Unit 5
Uploaded by
N.NevethanUnit 5
Uploaded by
N.NevethanCK COLLEGE OF ENGINEERING
AND TECHNOLOGY
BA4201– QUANTITATIVE TECHNIQUES FOR
DECISION MAKING
(I Year/ II Semester - MBA)
UNIT-5
QUEUEING MODELS SIMULATION
PREPARED BY
Dr. C.JESURAJ /Associate Professor
Department of S&H (Mathematics)
CK COLLEGE OF ENGINEERING AND TECHNOLOGY
Jayaram Nagar, Cuddalore, 607003.
235
4 INTRODUCTION
Queueing theory is a branch of applied probability theory that deals with queues or waiting
lines. A queueing system consists of customers requiring some kind of service at a service station
where service is provided, join a queue if service is immediately unavailable and eventually leave
after getting service. The purpose of queueing theory is to design and analyze a mathematical
model for various application areas such as manufacturing system, communication systems,
production management, health care management, business sector, transportation sector and
networking.
4.1 INTRODUCTORY CONCEPTS ON BIRTH AND DEATH PROCESS AND
MARKOVIAN QUEUEING MODELS
Queues or Waiting lines:
Queues are visible in all places of our daily life. A queue or waiting line is formed when
customers arrive at a system offering certain facilities and demand service. As examples, we may
consider subscribers' calls arriving at a telephone exchange, patients waiting in hospitals,
machines waiting to be repaired by mechanics and cars waiting at a traffic intersection etc.
Elements of a Basic Queueing System:
The basic characteristics of a queueing system are as follows
Input or arrival pattern of customers
Service pattern
Number of servers or service channels
Capacity of the system
Queue discipline
We describe these characteristics below
Types of Queues:
The service process should specify the number of servers and the arrangement of servers, as the
behavior of the queueing system depends on them also. The following diagrams represent the
framework of queueing systems in which only one queue is permitted to form.
Single server queueing system:
Multiple Servers (in Series) queueing system:
Multiple Servers (in parallel) queueing system:
236
Input or arrival pattern of customers
The input pattern means the manner in which the arrival occurs. It is specified by the inter-arrival
time between any two consecutive arrivals. A measure usually considered is the average length of
the inter-arrival time or its reciprocal, the average number of arrivals per some unit of time. We
deal the queueing systems in which the number of arrivals per unit of time has a Poisson
distribution with mean . In this case, the time between consecutive arrivals has an exponential
distribution with mean 1/ .
Pattern of service
By the pattern of service, we mean the manner in which the service is rendered. It is specified by
the time taken to complete a service. The time may be constant or it may be stochastic. If it is
stochastic, the pattern specification involves the distribution of service time of a unit. We shall
mostly deal the queueing systems in which the number of customer serviced per unit of time has a
Poisson distribution with mean or equivalently the inter-service time has an exponential
distribution with mean 1 / .
Number of servers or service channels
A system may have a single server or a number of parallel servers. An arrival who finds more than
one free server may choose at random any one of them for receiving service. If he finds all the
servers busy, he joins a queue common to all the servers. We discuss both single and multiple
server queueing systems.
Capacity of the system
A system may have an infinite capacity – that is, the queue in front of the server(s) may grow to
any length. If the system has finite capacity, there may be limitation of space, so that when the
space is filled to capacity, an arrival will not be able to join the system and will be lost to the
system. The system is called a delay system or a loss system, according to whether the capacity is
infinite or finite.
Queue discipline
The queue discipline indicates the manner in which the arrivals are taken for service. The most
common discipline is the first come, first served (FCFS), or first in first out (FIFO). The other
service disciplines as follows.
Last come first served (LCFS)
Service in random order (SRO)
Processor sharing (usually adopted in computer systems with a number of terminals)
Priority for receiving service with preemptive service discipline.
Birth and Death Queueing Processes
237
The class of all continuous time Markov chains has an important subclass formed by the birth and
death processes. These processes are characterized by the property that whenever a transition
occurs from one state to another, then this transition can be to a neighboring state only. Suppose
that the state space is S 0,1, 2,,i, , then transition, whenever it occurs from state i, can be
only to a neighboring state (i 1) or (i 1) .
A continuous-time Markov chain X (t ), t T with state space S 0,1, 2,,i, and with rates
qi ,i 1 i , i 0,1, 2,...,
qi ,i 1 i , i 1, 2,...,
qi , j 0, j i 1, i 0,1, 2,..., and
qi i i , i 0,1, 2,..., 0 0
is called (i) a pure birth process, if i 0 for i 1, 2,...,
(ii) a pure death process, if i 0 for i 0,1, 2,..., and
(iii) a birth and death process if some of the i ' s and some of the i ' s are positive.
Birth and Death process
If X(t) represents the number of individual present at time t in a population in which two types of
events occur- one representing birth which contributes to its increase and the other representing
death which contributes to its decrease, then the discrete random process {X(t)} is called the birth
and death process, provided the two events namely birth and death are governed by the following
postulates. If X(t) = n , (n>0)
(i) P[1 birth in (t, t+Δt)]= n (t )t o(t )
(ii) P[0 birth in (t, t+Δt)]= 1 n (t )t o(t )
(iii) P[2 or more births in (t, t+Δt)]= o(Δt)
(iv) Births occurring in (t, t+Δt) are independent of time since last birth.
(v) P[1 death in (t, t+Δt)]= n (t )t o(t )
(vi) P[0 death in (t, t+Δt)]= 1 n (t )t o(t )
(vii) P[ 2 or more deaths in (t, t+Δt)]= o(Δt)
(viii) Deaths occurring in (t, t+Δt) are independent of time since last death.
The birth and death occur independently of each other at any time. The above postulates are called
Birth and Death Poisson Postulates.
Transient and steady state of Queueing system
If the characteristics of a queueing system (input and output parameters) are independent of time
or equivalently if the behavior of the system is independent of time, the system is said to be in
steady-state. Otherwise it is said to be in transient-state.
Symbolic Representation of a Queueing Model (Kendall’s notation)
The notation introduced by Kendall (1951) is generally adopted to denote a queueing model. It
consists of the specifications of three basic characteristics: the input, the service time, and the
number of (parallel) servers. Symbols used to denote some of the common formulations are as
follows:
M - Exponential inter-arrival (Poisson input) and service time distribution (having Markov
property)
Ek - Erlang-k distribution
H - Hyper-exponential distribution
PH - Phase-type distribution
238
D - Deterministic (constant)(inter-arrival or service time)
G - Arbitrary (general) distribution.
Generally a queueing model is specified and represented symbolically in the form
(a / b / c) : (d / e) where
a - type of distribution of the number of arrivals per unit time;
b - type of distribution of the service time
c - number of servers
d - capacity of the system, viz., the maximum queue size.
e - queue discipline.
Markovian Queueing models
Queueing models in which both inter-arrival time and service time are exponentially
distributed are called Markovian Queueing models. Poisson arrivals and exponential service make
queueing models Markovian that are easy to analyze and get usable results. Markovian Queueing
models analyze the behavior of the queueing system to take intelligent decisions easily and
quickly in various application sectors.
Theorem 4.1
Derive the steady state probability for the general Poisson Queueing system or general birth
and death process.
Solution
Let Pn (t ) be the probability that there are n customers in the system at time t (n > 0). Let n be
the average arrival rate when there are n customers in the system (both waiting in the queue and
being served) and let n be the average service rate when there are n customers in the system.
Now, Pn (t t) is the probability of n customers at time t t .
The presence of n customers in the system at time t t can happen in any one of the following
four mutually exclusive ways.
(i) Presence of n customers at t and no arrival or departure during t time.
(ii) Presence of (n-1) customers at t and one arrival and no departure during t time.
(iii) Presence of (n+1) customers at t and no arrival and one departure during t time.
(iv) Presence of n customers at t and one arrival and one departure during t time. (more than one
arrival/departure during t time is ruled out)
Pn (t t ) P(i) P(ii) P(iii) P(iv)
Pn (t )(1 n t )(1 n t ) Pn 1 (t )(n 1t )(1 n 1t )
Pn 1 (t )(1 n 1t )( n 1t ) Pn (t )(n t )( n t )
Pn (t) (λ n μ n )Pn (t) t λ n 1Pn 1 (t) t
μ n 1Pn 1 (t) t (omitting higher powers oft )
Pn (t t ) Pn (t ) (n n ) Pn (t )t n 1Pn 1 (t )t n 1Pn 1 (t )t
Pn (t t ) Pn (t )
n 1 Pn 1 (t ) (n n ) Pn (t ) n 1 Pn 1 (t ) (1)
t
Take lim on both sides of (1), we get
t 0
Pn (t ) n1Pn1 (t ) (n n ) Pn (t ) n1Pn1 (t ) (2)
'
Equation (2) is not valid for n=0, as Pn 1 (t ) does not exist.
239
Hence we derive the differential equation satisfied by P0 (t ) independently.
P0 (t t ) P0 (t )(1 0 t ) P1 (t )(1 1t ) 1t (By the possibilities (i)& (iii) given above and no
service is possible when n = 0)
P0 (t t ) P0 (t )
0 P0 (t ) 1 P1 (t ) (3)
t
Take lim on both sides of (3), we get
t 0
P0 (t ) 0 P0 (t ) 1P1 (t ) (4)
'
In the steady- state , Pn (t ) and P0 (t ) are independent of time and hence Pn ' (t ) and P0 ' (t ) become
zero. Hence the differential equation (2) and (4) reduce to the difference equation.
n1Pn1 (n n ) Pn n 1Pn 1 0 (5)
0 P0 1 P1 0 (6)
From the equation (6), we get
P1 0 P0 (7)
1
Substitute n = 1 in (5), we deduce
0
2 P2 (1 1 ) P1 0 P0 (1 1 ) P0 0 P0 0 1 P0
1 1
0 1
P2 P (8)
1 2 0
Successively substitute n = 2,3,.. in (5) and proceeding similarly, we get
P3 0 1 2 P0
1 2 3
...
P4 0 1 2 3 P0 and so on Pn 0 1 2 n 1 P0 n 1, 2,..
1 2 3 4 1 2 3 ... n
We know,
012 .......n 1
012 .......n 1
Pn 1 P0 Pn 1 P0 P0 1 P0 1 1
n 0 n 1 n 1 1 2 3 ....... n i n 1 12 3 .......n i
1
P0 10
0 12 .......n 1
1
n 1 1 2 3 ....... n
Worked Examples 4.1(A)
Example 1
Give a real life situation in which (i) Customers are considered for service with last in first
Out queue discipline (ii) a system with infinite number of servers.
Solution:
The following are the real life situation where the given queue discipline used (i) In most Cargo
handling situations where the last item loaded is removed first because it reduces handling and
transportation cost, the last ones being easier to reach closer. (ii) Railway reservation counters, big
Shopping Malls etc .
Example 2
Define various distributions in arrival and service mechanism.
