stations along its route. Each train is programmed to arrive at a station, say,
every 15 minutes. Even with the constant service time, waiting lines develop while
riders wait for the next train or cannot get on a train because of the size of the
crowd at a busy time of the day
• In general, if no variability in the demand or service rate occurs and enough
capacity is provided, no waiting lines form

11
Variability

• Variability: The randomness associated with a process

• Any form of variability is measured based on the standard deviation (𝜎 𝑜𝑟 𝑆)
• Excel formula =STDEV.P, =STDEV.S
• Does a 5 minutes standard deviation indicate a high variability?

Whether a 5-minute standard deviation indicates high variability depends on the context and the (average)mean of the data.

12
Why averages do not work

• Decisions based on average numbers are wrong on average.

• HealthCeuticals sells a perishable antibiotic. Although demand for the drug varies,
for years the average monthly demand has been 5,000 units, so that’s the quantity
the company currently stocks.
• One day, the boss appears. “Give me a forecast of demand for next year,” he says
to his product manager. “I need it to estimate inventory cost for the budget.”
• The product manager responds, “Demand varies from month to month. Here, let
me give you a distribution.” But the boss doesn’t want a “distribution.” “Give me
a number!” he insists. “Well,” the manager says, “the average demand is 5,000
units a month. So, if you need a single number, go with 5,000.”
• The average demand may miss fluctuations from month to month,

13
Variability

14
Variability

15
 A Queueing System
Queueing System
Queuing refers to the process of
customers or entities waiting in Service area
line for service.

Server 1
Customer Exit
Arrivals

Queue (Waiting Line) Server n

16
Characteristics of a waiting line

• We examine the three components of a waiting line or queuing system:

1. Arrivals or Inputs to the System: These have characteristics such as population

size, behavior, and statistical distribution.
2. Queue Discipline or the Waiting Line Itself: Characteristics of a queue include
whether it is limited or unlimited in length and the discipline of people or items
within it.
3. Service Process: Its characteristics include the number of servers, and the service
time distribution.

17
 Characteristics of a waiting line

Characteristics of a waiting line

Arrival Service process Queue Discipline

characteristics

1. Population 2. Behavior of 3. Arrival

Size Arrivals Patterns

18
Arrival Characteristics 1. Population Size

Finite (limited) Infinite (Unlimited)

• Infinite when the number of customers or arrivals present at any given moment represents
only a small fraction of the potential total arrivals.
• Examples of unlimited populations include cars arriving at a large city car wash, shoppers
arriving at a supermarket
• A finite population has a specific, countable number of potential customers.
• Example: If you are in a class with nine other students, the total customer population for
meeting with the professor during office hours is ten students.
• As the number of students waiting to meet with the professor increases, the population of
possible new customers decreases. There is a finite limit as to how large the waiting line
can ever be.

19
 Patient
Customer Balking
Arrival Characteristics 2. Behavior of
Arrivals
Impatient Reneging
Customers
Jockeying

• Patient customers, whether people or machines, wait in the queue until they are served
and do not switch between lines.
• Balking: Customers who balk refuse to join the waiting line because it appears too long to
meet their needs or interests.
• Reneging: Customers who renege initially join the queue but then become impatient and
leave before completing their transaction.
• Jockeying: A customer may move, or jockey, from a seemingly longer queue to a shorter
one or the one with customers whose shopping carts are lightly loaded.

20
Arrival Characteristics Scheduled Arrivals
3. Arrival Patterns
Random Arrival
Random Arrivals
Distribution

• Customers arrive at a service facility either according to a known schedule, such

as one patient every 15 minutes or one student every half hour, or they arrive
randomly.
• Arrivals are considered random when they are independent of one another and
cannot be predicted exactly.
• In most cases, the arrival is random and is therefore characterized by a probability
distribution.

