QUEUING

THEORY
THE PROBABILISTIC ANALYSIS OF WAITING LINES.
Stay in Queue: Short Video
Something we can all relate to…
Characteristics of the Queuing System

The queue discipline (in what order

customers are served)

The nature of the calling population

(where customers come from)

The arrival rate (how often

customers arrive at the queue)

The service rate (how fast customers

are served)
Queue’s Discipline
FIFO/FCFS – First In First Out
/ First Come First Serve

LIFO/LCFS – Last In First

Out / Last Come First Serve

SIRO – Service in random

order

PQ – Priority Queuing

Processor Sharing
Customer Behaviors
Queuing System Design
Poisson Distribution
= for x = 0,1,2,3,4, . . .

Where:
𝑃(𝑥) = probability of x arrivals
x = number of arrivals per unit of time
𝜆 = average arrival rate
= 2.7183 (which is the base of the natural algorithm)
Kendall’s Notation
• The standard system used to describe and classify a queuing node.
• David George Kendall proposed describing queuing models using 3 factors
written A/S/c in 1953 where;
A = Time between arrivals to the queue
S = Service time distribution
c = Number of servers
It has been extended to A/S/c/K/N/D where;
K = Capacity of the queue
N = Calling population / population size to be served
D = Queuing discipline
Note: When the queuing parameters are not specified, it is assumed that;
K = ∞ N = ∞ D = FIFO
4 Models of Queuing Theory
1) M/M/1

2) M/M/m

3) M/D/1

4) M/M/1 with finite source

Model #1: M/M/1
Angie, the branch manager if Citibank and she wants to improve the service of the bank by
reducing the average waiting time of the bank’s client. She was able to determine the average
arrival and the average number of clients serviced per hour. The figures are summarized in the
table below.
How many clients are in the bank at any given time? How much time does a client spend in the
bank? How many clients are waiting to be served? How much time does a client spend waiting?
What is the probability that the teller is busy? What is the probability that there are no clients?

Data Table
Given Description Value
m Number of Tellers 1
𝜆 Arrivals / hr 11
μ Service / hr 12
Model #1: M/M/1
1. Compute the average number of clients in the system (L) at any given time:

2. Compute the average number of hrs a client spends in the system (W).

3. Compute the average number of clients waiting in line ().

4. Compute the average number of hrs a client waits in line ()

Model #1: M/M/1
5. Compute the Probability that the teller is busy ():

6. Compute the Probability that there are no client ().

Model #1: M/M/1
Statistics Table
Equations Descriptions Value
L Number of Clients in the system 11
W Average hrs client spends in the system 1
Average clients waiting on line 10.08
Average waiting hrs per client 0.92
Probability that the teller is busy 91.67%
Probability that there are no client 8.33%
Model #1: M/M/1
Angie, the branch manager of Citibank knows that there is also an opportunity cost for clients
who are idle while waiting in line. She was able to determine the teller’s labor cost as well as the
average opportunity cost of clients who are waiting.
How much is the total cost per shift?

Data Table
Given Description Value
Service cost per teller ($/hr) 5
Cost of waiting ($/hr) 6
h Working hours per shift 8
Model #1: M/M/1
1. Compute the total service cost ()
/ hr = m()
/ shift = / hr (h)
2. Compute for the Total Waiting Cost ($)
/ hr = ()()
/ shift = / hr) (h)
3. Compute the Total Cost ($)
/ hr = /hr + /hr
/ shift = /shift + /shift
Model #1: M/M/1
Cost Table
Costs Per hour Per shift
Total Service Cost ($) $5 $ 40
Total Waiting Cost ($) $ 60.50 $ 484.02
Total Cost ($) $ 65.50 $ 524.02

While total service cost is only $40/shift, Angie should expect a much larger
total cost of $524 because of the opportunity loss of $484.
Seat Work
In a bookstore, cashiers handle transactions with an average time of 2
minutes per transaction. Customers arrive at an average rate of 20
customers per hour. Assuming that arrivals follow a Poisson process
and service times follow an exponential distribution, let's determine:
a) The percentage of time the cashier is idle.
b) The average waiting time for customers.
c) The percentage of customers who have to wait in line before
reaching the cashier.
Model #2: M/M/m
Angie, the branch manager of Citibank and she wants to improve the service of the bank by
hiring an additional teller. She observed that the average arrival and the average number of
clients serviced per teller per hour remains the same. Figures are summarized in the table
below.
What is the probability that there are no clients? How many clients are in the bank at any given
time? How much time does client spend in the bank? How many clients are waiting to be
served? How much time does a client spend waiting? What is the probability that the teller is
busy?
Data Table
Given Description Value
m Number of Tellers 2
𝜆 Arrivals / hr 11
μ Service / hr 12
Model #2: M/M/m
1. Determine the Probability that there are no clients in the bank () from appendix C1. Note:
Appendix C1 is a table value of (multichannel for m=2).
= (m, 𝜆, μ) = (2,11,12) = 0.3714 ~ 37.14%

2. Compute the average Number of Clients in the system (L) at any given time:
𝜆μ ( x +

(11)(12) ( x + = 1.1604
Model #2: M/M/m
3. Compute the average hrs client spends in the system (W):

4. Compute the Average clients waiting on line :

= L - 0.2437
5. Compute the Average waiting hrs per client :
= =
6. Compute the Probability that the teller is busy :
= =
Model #2: M/M/m
Statistics Table
Equations Descriptions Value
Probability that there are no client 37.14%
L Number of Clients in the system 1.16
W Average hrs client spends in the system 0.11
Average clients waiting on line 0.24
Average waiting hrs per client 0.02
Probability that the teller is busy 45.83%
Model #2: M/M/m
Angie, the bank manager knows that with the additional teller, the number of clients waiting in
line will decrease. She observed that the teller’s labor cost as well as the average opportunity
cost of the clients who are waiting remain the same.
How much is the total cost per shift?

Data Table
Given Description Value
Service cost per teller ($/hr) 5
Cost of waiting ($/hr) 6
h Working hours per shift 8
Model #2: M/M/m
1. Compute the total service cost ()
/ hr = m()
/ shift = / hr (h)
2. Compute for the Total Waiting Cost ($)
/ hr = ()()
/ shift = / hr) (h)
3. Compute the Total Cost ($)
/ hr = /hr + /hr
/ shift = /shift + /shift
Model #2: M/M/m
Cost Table
Costs Per hour Per shift
Total Service Cost ($) $10 $80
Total Waiting Cost ($) $1.47 $11.72
Total Cost ($) $11.47 $91.72
Seat Work