240
Solution:
The mode of service is represented by means of the probability distribution of the number of
customers serviced per unit time follows a Poisson distribution and the inter-service time has an
exponential distribution.
The following are different Service process.
i) Single Server queueing system
ii) Multiple servers (in parallel) queueing system.
iii) Multiple servers (in series) queueing system
Example 3
Write down the formula for probability of n customers in the system in birth and death
queuing model.
Solution:
Probability of n customers in the system is defined as
...
Pn 0 1 2 n 1 P0 n 1, 2,..
1 2 3 ... n
1
P0
.......n 1
1 0 1 2
n 1 1 2 3 ....... n
The above probabilities are called steady state system size probabilities for the general birth and
death model.
Example 4
Draw the transition graph of birth and death process.
Solution:
Example 5
Which Queue is called Queue with discouragement
Solution
Customer is discouraged to join the queue expecting a long waiting time for having the impatience
in getting the service, then the queuing model is said to be Queue with discouragement. Customer
impatience to join the queue due to discouragement has a very negative impact on the queuing
system. This lead to the loss of potential customers which affects the system as a whole. We can
notice this typical queue in situations involving impatient telephone switchboard customers, the
hospital emergency rooms handling critical patients, and the inventory systems that store
perishable goods .
Example 6
Write the balance equation for the state 0 from the following diagram in steady state.
241
Solution
Let , be the average arrival rate and average service rate that there are n customers in the
system at time t (n>0).
From state 0 to 1, is the arrival rate
From state 0 to 2 , is the arrival rate
2
From state 1 to 0, & 2 to 0, is the service rate
The balance equation for the state 0 from the above diagram in steady state is
P0 P1 P2
2
Example 7
What is the relation between Poisson and Exponential distribution in Queueing Process?
Solution
If the number of arrivals in a time interval of length (t) follows a Poisson Process, then
corresponding inter-arrival time follows an 'Exponential Distribution'.
If the inter-arrival times are independently, identically distributed random variables with
an exponential probability distribution, then the number of arrivals in a time interval of length
(t) follows a Poisson Process with arrival rate identical with parameter of the Exponential
Distribution.
Example 8
Why and where queues are formed?
Solution
Queues or waiting lines arise when the demand for a service facility exceeds the capacity of that
facility, that is, the customers do not get service immediately upon request but must wait, or the
service facilities stand idle and wait for customers. Some customers wait when the total number of
customers requiring service exceeds the number of service facilities, some service facilities stand
idle when the total number of service facilities exceeds the number of customers requiring service.
Typical examples might be:
Banks/supermarkets - waiting for service
Computers - waiting for a response
Failure situations - waiting for a failure to occur e.g. in a piece of machinery
Public transport - waiting for a train or a bus.
Example 9
What are the limitations in the usage of queueing theory?
242
Solution
Major Limitations of queuing theory are most of the queuing models are very complex and cannot
be easily understood. The element of uncertainty is there in almost all queuing situations.
Uncertainty arises due to:
i) We may not know the form of theoretical probability distribution which applies.
ii) We might not know the parameters of the process even when the particular distribution is
known.
iii) We would simply be known only the probability distribution of out-comes and not the
distribution of actual outcomes even when (i) and (ii) are known.
Example 10
What does the term customer and Server mean in a queueing system?
Solution
System Customers Servers
Bank, Post offices Customers arriving Tellers serve the customers
Pizza parlor Request for Pizza delivery Pizza parlor sends out trucks to
deliver Pizza
Airports Aero planes Runway
Water service stations Cars Bays
In the following section we discuss about various queueing models both in single and multiple
server considering the system capacity as finite or infinite. We derive the system size probabilities
P0 & Pn and also we find various characteristics for the models by using little’s formula.
4.2 SINGLE SERVER QUEUEING MODEL WITH INFINITE CAPACITY
[(M/M/1): ( /FIFO)]
Introduction
In M/M/1 queueing system, the arrivals occur from an infinite source in accordance with a Poisson
process with parameter that is, the inter-arrival times are independent exponential with mean
1/ ; the service times are independently and exponentially distributed with parameter (say, );
and there is only one server. The queue discipline is FCFS; the utilization factor is / .
Little’s Formula
If any one of the quantities E[ Ns ], E[ Nq ], E[Ws ]& E[Wq ] is known, the other three can be find
using the following relations.
Ls E[ N s ] E[Ws ]
2
Lq E[ N q ] E[Wq ](or) E[ N q ] E[ N s ]
1
E[Ws ] E[Wq ]
1
E[Wq ] E[N q ] (or ) E[Wq ] E[N s ]
The above relations are called Little’s formulas.
Theorem 4.2
Derive the steady state probability for the M/M/1 Poisson Queueing model. (or)
Derive the system size probability for the [(M/M/1): ( /FIFO)] model.
243
Solution
From theorem 4.1, we know
...
Pn 0 1 2 n 1 P0 n 1, 2,.. (1)
1 2 3 ... n
1
P0 2
0 12 .......n 1
1
n 1 1 2 3 ....... n
In M/M/1 queueing model, we assume n & n
i.e. 1 2 ...n & 1 2 ...n
From equation (1), we get
1 1 1
P0 1
1
1 n n 1
n n
n 1 n 0
Probability of no customers in the system is P0 1 1
From equation (1), we get
n
n
Pn n P0 1
n
Probability n customers in the system is Pn 1 n 1 , where
Some Important characteristics for M/M/1 model
Probability of the system is idle: P0 1 1
n
n
Probability of n customer in the system: Pn n P0 1 n 1
Average number of customers in the system: E[ N s ]
2
Average number of customers in the queue: E[ N q ]
( )
1
Average waiting time of a customer in the system: E[Ws ]
Average waiting time of a customer in the queue: E[Wq ]
The above characteristics can be derived by using the following little’s formula
244
Ls E[ N s ] E[Ws ]
2
Lq E[ N q ] E[Wq ](or) E[ N q ] E[ N s ]
1
E[Ws ] E[Wq ]
1
E[Wq ] E[N q ] (or ) E[Wq ] E[N s ]
In the above four characteristics, if we know any one of the result, we can find other by using
little’s formula.
Average number of customers in non-empty queues: L w
k 1
Probability that the number of customers in the system P ( N k )
Probability that the waiting time of a customer in the system exceeds t P(Ws t ) e( )t
Probability density function of the waiting time in the system( Ws ) f ( ) ( )e ( )
Probability density function of the waiting time in the system( Wq )
( )
( )e , 0
g( )
1 , when 0
1
Average waiting time of a customer in the queue, if he has to wait E Wq / W 0
Worked Examples 4.2(A)
Example 1
Draw the state transition diagram for M/M/1 queueing model.
Solution
Example 2
What effect does doubling and have on Ls and Ws for an (M/M/1): (FIFO/ / ) queueing
model ?
Solution
245
The formula for Average number of customers in the system is Ls
2 2
When and are doubled, we have Ls
2 2 2
Ls
Then the formula for Average waiting time of customers in the system is Ws
When and are doubled, there will be no effect on Ls and Ws for an (M/M/1): (FIFO/ / )
queueing model.
Example 3
Consider an M/M/1 queueing system. Find the probability that at least ‘n’ customer in the
system.
Solution
n 1
The Probability that at least n customers in the system P( N n)
n
i.e P ( N n) Pn P0
n n 1 n n 1
n
1 P0 1
n n 1
n 1 n2
1 ......
n 1
2 n 1 1 n 1
1 1 .... 1 1
n 1
P ( N n)
n 1
Therefore the Probability that at least n customers in the system is
Example 4
Define Traffic intensity.
Solution
Utilization factor or traffic intensity is the average function of time that the servers are being
utilized while serving customers.
Mean arrival rate
Mean service rate
Example 5
A duplicating machine maintained for office use is operated by an office assistant who earns
Rs. 5 per hour. The time to complete each job varies according to an exponential
distribution with mean 6 min. Assume a Poisson input with an average arrival rate of 5 jobs
per hour. If an 8 hour day is used as a base, determine the % idle time of the machine.
Solution
246
Servers - Duplicating machine - Single Server
System capacity – jobs (infinite)
Model: (M/M/1): ( /FIFO)
1 60
5 / hour , / min s 10 / hour
6 6
λ 5 1
P (the machine is idle) = P (N=0) = P0 1 = 1 =1
μ 10 2
Percentage of idle time of the machine = 50
Example 6
Suppose that customers arrive at a Poisson rate of one per every 12 minutes and that the
service time is exponential at a rate of one service per 8 minutes
(i) What is the average number of customers in the system.
(ii) What is the average time of a customer spends in the system?
Solution
Model: (M/M/1): ( /FIFO)
1 1
/ min , / min
12 8
a) Average number of customers in the system is given by
1/12 0.0833
Ls 2
1/ 8 1/12 0.0417
b) Average waiting time of a customers in the system is given by
L
Ws s (By Little’s formula)
2
Ws 2 12 24 min s
1/12
Ws 24 min s
Example 7
In a railway marshalling yard, goods trains arrive at a rate of 30 trains per day. Assuming
that the inter-arrival time follows an exponential distribution and the service time
distribution is also exponential with an average 36 minutes. Calculate the mean queue size.
Solution
Servers - A railway marshalling Yard – Single Server
System capacity – goods Trains - Infinite Capacity
Model: (M/M/1): ( /FIFO)
Mean arrival rate 30 trains per day
Mean service time = 36 minutes
1 1 1
/ min 60 / hr 60 24 / day 40 / day
36 36 36
Mean service rate
40 trains per day
30 3
40 4
247
3 3
/
3 Ls 3
4 4
W.K.T Ls
1 / 1
3 1
1
4 4
The Mean queue size is given by
3 9
Lq Ls 3 2.25
4 4
Lq 2.25 trains 2 trains
Example 8
In the usual notation of an M/M/1 queueing system, of = 12 per hour and = 24 per hour,
find the average number of customers in the system.
Solution
Model: (M/M/1): ( /FIFO)
Given 12 / hour and 24 / hour
Average number of customers in the system is given by
12
Ls 1
24 12
Example 9
Arrivals at a telephone booth are considered to be Poisson with an average time of 12 min.
between one arrival and the next. The length of a phone call assumed to be distributed
exponentially with mean 4min. Find the average number of persons waiting in the system.
Solution
Given: Telephone booth – Single server
Telephone calls – Infinite system capacity
Model: M / M /1 : / FIFO
Mean inter arrival time = 12 minutes
Mean arrival rate 1/ 12 per minute
Mean service time = 4 minutes
Mean service rate 1/ 4 per minute
1/12 1
1/ 4 3
The average number of persons waiting in the system is given by
/ 1/ 3 1
Ls
1 / 1 1 1 1/ 3 2
Example 10
Arrival rate of telephone calls at a telephone booth is according to a poisson distribution
with an average time of 9 minutes between two consecutive arrivals. The length of a
telephone call is assumed to be exponentially distributed with mean 3 minutes. Determine
the probability that a person arriving at the booth will have to wait.