21
 Scheduled Arrivals
3. Arrival Patterns
Arrival Characteristics
Random Arrival
Random Arrivals
Distribution

• To analyze arrivals at a service facility, we can take two approaches:

✓Inter-Arrival Time Analysis: We examine the time between successive arrivals to see if it
follows a statistical distribution, , such as the Exponential Distribution. It also helps
estimate the average arrival rate (λ), where
λ=1/Average Interarrival time​
o If the average inter-arrival time is 5 minutes, the arrival rate is λ =1/5=0.2 arrivals per
minute.
✓Arrival Count Analysis: We set a time length, t , and determine how many arrivals might
enter the system within this period. λ represents the average arrival rate per unit of time The
number of arrivals per time unit is usually assumed to be Poisson distributed.

22
The distribution of customer arrival:

• In scheduled arrival the arrival distribution is generally constant with exactly the
same time period between successive arrivals. It would have same inter-arrival
time (t) and with variance zero (0).

23
The distribution of customer arrival:

• But in case of random arrivals, the time between arrivals would vary and would
have a positive variance.

24
Inter-Arrival Times

• Arrival Time (Ai): The exact time at which the call center receives the i-th call.
• Example: If the first call is received at 9:00 AM, A1=9:00
• Inter-Arrival Time (IA): The time interval between two consecutive calls,
calculated as: IAi=Ai+1−Ai
• By tracking Ai, we accumulate a large dataset of IAi, which provides insights into
patterns and trends in call arrival behavior.​

Customer 1 Customer 2

Inter-arrival time Time

25
Exponential Inter-Arrival Times (for clarification purposes only and
will not be included in the exam.)

• Inter-arrival times often follow an exponential distribution, a common model for

random arrival processes.

• Key Property: The exponential distribution is memoryless, meaning the

probability of future arrivals is independent of the time since the last arrival.

• Most queueing models assume that the form of the probability distribution of
interarrival times is an exponential distribution.
• Note: Exponential distribution is also referred to as the negative exponential
distribution

26
(for clarification )

• When arrivals at a service outlet occur on a purely random fashion, a plot of the
inter-arrival times yield an exponential distribution

27
Time between arrivals(for clarification )

• IA = Time between 2 arrivals

• IA is a random variable

Customer 1 Customer 2

Inter-arrival time = A Time

Average arrival rate λ= 1/average inter-arrival time

28
Exponential distribution (for clarification )

If A is exponentially
distributed with parameter λ
• 1/λ= average inter-arrival
time

29
Example: Time between arrivals

• At the drive-thru of a McDonald’s

• A car arrives every 5 minutes on average
• Average time between arrival= 5 min
• Average arrival rate λ = 1/5

• At a walk-in clinic with 3 doctors

• On average, patients arrive every 15 minutes
• λ = _________
1/15

30
Arrival rate

• Arrival rate is the number of arrivals per unit of time, while interarrival time is the
time between arrivals. The two are related by the equation

• Average Arrival rate = 1/average inter arrival time.

• Average Inter arrival time = 1/average arrival rate.
• Average Service Rate=1/average service time

31
 Arrival Characteristics (Summary)

Arrival
characteristics

1. Population
2. Behavior 3. Arrival
Size
of Arrivals Patterns

Finite Infinite
(limited) Patient Impatient
(Unlimited) Random Scheduled
Customer Customers
Arrivals Arrivals

Balking Reneging Random

Arrival
Distribution
32
Example

• Which of the following is an example of a finite population in a queuing

system?
• a) Customers arriving at a shopping mall.
b) Vehicles entering a toll booth on a busy highway.
c) Students queuing in a classroom with 30 students.
d) Calls arriving at a call center from an entire city.

33
Example

• What is 'balking' in the context of queuing systems?

• a) Switching between queues to reduce waiting time.
b) Leaving the queue after waiting for some time.
c) Deciding not to join the queue after seeing its length.
d) Arriving in groups rather than individually.

34
Example

• Which of the following describes an 'infinite population' in a queuing system?

• a) Customers arrive at a fixed rate, and no new arrivals occur after the system
reaches capacity.
b) The number of potential arrivals is so large that arrivals are not affected by the
current number in the system.
c) Only a limited number of entities can enter the system at any given time.
d) Arrivals occur in batches instead of individually.

35
Example

• What does the interarrival time represent in a queuing system?

• a) The time it takes to serve one entity.
b) The average time entities spend in the system.
c) The time between two consecutive arrivals.
d) The time entities spend waiting in the queue.