Solution
Servers – Telephone booth – Single server
System capacity – Telephone calls – Infinite capacity
Model: M / M /1 : ( / FIFO)
248
Mean inter arrival time = 9 mins
Mean arrival rate 1/ 9 per min
Mean service time = 3 minutes
Mean service rate 1/ 3per min
1/ 9 1
1/ 3 3
The probability that a person arriving at the booth will have to wait
= P(the system is busy) = 1- P(the system is idle) = 1 P0 1 (1 ) 1/ 3
Example11
What is the probability that a customer has to wait more than 15 minutes to get his service
M/M/1 : /F IFO queue system if = 6 per hour and = 10 per hour?
Solution
Model: M / M /1 : ( / FIFO)
15 1
6 / hr , 10 / hr , t 15 min s hr hr
60 4
The probability that the waiting time of customers in the system exceeds time t
P(Ws t ) e
t
P (a customer has to wait more than 15 mins in the system)
1
(106)
P(Ws 15) P(Ws 1/ 4) e e1 0.3679
4
Example12
Consider an M/M/1queueing system. If = 6 and = 8, find the probability of at least 10
customers in the system.
Solution
Model : M / M /1 : ( / FIFO)
6, 8
P (at least n customers in the system) is
n
P ( N n)
P (at least 10 customers in the system) is
10
6
P( N 10) (0.75)10 0.0563
8
Example 13
A Super market has single cashier. During peak hours, customers arrive at a rate of 20 per
hour. The average number of customers that can be serviced by the cashier is 24 per hour.
Calculate the probability that the cashier is idle.
Solution
Server - Cashier – Single server
System capacity – Customers – infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 20 / hour
Mean service rate 24 / hour
249
20
24
20 4
P (the cashier is idle) = P0 1 1
24 24
Example14
If people arrive to purchase cinema tickets at the average rate of 6 per minute, it takes an
average of 7.5 seconds to purchase a ticket. If a person arrives 2 min before the picture
starts and if he takes exactly 1.5 minutes to reach the correct seat after purchasing the
ticket, Can he expect to be seated for the start of the picture?
Solution
Servers-Single server
System capacity - infinite capacity
Model: M / M /1 : ( / FIFO)
6 / min s , Mean service time = 7.5 seconds
1 60
Mean service rate per sec / min 8 / min
7.5 7.5
8 / min
Average waiting time of customers in the system is given by
1 1 1
E Ws min s
86 2
E [Total time required to purchase the ticket and to reach the seat]
1
= Ws +1.5min= +1.5 = 2 mins
2
Hence, he can be just seated for the start of the picture.
Example15
In a given M/M/1 / /FCFS queue, = 0.6, what is the probability that the queue contains
5 or more customers?
Solution
Model: M / M /1 : ( / FIFO)
n
The probability of at least n customers in the system is given by P ( N n)
The probability of 5 or more customers in the system is given by
P( N 5) (0.6)5 0.0778
5
Example16
In a given M/M/1 /FCFS queue, 0.7 , Find the expected number of customers in the
queue .
Solution
Model: M / M /1 : ( / FIFO)
Expected number of customers in the queue Lq Ls
0.7
We know Ls 2.33
1 1 0.7
250
Lq 2.33 0.7 1.63
Expected number of customers in the queue Lq 1.63
Example17
In the usual notation of an M/M/1 queuing system, if 3 / hr and 4 / hour ,
find PX 5 where X is the number of customers in the system.
Solution
Model: (M/M/1): ( /FIFO)
Given 3 / hour and 4 / hour
3
4
n
P (X n)
5
3
P(X 5) 0.2373
5
4
Example18
Show that for the (M/M/1) : (FCFS/ / ) , the distribution of waiting time in the system is
w(t ) e t , t 0
Solution
Let Ws be the continuous random variable that represents the waiting time of a customer in the
system.
Let the pdf of Ws be f ( w) and let f (w/ n) be the density function of Ws subject to the condition,
that there are n customers in the system .
Then f ( w) f ( w / n) Pn
n 0
f(w / n) = pdf of sum of (n +1) service times , the random variables is exponentially distributed
with parameter
n 1
e w wn , w 0 which is the pdf of Erlang Distribution.
n!
n
n 1 w 1
f ( w) w
w
n
e n
w 1 e 1
n 0 n! n 0 n !
1 e we w , by exp onential distribution
f ( ) e
w
Example19
Arrivals at a telephone booth are to be Poisson with an average time of 10 min between one
arrival and the next. The length of a phone call is assumed to be distributed exponentially
with mean 3 min. What is the average length of the queue that forms from time to time?
Solution
251
Servers – A telephone booth - Single server
System capacity – arrivals – Infinite capacity
Model: M / M /1 : ( / FIFO)
1
Mean arrival rate / min,
10
1
Mean service rate / min
3
1
10 3
0.3min
1 10
3
Average length of queue that forms from time to time = Average number of customers in non-
1 1
empty queues: L w = 1.43
1 1 0.3
1
Example20
Customers arrive at one- man barber shop according to a Poisson process with a mean inter arrival
of 20 min. Customers spend an average of 15 min. in the barber’s chair. The service time is
exponentially distributed. If an hour is used as a unit time, then (i) what is the probability that a
customer need not wait for a haircut?
Solution
Server – One man barber shop – single server
System capacity – customers – infinite capacity
Model M / M /1 : ( / FIFO)
Mean arrival time = 20 mins
Mean service time = 15 mins
1 1
Mean arrival rate per min 60 per hr 3 / hr
20 20
1 1
Mean service rate per min 60 per hr 4 / hr
15 15
3
4
Probability that a customer need not wait for a haircut = P(no customer in the system)
3 1
P0 1 = 1
4 4
Worked Examples 4.2(B)
Example 1
Calculate any four measures of effectives of M / M / 1 queueing model.
Solution:
Average number L s of customers in the system:
Let N denote the number of customers in the queuing system (i.e.,those in the queue and
the one who is being served) and which takes values 0, 1, 2,… such that
252
n
λ
P(N=n)= Pn = P0
μ
We know
1 1 λ
P0 1
λ
n
λ
n
μ
1
n 1 μ n 0 μ
n
λ λ
Pn 1
μ μ
λ
n 1 2 2
λ λ λ λ λ λ λ λ λ
LS = 1 n 1 1 2 3 ... = 1 1
μ μ n 1 μ μ μ μ μ μ μ μ
LS
1
Average number of customers in the queue or Average length of the queue
If N is the no of customers in the system, then the no of customers in the queue is (N-1)
n
E N q Lq E ( N 1) (n 1) Pn (n 1) 1
n 1 n 1
n
λ λ
1 (n 1)
μ n 1 μ
λ
2 n 2 2 2
λ λ λ λ λ λ
= 1 (n 1) 1 1 2
3 ...
μ μ n 2 μ μ μ μ μ
2 2
λ λ λ
= 1 1 (U sin g Binomial Expansion)
μ μ μ
2
2
Lq
( )
1
Average number L w of customers in nonempty queues
LW E N 1 / N 1 0 , since the queue is non-empty
E(N 1) λ2 1 λ2 1 λ2 1
P(N 1 0) μ(μ λ) P(N 2) μ(μ λ) μ(μ λ)
n
λ λ
P n μ 1 μ
n 2
n 2
253
λ2 1
n 2
μ(μ λ) λ 2
λ λ
μ 1 μ μ
n 2
1
Lw
1
1 1
Probability that the number of customers in the system exceeds k
n
P ( N k ) Pn 1
n k 1 n k 1
λ λ λ
k 1 n (k 1) k 1 2
λ λ λ λ
1 1 1 ...
μ μ n k 1 μ μ μ μ μ
k 1 1 k 1
P( N K ) 1 1
Example 2
Arrivals at a telephone booth are considered to be Poisson with an average time
of 12 min between one arrival and the next. The length of a phone call is assumed to be
distributed exponentially with mean 4min. (i)Estimate the fraction of the day when the
phone will be in use. (ii) The telephone department will install a second booth, when
convinced that an arrival has to wait on the average for at least 3 min for phone. By how
much the flow of arrivals should increase in order to justify a second booth?(iii) What is
the average length of the queue that forms from time to time?
Solution:
Servers - Telephone booth – Single server
System capacity – Arrivals at a telephone booth – Infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival time = 12 minutes
1
Mean arrival rate / min
12
Mean arrival time = 4 minutes
1
Mean service rate / min
4
1/12
1/ 3
1/ 4
(i)P(Phone in use) = P(Phone is busy) = 1- P(Phone is in idle)
1
1 P0 1 1
3
(ii) The second phone will be installed if E(Wq ) 3
R
i.e. 3 , Where R is the required arrival rate.
( R )
R 3 ( R )
254
11 3 3 3 3 7 3 3 3
R 3 R R R R R R 7R R
44 16 4 4 16 4 16 4 28
3 1 1
Difference in the arrival rate is / minute
28 12 42
1
Hence the arrival rate should increase by / minute in order to justify a second booth.
42
(iii)Average length of queue that forms from time to time is Average length of queue in non empty
queue.
1/ 4 1/ 4 3
Lw
(1/ 4) (1/12) 2 /12 2
Example 3
Customers arrive at a sales counter manned by a single person according to a Poisson
process with mean rate of 20 per hour. The time required to serve a customer has an
exponential distribution with a mean of 100 seconds. Find the average waiting time of a
Customer.
Solution:
Servers - Single sales person – Single server
System capacity – customers – Infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 20 per hour
Mean service time = 100 seconds
1 1 1
Mean service rate / sec 60 / min 60 60 / hr 36 / hr
100 100 100
i) Average waiting time of customers in the system is given by
1 1 1
E Ws hour or 225sec onds
36 20 16
ii) Average waiting time of customers in the queue is given by
20 5
E Wq hour or 125sec onds
36 36 20 144
Example 4
Customers arriving at a watch repair shop according to Poisson process at a rate of one per
every 10 minutes and the service time is an exponential random variable with mean 8
minutes. (i)Find the average number of customers Ls in the shop.(ii)Find the average time a
customer spends in the shop Ws.(iii) Find the average number of customer in the queue L q.
(iv) What is the probability that the server is idle?
Solution
Single server - one repair man
System capacity – Customers - infinite
Model: M / M /1 : ( / FIFO)
Mean arrival time 1/10 customers per minute
Mean arrival rate 60 /10 per hr 6 / hr
Mean service time 8 customers per minute
255
1
Mean service rate 1/ 8 per min 60 per hr 7.5 / hr
8
6
0.8
7.5
i ) The average number of customers in the shop:
0.8
Ls 4
1 1 0.8
ii ) The average time a customer spends in the shop:
1 1 2
E Ws 0.6666 hour or 40 mins
7.5 6 3
iii ) The average number of customers in the queue :
By using Little’s formula we get,
Lq Ls 4 0.8 3.2 3
iv) Probability that the server is idle :
P(server is idle) = P(no customer in the system)= P0 1 1 0.8 0.2
Example 5
A T.V repairman finds that the time spent on his job has an exponential distribution with
mean 30 minutes. If he repairs cars in the order in which they come, which follow a Poisson
arrival pattern with average rate of 10 per 8 hour day. (i)What is the repairman’s expected
idle time each day? (ii) How many jobs are ahead of an average sets brought in?(iii) What is
the average number of jobs in a non-empty queue?