36
 Characteristics of a waiting line

Characteristics of a waiting line

Arrival Service process Queue Discipline

characteristics

1. Population 2. Behavior of 3. Arrival

Size Arrivals Patterns

37
Service Process

1. Number of servers Service area

2. Distribution of service time

Server 1

Exit

Queue (Waiting Line) Server n

38
Service Process Service process

Distribution of
# of servers
service times

Single server multi server

• Number of Servers: The total number of service channels available to handle

customer requests or tasks.
• Distribution of Service Time: The time taken by a server to complete the service
for a customer.

39
Number of servers (single vs multi server)

40
Service Process

• Single-Channel (server) Queuing System: This system features one server. An example
is a drive-in bank with only one open teller.
• Multiple-Channel (server) Queuing System: This system has several servers with
customers waiting in a common line for the first available server. Examples include most
modern banks, large barbershops, airline ticket counters, and post offices.
• Single-Phase System: In this system, the customer receives service from only one station
and then exits. Examples include a fast-food restaurant where the same person takes your
order, brings your food, and takes your payment, or a driver’s license agency where one
person handles the entire process.
• Multi-Phase System: Here, the customer must go through multiple service stations. For
instance, in a fast-food restaurant where you place your order at one station, pay at
another, and pick up your food at a third. Similarly, a busy driver’s license agency might
require you to wait in one line to complete your application, queue again to have your test
graded, and finally go to a third counter to pay your fee.

41
Example

• Which of the following is an example of a Multi-Phase System?

a) A single barber providing all services to a customer.
b) A drive-thru where you order, pay, and pick up your food at separate windows.
c) A bank with multiple tellers serving customers from a single waiting line.
d) A self-checkout counter at a grocery store.

42
Example

• What type of queuing system is a post office with multiple clerks serving
customers from a single waiting line?
a) Single-Channel, Single-Phase
b) Multi-Channel, Single-Phase
c) Multi-Channel, Multi-Phase
d) Single-Channel, Multi-Phase

43
Example

• Which of the following is NOT an example of a Multi-Phase System?

a) A fast-food restaurant where one person takes your order, delivers food, and
processes payment.
b) A drive-thru with separate windows for ordering, payment, and food pickup.
c) A driver's license agency requiring customers to visit multiple counters for
application, testing, and payment.
d) A car wash with separate stations for washing, waxing, and drying.

44
Service Process

• Service time Just as the exact arrival time of an individual is

difficult to predict, so is the actual duration of
the service.

Service starts Service ends

• S = Service time (for one server). S is a random variable

• Service rate (for 1 server) μ = 1/average service time

• Note: It is also common to model service time using exponential distribution

45
 Example: Walk-in clinic

• At a walk-in clinic with 3 doctors

• On average, patients arrive every 15 minutes
• Average service time by each doctor is 20 minutes

• λ = 1/15

• μ = 1/20

46
 Characteristics of a waiting line

Characteristics of a waiting line

Arrival Service process Queue Discipline

characteristics

1. Population 2. Behavior of 3. Arrival

Size Arrivals Patterns

47
Queue Discipline

• Queue Discipline refers to the set of rules or policies that determine

the order in which customers in a queue are served.

• Queueing discipline (Which customers to serve first)

• First-come-first-served (FCFS, a.k.a. FIFO)
• Priority-based rules
• Random order

48
Queue Discipline

The most commonly used rule in service systems is FCFS, however, certain
situations may require different priorities, preempting the FCFS rule.
• Emergency Rooms: In hospital emergency rooms, critically injured patients are
treated ahead of those with less severe conditions, such as broken fingers or noses.
• Express Checkout Lines: In supermarkets, shoppers with fewer than 10 items
may enter an express checkout queue, but within this queue, they are served on a
FIFO basis.

49
• Which of the following is an example of the First-Come-First-Served (FCFS)
discipline?
• a) Patients in an emergency room are treated based on the severity of their
condition.
b) Customers at a retail store are served in the order they arrive at the checkout
counter.
c) Passengers boarding a flight in priority groups.