Solution
One T.V repair man ---- single channel
Arrival of T.V sets ---any number (infinite capacity)
Model: M / M /1 : ( / FIFO)
Mean Arrival rate 10 sets / day (8 hour )[Assume1day 8 hr]
1
Mean service time 30 min/ set hr
2
Mean Service rate 1/ (1/ 2) 2 / hr 2 8 / day 16 / day
10 5
16 8
5
(i) P (the repairman’s idle time) = P0 1 1 0.375
8
Expected idle time 8 hrs 8 x 0.375 3hrs
(ii) Average number of jobs ahead of an average set brought in:
5/8 5
Ls
1 1 (5 / 8) 3
(iii) The average number of jobs in a non-empty queue
16 16 8
Lw 2.667 jobs
16 10 6 3
256
Example 6
A telephone exchange receives one call every 4 min and connects one call every 3 min. if the
rate of arrival follows Poisson distribution and the service rate follows exponential
distribution, find (i) The average waiting time for a call in the queue(ii) The average waiting
time for a call in the queuing system.
Solution
Server - A telephone exchange-single server
System capacity - Number of calls received - infinite
Model: M / M /1 : ( / FIFO)
Mean arrival rate 1 call / 4 min ( convert to 1min)
1
/ min
4
Mean service rate : 1 call / 3 min
1
/ min
3
1/ 4 3
1/ 3 4
3/ 4
Ls 3
1 1 (3 / 4)
3
Lq Ls 3 2.25
4
(i) The average waiting time for a call in the queue:
Lq 2.25
Wq 9 min
(1/ 4)
(ii) The average waiting time for a call in the queueing system:
1 3
ws Ls 12 min
(1/ 4)
Example 7
Customers arrive at an office window, being manned by a single individual according to
Poisson input process with a mean rate of 30 per hour. The time required to serve a customer
has an exponential distribution with a mean of 90 seconds. Find the average waiting time of a
customer in the queue and the average number customer in the system.
Solution
Server - One customer arrive at office – single server
System capacity -Customers – infinite server
Model: (M/M/1) :( ∞/FIFO)
Mean arrival rate λ= 30/hr=30/60 per min = 1/2 per min
Mean service time = 90 seconds
Mean service rate µ = 1/90 x 60 = 2/3 mins
1/ 2 3
2/3 4
(i) Average number of customer in the system
257
3/ 4
E Ns 3
1 1 (3 / 4)
ii) Average waiting time of a customer in the queue
1 1 3
E Wq E N s 3 4.5
( ) 1 2
Example 8
Arrivals at a telephone booth are considered to be Poisson with an average time of 12 min.
between one arrival and the next. The length of a phone call is assumed to be distributed
exponentially with mean 4min. (i) Find the average number of persons waiting in the
system.(ii)What is the probability that a person arriving at the booth will have to wait in the
queue?(iii) What is the probability that it will take him more than 10min. altogether to
wait for the phone and complete his call?
Solution
Servers - A telephone booth
System capacity – Customers – infinite
Model: M / M /1 : ( / FIFO)
Mean inter arrival time 12 min s
1
Mean arrival rate per min
12
Mean service time 4 min s
1
Mean service rate per min
4
(i)Average number of persons waiting in the system
1 1 12
E Ws 6
(1/ 4) (1/12) 2
E N s E Ws (1/12) 6 0.5
(ii) Probability that a person arriving at the booth will have to wait in the queue =
λ 1/12 1
P(thesystem is busy) 1 P(thesystem is idle) 1 P0
μ 1/ 4 3
(iii) P (it will take more than 10 mins altogether to wait for the phone and complete his call) is
1 1 5
10
( ) t
P(WT t ) e e
P(WT 10) e e 3 0.1889
(μ λ)t 4 12
Example 9
Customers arrive at one- man barber shop according to a Poisson process with a mean inter
arrival of 12min. customers spend an average of 10 min. in the barber’s chair. (i) Calculate
the % of customers who have to wait prior to getting into the barber’s chair. (ii)What is the
probability that more than 3 customers are in the system? (iii) What is the average time
customers spend in the queue?
Solution:
Server : One man Barber shop- (Single server)
System capacity - Customer – Infinite
Model : M / M /1 : ( / FIFO)
Mean arrival rate: 1/12 per min
258
Mean service rate: 1/10 per min
λ 1/12
ρ = 5/6
μ 1/10
λ 5
(i)P (a Customer has to wait) = P(thesystem is busy) 1 P(thesystem is idle) 1 P0
μ 6
The percentage of customers who have to wait is (5 / 6) 100 83.33%
(ii) Probability more than 3 customers in the system = P(N > 3)
We know, P( N k ) k 1 P(N 3) ρ 31 (5 / 6)4 0.4823
(iii) Average waiting time per customer spend in the queue is
1 / 12 10 120
E Wq = = 50 min
1 1 1 12 2
-
10 10 12
Example 10
An airport has a single runway. Airplanes have been found to arrive at the rate of 15 per
hour. It is estimated that each landing takes 3 min. Assuming a Poisson process for arrivals
and an exponential distribution for landing times. Find the expected number of airplanes
waiting to land, expected waiting time. What is the probability that the waiting will be more
than 5 minutes?
Solution:
Servers – Runway – single Runway
System capacity – Air port – infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 15 per hr
1
Mean service rate 1/ 3 per min 60 per hr 20 per hr
3
15 3
20 4
(i) Expected number of Airplanes waiting to land:
15 15
Lq Ls where Ls 3
20 15 5
3 9
Lq 3 2.25
4 4
(iii) The Expected waiting time:
Lq 2.25 L 3
Wq 0.15 Ws s 0.2
15 15
(iii)The probability that the waiting time will be more than 5 minutes.
t
P(Ws t ) e
20155
P(Ws 5) e e25 0.4944
Example11
A concentrator receives messages from a group of terminals and transmits then over a single
transmission line. Suppose that messages arrives according to a Poisson process at a rate of
one message every 4 milliseconds and suppose that message transmission times are
259
exponentially distributed with mean 3 ms. Find the mean number of messages in the system
and the mean total delay in the system. What percentage increase in arrival rate results in a
doubling of the above mean total delay?
Solution:
Servers - A concentrator-Single server
System capacity – Messages- Infinite
Model: M / M /1 : ( / FIFO)
Mean total delay = Average waiting time in the system
Mean arrival rate 1/ 4 per ms & Mean service rate 1/ 3 per ms
Mean total delay in the system is
1 1
E Ws 12ms
(1/ 3) (1/ 4)
Double the mean waiting time = 2 *12 = 24
1 1 1 1 1 1 7
Let the new arrival rate be ,then 24
24 3 24 3 24 24
7 1 1
Increase in arrival rate
24 4 24
1/ 24 1
Rate of increase
1/ 4 6
1
Increase in arrival 100 16.6 17%
6
Example12
A duplicating machine maintained for office use is operated by an office assistant who earns
Rs. 5 per hour. The time to complete each job varies according to an exponential
distribution with mean 6 min. Assume a Poisson input with an average arrival rate of 5 jobs
per hour. If an 8 hour day is used as a base, determine (i) The % idle time of the machine (ii)
the average time a job is in the system and (iii) the average earning per day of the assistant.
Solution:
Server: Duplicating machine - Single server
Customers: jobs - Infinite
Model: M / M /1 : ( / FIFO)
Mean arrival rate λ = 5 / hour and Mean service time = 6 minutes
1 1
Mean service rate μ = / min 60 / hr 10 / hour
6 6
λ 5 1
(i) P(the machine is idle)=P(N=0)= P0 = 1 = 1
μ 10 2
Percentage of idle time of the machine = (1/2) x 100=50%
(ii) Average time a job in the system
1 1 1
E Ws = = h or 12 min
10 - 5 5
(iii)E(earning per day ) = E(number of jobs done/day) * earning per job
=E(number of jobs done/day) * E(time in
hour/job)*earning/hour
260
1
= (8 * 5) * * 5 =Rs.40.
5
E earning per day Rs 40
Example13
If people arrive to purchase cinema tickets at the average rate of 6 per minute, it takes an
average of 7.5 seconds to purchase a ticket. If a person arrives 2 min before the picture
starts and if he takes exactly 1.5 minutes to reach the correct seat after purchasing the
ticket, (i) Can he expect to be seated for the start of the picture? (ii) What is the probability
that he will be seated for the start of the picture? (iii) How early must he arrive in order to
be 99% sure of being seated for the start of the picture?
Solution:
Servers – Ticket counter - Single chennal
System capacity - Customers – Peoples -infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate: λ =6/minute
Mean service time =7.5 sec
Mean service rate = 1/7.5 /sec = (1/7.5) x 60 / min = 8/minute
(i) If a person arrives 2 min before the picture starts and if he takes exactly 1.5 minutes to reach
the correct seat after purchasing the ticket
E(total time required to purchase the ticket and to reach the seat) Ws = 1.5
1 1 1
Wq =
86 2
1
Ws = 1.5 =2 min
2
Hence he can just be seated for the start of the picture.
(ii) P(the person will be seated for the start of the picture) is
1 1
P(total time < 2 min) = P Ws 1 P Ws = 1- e-(8-6)(1/2) = 1 e 1 0.63
2 2
(iii) P W t = 0.99
1- P W t =0.99 e (μ λ)t = 0.01
Take log on both sides
(μ λ) t log(0.01) - 8 - 6 t = -4.605 -2t = -4.605
t = 2.3 mins
P (ticket purchasing time < 2.3) =0.99
P[total time to get the ticket and to go to the seat < (2.3+1.5)]=0.99
Therefore the person must arrive at least 3.8 min early so as to be 99 % sure of seeing the start of
the picture.
Example 14
Customers arrive at one- man barber shop according to a Poisson process with a mean inter
arrival of 20 min. Customers spend an average of 15 min. in the barber’s chair. The service
time is exponentially distributed. If an hour is used as a unit time, then (i) what is the
probability that a customer need not wait for a haircut? (ii) What is the expected number of
customer in the barber shop and in the queue? (iii) How much time can a customer expect to
261
spend in the barber shop? (iv) Find the average time that a customer spends in the queue. (v)
Estimate the fraction of the day that the customer will be idle? (vi) what is the probability
that there will be 6 or more Customers? (vii) Estimate the percentage of Customers who
have to wait prior to getting into the barber’s chair.