50
• What is the queue discipline where customers are served based on their
importance or urgency?
• a) First-Come-First-Served (FCFS).
b) Last-In-First-Out (LIFO).
c) Priority-based rules.
d) Random order.

51
Parts of a waiting line

52
Little’s Law

• We consider here a famous and very useful law in queueing theory called Little's
Law.
• Given a queueing system, there are three questions we can ask:

1. What is the average number of items in the system?

2. What is the average waiting time for each item that has entered the system?
3. What is the average arrival or throughput rate for the system?

• Little's law is a theorem in queuing theory that describes the relationship between
the average number of items in a system, the average arrival rate, and the average
time spent in the system

53
Little’s Law

• Little’s Law applies to any stable system, regardless of the distribution of arrival
or service times, as long as the system is in steady state.
• It provides a direct relationship between the average number of items in the
system (L), arrival rate (λ), and average time in the system (W).

• In a factory where items arrive at a constant rate and spend a known amount of
time in the system, Little’s Law provides the average number of items in the
system without needing queue-specific details.

54
Little’s Law

• The arrival rate (λ): the average number of entities (such as customers, vehicles,
calls, or items) that arrive at a system per unit of time.
• The waiting time (W): this is the time between arrival and departure. We typically
have a queue in front of a server. The waiting time includes the serving time so it
is not just the time spent in the queue.
• The queue length (L): The number of units waiting in a queue or system

55
Little’s Law

𝑳 = 𝝀 × 𝒘

Little's Law states that the average number of items within a system equals the average
arrival rate of items into and out of the system, multiplied by the average amount of time an
item spends in the system.

Because of its simplicity and generality, the formula is extremely handy to measure the
effectiveness of system performance.

56
Assumptions

✓Little’s law is applicable if the system is stable, meaning that the total number of items
arriving into the system equals the total number of items leaving, over the long run, and
every customer gets served eventually.
➢ If the arrival rate exceeds the departure rate, the queue grows indefinitely, and the
system does not reach a steady state.
➢ If the departure rate exceeds the arrival rate, the queue will eventually be empty, and
the system will not operate continuously.
✓Make sure your processes are stable and easy to calculate so that you know the average
variables of W, L, and lambda.
✓The unit of measure of these variables should be consistent. If one of them is measured in
weeks, the other two should remain the same.
• Little’s law applies no matter what the distribution of the arrival times is. It could be
regular, random, Poisson or whatever.

57
 Each customer spends 1/3 of an hour in the restaurant.
Therefore, the current capacity per hour is ( 12/3 = 36 customers per hour ).
Example: A Restaurant Since the arrival rate (21 customers per hour) is less than the current capacity (36
customers per hour), Matthew's restaurant can handle the current customer flow without
needing additional tables.

• Matthew owns a small restaurant. He wants to know how many

tables he needs to add to accommodate more customers.
• Currently, the restaurant has 3 tables which can accommodate no
more than 12 customers.
• Matthew measured that, on average, 21 customers arrive at his
restaurant every hour.
• He also determined that, on average, a customer spends around 20
minutes in his restaurant for lunch.
• Does Matthew not to create more space in the restaurant to
accommodate more queuing customers?

58
Solution

• Matthew can find the average number of customers queuing in his restaurant by applying
Little's Law:

• 𝐿: Average # of customers in the restaurant

• 𝜆= Average arrival rate of customers=21 customer per hour
1
• 𝑊= Average time spent in the system= 20 minutes or hour
3

𝐿 =𝜆×𝑊
1
𝐿 = 21 × = 7
3
on average, 7 customers will be queuing in the restaurant. So he does not need to create
more space in the restaurant to accommodate more queuing customers.

59
Example

• An outpatient clinic works 12 hours a day. On an average in this clinic it is

observed that 542 patients visit each day and seek consultation. Typically a patient
spends about 28 minutes in the clinic. Find out how many patients are present in
the clinic on average at a given point of time during the working hours (working
hours=12 hours per day)?
𝐿 =𝜆×𝑊

60
Example

• A bread-manufacturing line typically produces 150 loaves of bread during an 8-

hour shift. It's observed that on average, there are 30 loaves being processed in the
line at any given time. What is the average time it takes to produce a loaf of bread
from the raw ingredients?

61
Example

• John owns a small coffee shop. He wants to know the average number of
customers queuing in his coffee shop, to decide whether he needs to add more
space to accommodate more customers. Currently, his queuing area can
accommodate no more than eight people. John measured that, on average, 40
customers arrive at his coffee shop every hour. He also determined that, on
average, a customer spends around 6 minutes in his store (or 0.1 hours).

62
Review: Flow Rate

• Flow rate, also known as throughput, is a key metric in operations management

that measures the rate at which a process delivers output.
• Flow Rate (Throughput):The average rate at which flow units pass a specific
point in the process. The maximum throughput rate is the process capacity.
• Flow rate=min{Available Input, Demand , Process Capacity}
• Note: We assume that we have sufficient Available Input for our analysis

63
Adjustment
• We can make some slight adjustments to the formula and replace some terms:

The average number of items in a work in progress (WIP): items or materials that are in the
queuing system (L) process but are not yet finished products.
The average arrival rate (or departure
rate) of items in the system (𝜆) Average Flow Rate (Throughput )

the average waiting time of an item in Response time or lead time or flow time
the system (𝑊)

The formula can be presented as: 𝑊IP = Average Flow Rate × 𝑓𝑙𝑜𝑤 𝑡𝑖𝑚𝑒
Throughput calculations can help a business determine the rate at which deliverables reach
consumers.

64
Example
• Imagine you manage a pizza shop. There are five pizzas in the queue waiting for
baking and three pizzas are in the oven being pre-seated.

• The total time, on average for making a pizza, is 40 minutes

• Calculate the average number of pizzas served per hour

65
Example

• Consider a repair shop that averages 2 vehicle deliveries a day. This shop has 10 cars
in it. Calculate the average waiting time for the vehicles in the system?
• Flow rate= 2/day, WIP=10
𝑊𝐼𝑃
• 𝑊IP = Average Flow Rate × 𝑓𝑙𝑜𝑤 𝑡𝑖𝑚𝑒 → 𝑓𝑙𝑜𝑤 time = =10/2= 5
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒
days average wait time.
• What would the average flow time if there are 20 cars in the shop?
𝑊𝐼𝑃
• 𝑓𝑙𝑜𝑤 time = =20/2=10 days average flow time
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒
• In order to reduce flow time, you often first must reduce your WIP.

66
Example: Calculate WIP

67
Example

• ABC Body Shop produces 4 cars a day on average and wishes to reduce its flow time to an average
of seven days per day. ABC Body Shop must maintain a work in process (WIP) count of ……….
cars to achieve it flow time goal of 7 days.
• Recommendation: set your goals to reduce WIP incrementally as your system improvements allow.

• As you can see, this is a simple law, but yet very powerful in describing the behavior of any
production system, including the delivery of a project.
𝑊𝐼𝑃
• Using the equation Average Flow Rate = flow 𝑡𝑖𝑚𝑒 , you can make two conclusions:
✓Average Flow Rate can be increased either by increasing WIP, decreasing flow time or both, and
✓The same Average Flow Rate can be achieved with “large WIP & long flow time” or “small
WIP & short flow time”

68
Example

• A retailer stocks a popular electronic gadget and wants to ensure they meet
customer demand without overstocking. On average, the warehouse holds 1,200
units of the gadget in inventory. The average flow time (from when an order is
placed with the supplier until it is received and stocked) is 10 days. The retailer
operates 7 days a week.
1. Using Little’s Law, calculate the daily demand rate (units sold per day).
2. If the retailer wants to reduce inventory to 900 units while maintaining the same
flow time, what should the daily demand rate be to avoid stockouts?
3. If the flow time increases to 15 days, how much inventory should the retailer
maintain to meet the original demand rate?

69
Example

• Let’s say it’s a hot summer, and you own an ice cream shop. You want to know the
average number of customers you’ll have in your store at any given time to ensure
you have enough employees to handle business and the room in your store to serve
them.
• You have about 20 customers arriving every hour. It takes each customer an
average time of about 15 minutes to pick a flavor, pay, and leave with their ice
cream.

70