Solution:
Servers – One man barber shop – single server
System capacity – customers – infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival time = 20 mins
Mean service time = 15 mins
1 1
Mean arrival rate / min 60 / hr 3 / hr
20 20
1 1
Mean service rate / min 60 / hr 4 / hr
15 15
3
4
(i) Probability that a customer need not wait for a haircut is equal to probability that no customer
in the system. i.e
3 1
P0 1 1
4 4
(ii) Expected number of customer in the barber shop and in the queue:
E Ns
3
3
43
E N q E N s
3 9
3 2.25
4 4
(iii) Expected time a customer spends in the shop:
L
E Ws s 1 hr
3
3
(iv)Average time a customer spends in the queue:
E Nq
E Wq
2.25
0.75hr
3
(v) The fraction of the day that the customer will be idle
3 1
P0 1 = 1
4 4
(vi)Probability that there will be 6 or more Customers
n
We Know P ( N n)
6
3
P( N 6) 0.1779
4
(vii) The percentage of Customers who have to wait prior to getting into the barber’s chair.
1 3
P(W 0) 1 P(W 0) 1 P0 1
4 4
Example15
262
Customers arrive at the first-class ticket counter of a theatre at the rate of 12 per hour.
There is one clerk servicing the customers at the rate of 30 per hour. (i) What is the
probability that there is no customer at the counter? (ii) What is the probability that there
are more than 2 customers at the counter?(iii) What is the probability that there is no
customer waiting to be served?(iv) What is the probability that a customer is being served
and nobody is waiting? (v) Probability that a customer has to wait for at most 4 minutes in
the queue.
Solution:
Server - One clerk – single server
System capacity -Customers – infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 12 / hour
Mean service rate 30 / hour
12 2
30 5
2 3
P0 1 1 0.6
5 5
(i) Probability that there is no customers at the counter (i.e. in the system) is P0 0.6
(ii) Probability that there are more than 2 customers at the counter (i.e. in the system) is
3
2
P( N 2) (0.4)3 0.0664
2 1 3
5
(iii) Probability that there is no customer waiting to be served
= P [there is atmost one customer in the system]
P( N 1) P N 0 or N 1
P0 P1 Pn n P0
2
2
P0 P0 P0 1 1 1 1 2
1 0.84
5
P ( N 1) 0.84
(iv) Probability that a customer is being served and nobody is waiting is
2
P [there is exactly one customer in the system who is being served] P1 P0 (0.6) 0.24
5
(v) Probability that a customer has to wait for at most 4 minutes in the queue
1 4 1
P Wq 4mins = hrs = hrs
15 60 15
2 (3012)15
1 18
1
1 P Wq 1 e 1 (0.4)e 1 (0.4)e1.2 1 0.1205 0.8795
15
15 5
( ) t
P Wq t e
Example16
Customers arrive at one- man barber shop according to a Poisson process with a mean inter
arrival of 12min. customers spend an average of 10 min. in the barber’s chair.(i) How much
263
time can customer expect to spend in the barber’s shop? (ii) Management will provide
another chair and here another barber, when a customer’s waiting time in the shop exceeds
1.25h. (iii) How much must the average rate of arrivals increase to warrant a second barber?
(iv) What are the average time customers spend in the queue? (v) What is the probability
that the waiting time in the system is greater than 30 min?
Solution:
Server: One man Barber shop- Single server
System capacity - Customer – Infinite
Model: M / M /1 : ( / FIFO)
Mean arrival rate: 1/12 per min
Mean service rate: 1/10 per min
λ 1/12
ρ = 5/6
μ 1/10
Expected Number of customers in the barber shop
56 56
Ls 5
1 5 6 1 6
Expected number of customers in the queue
5 25
Lq Ls 5
6 6
i) Expected time a customer spend time in barber’s shop
L
E Ws s
1
5 60 mins
1 12
ii) Average rate of arrivals increase to warrant a second barber
1.25hr 1.25 60 75min s
R is the required arrival rate
1 1 1 1 1 1
Given 75 R R R R
R 75 75 75 10 75
13
R
150
13 1
R 0.003
150 12
Hence to warrant a second barber, the average arrival rate must increase by 0.003
iii) Probability that the waiting time in the system > 30 mins
i.e., P[W > 30]
t
P(Ws t ) e
1 1
30
P[Ws 30] e e 0.5 0.6065
10 12
P[W 30] 0.6065
P( a customer walk straight into the barbers chair without having to wait ) is
1
P (No customer in the system) = P0 = 1- ρ = 1- 5 / 6 = = 0.1667
6
Example17
In a given M/M/1 queueing system the average arrivals is 4 customers per minute and
264
= 0.7. Calculate the following. (i) Mean number of customer Ls in the system (ii) Mean
number of customers Lq in the system. (iii) Probability that the server is idle. (iv) Mean
waiting time Ws in the system.
Solution:
Given Mean arrival rate λ = 4/min
= 0.7
4 4
i.e., 0.7 0.7 5.714
0.7
(i)Mean number of customers in the system
0.7 0.7 7
Ls 2.333
1 1 0.7 0.3 3
(ii)Mean number of customers in the system
Lq Ls 2.333 0.7 1.633
(iii)Probability that the server is idle
P0 1 1 0.7 0.3
(iv)Mean waiting time in the system
1 1
Ws Ls 2.333 0.583
4
Example18
Customers arrive at the express checkout lane in a supermarket in a Poisson process with
rate of 15 per hour. The time to check out a customer is an exponential random variable
with the mean of 2 minutes. Find the average number of customers present. What is the
expected waiting time for a customer in the system?
Solution:
Servers - A express checkout lane – Single server
System capacity – Customers – infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 15 per hr
1 60
Mean service rate / min / hr 30 per hr
2 2
i) The average number of customers present
15 15
Ls 1
30 15 15
ii) Expected waiting time for a customer in the system
L 1
Ws s hr or 4 min s
15
Example19
Arrivals at a telephone booth are to be Possion with an average time of ten min between one
arrival and the next. The length of a phone call is assumed to be distributed exponentially
with mean 3 min. (i) what is the probability that an arrival will have to wait for more than
10 min before the phone is free. (ii) What is the probability that he will have to wait for more
than 10 min before the phone is available and is the call is also complete. (iii)Find the
fraction of a day that the phone will be in use.
Solution:
265
Servers – A telephone booth - Single server
System capacity – arrivals – Infinite capacity
Model: M / M /1 : ( / FIFO)
Mean arrival rate 1/10 per min,
Mean service rate 1/ 3 per min
1/10 3
0.3min
1/ 3 10
(i)P (an arrival will have to wait for more than 10 minutes before the phone is free)
1 1 7
10
P(Wq 10) e ( )10 (0.3)e 3 10 (0.3)e 3 0.029
(ii) P (A customer will have to wait for more than 10 mins before the phone is available and the
call is also complete)
7
( )10
P (Ws 10) e e 0.97
3
(iii)Fraction of the day the phone will be in use
1
3
P(the phoneis busy) 1 P(the phoneis idle) 1 P0 10 0.3min
1/ 3 10
Example20
Arrivals at a telephone booth are considered to be Poisson with a mean inter arrival time of
12 min. The length of a phone call assumed to be distributed exponentially with mean 10
min. (i) what is the expected number of arrivals in the telephone booth and in the queue? (ii)
Calculate the percentage of time an arrival can walk straight in to the telephone booth
without having to wait. (iii) How much time can an arrival expect to spend in the telephone
booth? (iv) What is the probability that the waiting time in the system is greater than 30
min?
Solution:
Server: Telephone booth- Single server
System capacity - Customer – Infinite
Model: M / M /1 : ( / FIFO)
Mean arrival rate: 1/12 per min
Mean service rate: 1/10 per min
λ 1/12
ρ = 5/6
μ 1/10
Expected Number of customers in the Telephone booth
56 56
Ls 5
1 5 6 1 6
Expected number of customers in the queue
5 25
Lq Ls 5
6 6
P(an arrival can walk straight in to the telephone booth without having to wait) =
1
P (No customer in the system) = P0 = 1- ρ = 1- 5 / 6 = = 0.1667
6
266
Expected time a customer spend time in the Telephone booth
L 5
Ws s 60 mins
1 12
Probability that the waiting time in the system > 30 mins
i.e., P[W > 30]
1 1
t
e 0.5 0.6065
30
P[W 30] e 10 12 P (Ws t ) e
Example21
In a city Airport, flights arrive at a rate of 24 flights per day. It is known that the inter-
arrival time follws an exponential distribution and the service time is also exponential with
an average of 30 min. find the following: (i) The probability that the system will be idle. (ii)
The mean queue size (iii) The average number of flights in the queue (iv) The probability
that the queue size exceeds 7.
Solution:
City Airport – one server – Single server
System capacity - infinite capacity
Model: M / M /1 : ( / FIFO)
Mean Arrival rate 24 / day 1/ hr
Mean Service time = 30 min
Mean service rate: μ = 1/ 30 per min 1/ 30 60 per hr 2 / hr
λ 1
ρ
μ 2
i) The probability that the system will be idle:
1
P0 1 1 0.5
2
(ii) The mean queue size:
1
Ls 1 flight
2 1
(iii) The average number of flights in the queue:
1
Lq 0.5
( ) 2(2 1)
(iv) The probability that the queue size exceeds k is:
P( N k ) k 1
Therefore the probability that the queue size exceeds 7 is
7 1
1
P( N 7) 0.0039
2
Example22
At what average rate must a clerk in a supermarket work in order to ensure a probability of 0.90
that the customer will not wait longer than 12 min? It is assumed that there is only one counter at
267
which customers arrive in a Poisson fashion at an average rate of 15 per hour and that the length of
the service by the clerk has an exponential distribution.
Solution:
Servers – A clerk – single server
System capacity – customers – infinite
Model: M / M /1 : ( / FIFO)
Mean arrival rate 15 / hr Let R / hr
Let the required service rate be
12 min = 12/60 per hr = 1/5 per hr.
P Wq 12 min P Wq 1/ 5hr 1 P Wq 1/ 5hr 1 0.90 0.10
P Wq 1/ 5hr 0.10
g (w)dw 0.10
0.2
λ
μ μ λ e dw 0.10 g(ω)is the probability of waiting time of a customer in thesystem
μ λ w
0.2
e w w
e 0.1
0.2 0.2
15
R
e
15 0.2
0.1
15 e 15 0.2
0.1
15 (0.2) 0.2 3
e 0.10 e 0.10 e 0.2 3 0.10 e 0.2 3
15 15 15 150
Take log on both sides
log e 0.2 3 log
150
(0.2 3) log e log log150
0.2 3 log log150
log 0.2 3 log150
log 0.2 3 2.1760 3 5.1760
log 0.2 8.0106
By trial and error 24
That is the clerk should, on an average, serve 24 customers per hour.
Exercise 4.2
1. At one man barber shop, customers arrive according to Poisson distribution with a mean arrival
rate of 5 per hour and his hair cutting time is exponentially distributed with an average hair cut
taking 10 minutes. It is assumed that because of his excellent reputation, customers are always
willing to wait. Calculate the following (i) Average number of customers in the shop and average
number of customers waiting for a hair cut. (ii) The percentage of time an arrival can walk right in
without having to wait. (iii) The percentage of customers who have to wait prior to getting into the
barber’s chair. Ans: (i) 5 and 4.1667 (ii)16.67% (iii)83.3%
268
2. A departmental secretary receives an average of 8 jobs per hour. Many are short jobs, while
others are quite long. Assume, however , that the time to perform a job has an exponential
distribution with mean of 6 min. (i) what is the average elapsed time from the time the secretary
receives a job until it is complicated? (ii) Calculate E(N), E Wq , P(W 2h), P(N 5) and the
fraction of time the secretary is busy. Ans (i) 30 min (ii) 4, 24 min, 0.0183, 0.2621, 0.8.
3. A duplicating machine in an office is used by people in the office who make copies, mostly by
secretaries. The service rates follows Poisson distribution with a mean service rate of 10 jobs /
hour. Requirements for use are random over the 8 – hour work day, but arrive at a rate 5 / hour. If
the time of a secretary is valued at Rs 3.50 / hour, determine (i) Equipment utilization (ii) The
percentage of time that an arrival has to wait. (iii) The average system time (iv) Average cost due
to waiting and operating the machine. Ans : (i) 0.5 (ii) 50% (iii) 0.2 hour (iv) Rs 0.7 / job and Rs
28 / day
4. The capacity of a communication line is 2000 bits per second. The line is used to transmit eight-
bit characters, so the maximum rate is 250 characters per second. The application calls for traffic
from many devices to be sent on the line with a total volume of 12000charecters per minute. Find
(i) Line utilization (ii) Average number of characters waiting to be transmitted (iii) Average
transmission time per character, including queueing delay. Ans: (i) 0.8 (ii) 3.2 (iii) 0.02
5. A self - service store employs one cashier at its counter. Eight customers arrive on an average
every 5 minutes while the cashier can serve 10 customers in the same time. Assuming poisson
distribution for arrival and exponential distribution for service rate, determine (i) Average number
of customer in the system (ii) Average number of customers in queue or average queue length (iii)
Average time a customer a spends in the system (iv) Average time a customer waits before being
served. Ans: (i) 4 (ii) 3.2 (iii) 2.5 (iv) 2 min
4.3 MULTIPLE SERVER QUEUEING MODEL WITH INFINITE CAPACITY
[(M/M/C): ( /FIFO)]
Introduction
Consider a queue with Poisson input and with s, (1 c ) parallel service channels having
independent and identically distributed with exponential service time distribution, each with rate
say . If there are n units in the system, and n is less than s, then, in all, n channels are busy and
the interval between two consecutive service completions is again exponential with rate n . If
there are n (n > c) in the system, then all the c channels are busy and the interval between two
consecutive service completions is exponential with rate c . Thus, we have a birth-death model
having constant arrival (birth) rate and state-dependent service rate.
n , n 0,1, 2,..., c
n Denote the traffic intensity / c
c , n c 1, c 2,...
Theorem 4.3
Derive the steady state probability for the [(M/M/C): ( /FIFO)] model.
Solution
From Theorem 4.1, we know
...
Pn 0 1 2 n 1 P0 n 1, 2,.. (1)
1 2 3 ... n
1
P0 2
0 12 3 .......n 1
1
n 1 1 2 3 ....... n
269
For the [(M/M/S): ( /FIFO)] model,
for n 0,1,...k 1
n
0 for n k , k 1,...
n for n 0,1,...c 1
n
c for n c,
Using these values of λn and µn in (2)
1
P0
0 0 1 012 ...
.......n 1
1 .... 0 1 2 s 1 0 1 2 3
1 12 12 3 12 3 ... s n c 12 3 ....... n
1 2 c 1
P0 1 ...
1! 2! (c 1)! c 1
2
c c 1 k
... k c 1
(c 1)! c (c 1)! c (c 1)! c 1 c
c 1 c 1 2
1
c 1
c
n
2
k c
1 ...
n0 n ! c! c c c c
1 1
c 1 1 n 1 c k c 1 1 n 1
c
P0 (or ) P0
n c
n 0 n ! c! n c n 0 n ! c!1
Using these values of λn and µn in (1), we get
1 n
P0 , nc
n!
Pn n
1
c!c n c P0 , n c
Some Important characteristics for [(M/M/S): ( /FIFO)] model
Average number of customer in the system
c
c
E N s Ls P0 c , where , c number of servers
1 c! c
2
By using the above we can find other important characteristics by using the following little
formula
c 1
P0
1
E N q Lq Ls
c.c! 1 2
Ls
E Ws
Lq
E Wq
270
The probability that an arrival has to wait
c
P0
1
P( N c)
c! 1
Probability that an arrival enters the service without waiting .
c
P0
1
1
c! 1
Mean waiting time of a customer in the queue for those who actually waits
E Wq / Ws 0
1
c
Average number of customers (in the nonempty queue ) who have to wait actually
Lw
1
Probability of waiting time of a customer exceed t
c t c 1
1 e P0
P Ws t e t 1
c!1 c 1
Worked Examples 4.3(A)
Example 1
Draw the state transition diagram of an M/M/S queuing model.
Solution
0 1 2
n-1 n n+1
2 n (n+1)
Example 2
Define M/M/2 queuing model why the notation M is used?
Solution
271
M/M/2 is the two server Poisson queue model, where the arrival rate follows Poisson distribution
and the service rate follows Exponential distribution. This model also called multi server queueing
model with infinite capacity.
The notation M stands for Markov which means the arrival and service follows Markov property.
Example3
Consider an M/M/C queueing system. Find the probability that an arriving customer is
forced to join the queue.
Solution
The probability that an arrival is forced to join in the queue
= the probability that an arrival has to wait
= probability that there are s or more customers in the system is P ( N c ) Pn
n c
n c n c
1 1
P( N c) n c P0 P0 Let C c
n c c!c c! n c c
1
c
2 1 c
1
P0 1 ...... P0 1
c! c c c! c
c c
P0 P0
1 1
P( N c)
c! c! 1
1 c
Example 4
Find PX c + n for an M/M/C queuing system.
Solution
We know that
n
P[ X c n] Pn P ( N c )
n c
c
c
c
0 0
P P P0
P( N c) 1
c ! 1 c! c! 1
c c
P0
P0
1
1
c! 1 c! 1
1
c 1 1 n 1 c k
where, P0 &
n c
n 0 n ! c! n c c
Example 5
What is the probability that an arrival to an infinite capacity 3 server Poisson queue with
2 1
and p 0 enters the service without waiting?
c 3 9
272
Solution
Servers - 3 servers c = 3
System capacity – infinite capacity
Model : M / M / C : ( / FIFO)
λ 2 1
Given ρ and p0
cμ 3 9
2
2
3 3
The probability that an arrival enters the service without waiting is P ( N c)
1 / P0
c
P( N c) 1 P( N c) 1
c! 1
23 1/ 9 8/9 4 5
P( N c) 1 1 1
3!1 (2 / 3) 6(1/ 3) 9 9
Probability that an arrival enters the service without waiting = 5/9
Example 6
What is the steady state condition for M/M/C queueing model?
Solution
Steady state conditions are
1 n
P0 , nc
n!
Pn n
1
c!c n c P0 , n c
For n = 0
1 1
c 1 1 n 1 c k n c c 1 1 n 1
c
P0 (or ) P0
n 0 n ! c! n c c n 0 n ! c!1
Example 7
There are three typist in an office. Each typist can type an average of 6 letters per hour. If
letters arrive for being typed at the rate of 15 letters per hour. What fraction of time all
typist will be busy?
Solution
Servers - Typists - Multiple server
System capacity – Infinite capacity
Model : M / M / C : ( / FIFO)
Mean arrival rate 15 / hour
Mean service rate 6 / hour
Number of server c = 3
15 15
2.5 0.83
6 c 3 6
273
1 1
c 1 1 n
c
2 1 1 2
2.5
1
P0 2.5
n
n 0 n ! c!1 n 0 n ! 3!1 0.83
1
2.5 2.52
6.127 6.625 6.127 0.0784
1
1
1! 2!
c
P0
1
P (all typists are busy) = P( N c)
c! 1
1 2.5 0.0784
3
P( N 3) 1.201
3! 1 0.83
Hence, the fraction of time all the typist will be busy = 1.201
Example 8
If there are two servers in an infinite capacity Poisson queue system with
λ 10 / hr & μ 15 / hr , What is the percentage of idle time for each server?
Solution
Servers - 2 servers
System capacity – infinite capacity
Model : M / M / C : ( / FIFO)
Mean arrival rate 10 / hour
Mean service rate 15 / hour
10
c = 2, 15 0.67
10
0.33
c 2 15
1 1
c 1 1 n
c
1 1 1 2
0.67
1
P0 0.67
n
n 0 n ! c!1 n 0 n ! 2!1 0.33
1 0.67 0.335 1.67 0.335
1
2.005 0.4988 0.5
1 1
P0 0.5
1
The Percentage of Idle time for each server 100 50%
2
Example 9
A two-server queueing system is in a steady-state condition and the steady state probabilities are
1 4 6 4 1
p0 , p1 , p2 , p3 , p4 , pn 0 if n 4 . Calculate the number of customers
16 16 16 16 16
in the system and in the queue. (GATE)
Solution
274
k 4
The number of customers in the system Ls n Pn n Pn 0 1P1 2 P2 3P3 4 P4
n0 n 0
4 1
(2) (3) (4) 0.25 0.75 0.75 0.25 2
6 4
16 16 16 16
The number of customers in the queue Lq (n 1) Pn
n 1
1
n Pn Pn Lq 2 1 P0 1 1.06
n 1 n 1 16
Example 10
Management of a shop plans to open during Sundays. The shop will have three cash registers with
an expected service time 1 minute, projected mean arrival rate is expected to be 1.2 customers per
minute. Compute the average queue length. (GATE)
Solution
First, = 1.2/min, µ = 1/min, c = 3
1.2 1.2
1.2 The system utilization 0.4 , 1 1 0.4 0.6
1 c 3
1 1
c 1 1 n 1
c
2 1 1 3
. P0
n
1.2
3! 0.6 1.2
n 0 n ! c!1 n 0 n !
1
1
1 1.2 (1.2) 2 0.48 0.29
2
c 1
P0
1 1.2 (0.29)
31
1
L
The average queue length at each counter is q 0.8
c.c! 1 2 3.3! 0.6
2
Worked Examples 4.3(B)
Example 1
Obtain the expression for the average number of customers in the queue, mean number of
customers in the system and waiting time of a customer in the system and in the queue
,Probability that an arrival has to wait for an (M/M/S) : (FCFS/ / ) queueing system .
Solution
Model : M / M / C : ( / FIFO)
The average number of customers in the queue
Lq E[ N q ] ( n c )Pn
ns
xPx c Put x n c, n x c
x 0
xc c x
1 1
x x P0 P0 x
x 0 c!c c! x 0 c
275
1
c
2
3
P0 2 3 ....
c! c c c
1 c
c 2 2
1
P0 1 2 3 .... P0 1
c! c c c c! c c
c 1 c 1
1 1
Lq P P0
cc !
2 0
cc ! 1 2
1 c
Average number of customers in the system:
k
Ls n Pn
n0
Using Little’s formula we get,
c
c
E N s Ls Lq P0 c , where , c number of servers
1 c!
2
c
Waiting time of a customer in the system
E Ns c
s
c
E Ws P0
1
2
c!
Waiting time of a customer in the queue
c 1
Lq 1
E Wq P
cc ! 1 2 0
Probability that an arrival has to wait (ie busy system) P(Ws 0)
P(Ws 0) P( N c) Pn
n c
n
P0 c n c
1
P0
c! n c c
n c
n c c!c
1
c
2
P0 1 ....
c! c c
c 1
1
P0 1
c! c
c c
P0 P0
P( N c)
c!1
c! 1
c
276
Example 2
Four counters are being run on the frontiers of the country to check the passports of the
Tourists. The Tourist choose a counter at random. If the arrival at the frontier is Poisson at
the rate and the service time is exponential with parameter λ / 2 , find the average queue
length at each counter.
Solution
Servers - Four counters – Multiple server
System capacity – Tourists, passports – Infinite capacity
Model : M / M / C : ( / FIFO)
Mean arrival rate
Mean service rate
2
2 1
2, c 4 & 0.5
c 4 2
2
1 1 0.5 0.5
1
c 1 1 n 1
c
P0
n 0 n ! c!1
1
3 1 1 4
2 2
n
n 0 n ! 4! 0.5
1
2 22 23
1 1.33
1! 2! 3!
6.33 1.33 7.66 0.13
1 1
The average queue length at each counter is
c 1
P0
1 1 2 0.13 4.16
5
Lq 0.173
c.c! 1 2 4 4! 1 0.52 24
Example 3
A telephone exchange has 2 long distance operators. The telephone company finds that
during the peak load, long distance call arrive in a poisson passion at an average rate of 15
per hour. The length of service on these calls is approximately exponentially distributed with
mean length 5 min. (i) what is the probability that a subscriber will have to wait for his long
distance call during the peak hours of the day? (ii) If the subscribers will wait and are
serviced in turn, what is the expected waiting time?
Solution
Two operators-----multiple servers
Number of calls received (infinity) ------infinite capacity
Model : M / M / C : ( / FIFO)
c=2
Mean arrival rate: 15 per hr (convert to minute)
277
15 1
/ min / min
60 4
Mean service time: 5 min
1
Mean service rate
/ min,
5
1/ 4 5 1/ 4 5
1.25 & 0.625
1/ 5 4 c 2 / 5 8
1
c 1 1 n 1
c
P0
n 0 n ! c!1
1
1 1 n 1 2
P0 1.25 1.25
n 0 n ! 2!1 0.625
1
1.5625
4.3333 0.231
1
P0 1 1.25
0.75
(i) Probability that a subscriber will have to wait for his long distance call during the peak hours
of the day i.e P(a customer waits) is
c
P0
1
P( N c)
c ! 1
1.25 (0.231) 0.3609 0.4812
2
Therefore P(N≥ 2) =
2! 1 0.625 0.75
(ii) If the subscribers will wait and are serviced in turn, then the expected waiting time is
E Wq / Ws 0
1 1 1
4.06
c (2 / 5) 0.25 0.4962 0.25
Example 4
A petrol pump station has 4 pumps. The service times follow the exponential distribution
with a mean of 6 min. and cars arrive for service in a poisson process at the rate of 30 cars
per hour. (i) What is the probability that an arrival would have to wait in line?(ii) Find the
average waiting time, average time spent in the system and the average number of cars in
the system. (iii) For what % of time would a pump be idle on an average?
Solution
Servers - Petrol pumps----- Multiple server(s)
System capacity - Cars ----- Infinite capacity (∞)
Model M / M / C : ( / FIFO)
c= 4
Mean arrival rate λ =30 / hour
Mean service time =6 minutes
Mean service rate µ = 1/6 per minute = (1/6)*60 per hour =10 per hour
30
3
10
278
30 3
0.75 1 0.25
c 40 4
1
c 1 1 n 1
c
3 1 34
1
P0 3
n
n 0 n ! c!1 n 0 n ! 4!(0.25)
1
31 32 33
1 13.5 13 13.5 26.5 0.038
1 1
1! 2! 3!
c
c
0.75 34
E N s Ls P c (0.038) 3 12 0.128 3 4.536
1 c!
2 0
(0.25)2 4!
L
E N q Lq Ls 4.536 3 1.536 & E Ws s 2(4.53) 9.06 min
Lq
Wq 2(1.536) 3.072 min
(i) The probability that an arrival has to wait = The probability that there are s or more customers
in the system.
c
P0
1 1 3 0.0377 3.0537
4
P( N c) P N 4 0.509
c! 1 4! 0.25 6
ii) The average waiting time in the queue is 3.072 min
The average time spent in the system is 9.06 min
The average number of cars ion the system is 4.53 cars (approximately 5 cars)
3
(iii) The fraction of time when the pumps are busy = traffic intensity = 0.75
c 4
The fraction of time when the pumps are idle = 1-(3/4) =1/4
Hence the required % = 25%
Example 5
Consider a bank with two tellers. An average of 80 customers per hour arrives at the bank
and waits in the single line for an idle teller. The average time it takes to serve a customer is
1.2 minutes. Assume that the inter-arrival times and the service times are exponential.
Determine (i) The expected number of customers present in the bank. (ii) The expected
length of time a customer spends in the bank. (iii) The fraction of time that particular teller
is idle.
Solution
Server: 2 tellers (Multiple)
System capacity - Customers - Infinite
Model : M / M / C : ( / FIFO)
Number of servers c = 2
80
Mean arrival rate 80 / hr / min 1.33 / min
60
Mean service rate 1.2 /min
279
1.33 1.108
1.108 0.55
1.2 c 2
The probability of no customer in the system
1 1
c 1 1 n
c
1 1 1 2
1
P0
n
1.108
2!1 0.55 1.108
n 0 n ! c!1 n 0 n !
1
1
1 1.108 1.227
2(0.45)
2.108 1.3633 3.4713 0.288
1 1
P0 0.288
(i) The average number of customers in the bank is
Ls Lq
c 1
P0
1 1.108
2 1
1 1 1.360
2
where, Lq 0.288 0.288 0.483
c.c! 1 2
2.2! 1 0.55 4 0.2025
Ls Lq 0.4833 1.108 1.591
(ii) The average waiting time of customers is
1
Ws Wq
Lq 0.483
where, Wq = 0.3631
1.33
1
Ws 0.3631 1.1961
1.2
(iii) The fraction of idle time is
1
c 1 1 n 1
c
P0 = 0.288
n 0 n ! c!1
Example 6
A bank has 2 tellers working on savings accounts. The first teller handles withdrawals only.
The second teller handles deposits only. It has been found that the service time distributions
for both deposits and withdrawals are exponential with mean service time of 3 min per
customers. Depositors are found to arrive in a Poisson fashion throughout the day with
mean arrival rate of 16 per hour. Withdrawals also arrive in a Poisson fashion with mean
arrival rate of 14 per hour. What would be the effect on the average waiting time for the
customers if each teller could handle both withdrawals and deposits? What could be the
effect, if this could only be accomplished by increasing the service time to 3.5 min?
Solution
Server : 2 teller (Multiple)
280
System capacity - Customers - Infinite
Model : M / M / C : ( / FIFO)
Consider both the tellers serve separately. Assume c =1
Channel – I
When there is a separate channel for the depositors
Mean arrival rate λ1 = 16/hour
60
Mean service rate μ = 1/ 3per min = / hr = 20 / hr
3
Model M / M /1 : ( / FIFO)
To find Wq :
1 16 1
E( Wq for depositors) = = h or 12 min-------(1)
( 1 ) 20(20 16) 5
When there is a separate channel for the withdrawers,
Channel – II
Mean arrival rate λ 2 = 14/hour
60
Mean service rate μ = 1/ 3per min = / hr = 20 / hr
3
Model M / M /1 : ( / FIFO)
To find Wq :
2 14 7
E( Wq for withdrawers) = = h or 7 min------(2)
( 2 ) 20(20 14) 60
i ) If both tellers do both service, c = 2, μ = 20/hour, λ λ1 λ 2 =30/hour
To find Wq :
c
P0
1 1 1 1 (1.5) 2
E ( Wq for any customer) = = P0 = 0.45 P0
cc ! 1 2 20 2 2 (1 .75) 2
1 1
c 1 1 n 1
c
(1.5) 2
Where P0 = 1 1.5 2 0.25 1/ 7
n 0 n ! c!1
1
E ( Wq for any customer) = 0.45* h or 3.86 min---------(3)
7
From (1), (2) and (3) we conclude, if both tellers do both types of service, the customers get
benefited as their waiting time is considerably reduced.
ii) Now if both tellers do both types of service but with increased service time,
60 120
c=2, λ 30,μ =17.14 per hour
3.5 7
To find Wq :
7 1 (1.75) 2
E ( Wq for any customer)= p0 2.86 p 0 where
120 2 2 7 2
1
8
281
1
(1.75) 2 1 2.86
P0 = 1 1.75 E ( Wq for any customer) = h or 11.44 min
1 15 15
2
8
If this arrangement is adopted, withdrawers stand to lose as their waiting time is increased
considerably and depositors get slightly benefited.
Example 7
On every Sunday morning, patients arrive to a Dental clinic. As per the hospital rules 3
dentists who are equally qualified and experienced will be on duty. It takes on an average 10
min for a patient to get treatment and the actual time taken is known to vary approximately
exponentially around this average. The patients arrive according to the Poisson distribution
with an average of 12 per hour. The hospital management wants to investigate the
following. (i) The expected number of patients waiting in the queue. (ii)The average time
that a patient spends at the hospital.
Solution
Servers - 3 dentists – multiple server
System capacity – patients – infinite capacity
Model : M / M / C : ( / FIFO)
Mean arrival rate 12 /hour
Mean service rate 6/hour
Number of servers c = 3
12 12 2
2 and
6 c 18 3
The probability of no customer in the system is
1
1
c 1 1 n 1
c 2 1 1
P0 2 2
n 3
n 0 n ! c!1 n 0 n ! 3!1 2
3
1
2 1 1 22 8
1
2 2 1 2
n 3
n 0 n ! 6 1 2 2
3
(i) The average number of customers in the system is
s 1
P0
1 2
31
1 1 1(2) 4 1 16 8
Lq 2
0.8889
s.s ! 1 2 3.3! 2 2 9 1 9 18 9
1 (18)
3 3
1
1 2 2 4 9
1 1
9
(ii) The average waiting time of customers in the queue is
282
Lq 0.8889
E Wq 0.0740
12
(iii) The average waiting time of customers in the system is
1
E Ws E Wq
1
0.0740 0.2407 hour or 14.44 min s
6
Example 8
A supermarket has 2 girls attending to sales at the counters. If the service time for each
customer is exponential with mean 4 min. and if people arrive in Poisson (i) Fashion at the rate
of 10perhour, (ii) What is the probability that a customer has to wait for service? (iii) What is
the expected percentage of idle time for each girl? (iv) If the customer has to wait in the queue,
what is the expected length of his waiting time?
Solution
Servers: Two girls - (multiple servers)
System capacity - Customers: infinite capacity
Model : M / M / C : ( / FIFO)
Mean arrival rate: 10 / hr
Mean service time = 4 mins
1 60
Mean service rate / min / hr 15 / hr
4 4
Number of servers c = 2
10 2 10
& 1/ 3
15 3 c 2(15)
1
c 1 1 n 1
c
P0
n 0 n ! c!1
1 1
21 1 2 n 1 10
2 1
1 2
n
1 2
2
P0
n 0 n ! 3 2!1 1 15 n 0 n ! 3 2! 2 3
3 3
1 1 1
1 1 10 n 3 4 1 1 10 n 1 1 10 1 1 2 1
1
1 1
n 0 n ! 15 4 9 n 0 n ! 15 3 1! 15 3 3 3
21 1/ 2
(i) Probability that a customer has to wait for service
c
P0
1
P( N c)
c! 1
283
2
2
2
P0 1/ 2 2 / 9
P( N 2) 3 1/ 6
2!1 2!1 (1/ 3) 4 / 3
(ii) Fraction of time that a girl is idle = 1 - Fraction of time that a girl is Busy
10 1 2
1 1 1
c 2(15) 3 3
2
Percentage of idle time for each girl 100 67%
3
iii) Waiting time of a customer who has to wait
E Wq / Ws 0
1 1 1
hr 3min
c 2(15) 10 20
Example 9
There are 3 typists in an office. Each typist can type an average of 6 letters per hour. If
letters arrive for being typed at the rate of 15 letters per hour. (i) What fraction of the time
all the typists will be busy? (ii) What is the average number of letters waiting to be typed?
(iii) What is the average time a letter has to spend for waiting and for being typed? (iv)
What is the probability that a letter will take longer than 20 min waiting to be typed and
being typed?
Solution
Servers – 3 typists - (multiple)
System capacity - Customers – Letters(infinite capacity)
Model : M / M / C : ( / FIFO)
Mean arrival rate: 15 / hr
Mean service rate: 6 / hr
Number of servers c = 3
15 5 15 5
2.5 & 0.8333
6 2 s 3(6) 6
1 1
c 1 1 n 1
c
31 1 1 3
P0 2.5 2.5
n
n 0 n ! c!1 n 0 n ! 3!1 0.8333
1 1
2 1 1
15.625 1 2.5 (2.5) 15.3186 22.25 0.0449
1
2.5
n 2 1
n 0 n ! 1.02 2
P0 0.0449
(i) P (all the typists are busy) P( N 3)
s
P0
1
We know P ( N s )
s ! 1
284
3
P0 15.625 0.0449 0.7016
P( N 3) 0.7002
3!1 1.002 1.002
Hence, the fraction of the time all the typists will be busy = 0.70
(ii) Average number of letters waiting to typed is E N q
s 1
P0
1 1 2.5 0.0449
4
E N q Lq 3.5078
s.s ! 1 2 3.6 1 0.833332
(iii) Average time a letter has to spend for waiting and for being typed
L 1 1
E Ws s Lq 3.5078 2.5 0.4005 hr 24 min nearly
15
(iv) P(waiting time of a letter in the system is exceed t min)
c t c 1
1 e
P0
P Ws t e t 1
c !1 c 1
Given t = 20 min =20/60 per hr = 1/3 hr
P (waiting time of a letter in the system is exceed 20 min)
2.5 3 1 e 2( 0.5) 0.0449
6(1/3)
P Ws t = P Ws 20 = P Ws 1 / 3 e 1
6 1 0.03333 0.5
0.7016(1 e)
e 2 1
0.5
P Ws 1/ 3 0.4616
Example 10
A tax consulting firm has 3 counters in its office to receive the people who have problems
concerning their income and sales taxes. On the average 48 persons arrive in an 8 hour day.
Each tax advisor spends 15 min on an average and on an arrival. If the arrivals are poisson
distributed and service time are according to exponential, find the following. (i) Average
waiting time for a customer. (ii) The number of hours each week a tax advisor spends
performing his job.
Solution
Servers - 3 counters – multiple server
System capacity - Arrival of customers – Infinite
Model : M / M / C : ( / FIFO)
Mean arrival rate λ = 48/8 =6/ hr
Mean service time = 15 mins
Mean service rate µ = 1/15per min=(1/15) x 60 = 4 / hr
Number of Channels c = 3
285
6 1.5
1.5 & 0.5
4 c c 3
1 1
c 1 1 n 1
c
2 1 1 3
P0
n
1.5
3!1 0.5 1.5
n 0 n ! c!1 n 0 n !
1
1.5 1
3
2
1 1
1 1.5 1.5 1 1.5 1.125 3.375
2! 6 1 0.5 3
3.625 1.125 4.75 0.21
1 1
P0 0.21
Number of customers waiting in the queue is
s 1
P0
1 1 1.5 0.21
4
Lq 0.2363
s.s ! 1 2 3.3! 1 0.52
(i) Average waiting time for a customer
1
E Wq Lq
0.2363 0.0394hr 2.364 min s
6
(ii) The number of hrs each week a tax advisor spends performing his job utilization factor is
6 1
0.5 hr
c 3 4 2
Number of hours each day a tax advisor spends doing his job = 8 x 0.5= 4 hrs.
On the average he is busy 4 x 6=24 hrs in a week of 6 working days.
Example11
A petrol pump station has 2 pumps. The Service times follow the exponential distribution
with mean of 4 minutes and cars arrive for service is a Poisson process at the rate of 10 cars
per hour. Find the probability that a customer has to wait for service. What is the
probability that the pumps remain idle?
Solution
Servers - 2 pumps – multiple server
System capacity – Infinite capacity
Model : M / M / C : ( / FIFO)
1
Mean service time = min s
4
Mean arrival rate 10 / hr
60
Mean service rate 15 / hr
4
c =2
10 2 2 1 1
&
15 3 c 3 2 3
286
1
1
c 1 1 n
c 1 1 2 n 2
1 1 2
P0
n 0 n ! c!1 n 0 n ! 3 2!1 1 3
3
1
2 1 4 2 1
1
6
1
1
1 1
3 2 2 9 3 3 3 2
3
P0 0.5
c
P0
1
(i) P (a customer waits for service) = P( N c)
c! 1
2
2 4
0.5 0.5
P( N 2)
1 3 9 1
0.1667
2! 1 2 6
1 2
3 3
P( N 2) 0.1667
1
(ii) The proportion of time the pumps are busy = =
c 3
The proportion of time the pumps are idle = 2/3
Percentage of idle time = (2 / 3) 100 66.7%
Example 12
A post office has 3 windows providing the same services. It receives on an average 30
customers per hour. Arrivals are Poisson distributed and service time exponentially
distributed. The post office serves on average 12 customers per hour. (i) What is the
probability that a customer will be served immediately? (ii) What is the probability that a
customer will have to wait? (iii) What is the average number of customers in the system? (iv)
What is the average total time that a customer must spend in the post office?
Solution
Servers - 3 windows – multiple server
System capacity – Infinite
Model : M / M / C : ( / FIFO)
Mean arrival rate λ = 30 per hr
Mean service rate µ = 12 per hr
Number of windows c = 3
30 30
2.5 & 0.83, 1 0.17
12 c 3(12)
287
1
c 1 1 n 1
c
P0
n 0 n ! c!1
1 1
2 2.5 n 2.5
3
2.5 15.625
2
6.625 15.319 0.0456
1
1 2.5
n 0 n ! 3!(0.17) 2 1.02
P0 0.045
Probability that a customer will be served immediately is
/ P0
c
P[n c] 1 P[n c] 1
c !1
2.5 0.045 15.625 0.045 0.311
3
1 P[n 3] 1 1
3!(0.17) 1.02
P[n c] 0.311
(i)Probability that a customer will have to wait
P[a customer will be served immediately]
c
P0
1
P( N c)
c! 1
2.5 0.045 0.6893
3
P[ n 3]
3! 0.17
(ii)Average number of customers in the system
s 1
1 1 1 1
2.5 0.045 2.5
4
Ls P0
s!s 1 3!3 0.17
2 2
1
39.0625 34.602 0.045 2.5 3.379 2.5 5.879
18
Ls 5.879
(iii)Average total time that a customer must spend in the post office
L 5.879
E Ws s 0.19 hr 0.19 60 min 11.75 min
30
Example13
A tax consulting firm has 3 counters in its office to receive the people who have problems
concerning their income and sales taxes. On an average 48 persons arrive in an 8 hour day.
Each tax advisor spends 15 min on an average and on an arrival. If the arrivals are Poisson
distributed and service time are according to exponential, find the following. (i)Probability
that a customer has to wait before he gets service. (ii) The expected no of idle tax advisors at
any specified time.
Solution
Servers - 3 counters – multiple server
System capacity – Infinite
Model : M / M / C : ( / FIFO)
288
Mean arrival rate λ = 48/8 = 6 per hr
Mean service time = 15 mins
Mean service rate µ = 1/15 per min= 1/15 x 60 per hr = 4 / hr
6 1.5
Number of Channels c = 3, 1.5 & 0.5
4 c 3
1 1
c 1 1 n 1
c
2 1 1 3
P0
n
1.5
3!1 0.5 1.5
n 0 n ! c!1 n 0 n !
1
1.5 1
3
2
1 1
1 1.5 1.5 1 1.5 1.125 3.375
2! 6 1 0.5 3
3.625 1.125 4.75 0.21
1 1
P0 0.21
1
c 1 1 n 1
c
P0
n 0 n ! c!1
1
2 1.5 n 1.53
1 1.5 1.125 1.125 4.75 0.211
1 1
n 0 n ! 3!(0.5)
Probability that a customer has to wait before he gets service is
c
P0 1.53 0.21
1
P( N c) 0.236
c! 1 3!1 0.5
(i)The expected no of idle tax advisors at any specified time
We know that the probability of no customers in the system is given by P0. Since there are 3 tax
advisors, all of them should be idle. We have to determine P1 & P2.
P ( 3 advisors are idle ) = P(empty system ) = P0
P ( 2 advisors are idle ) = P(one customer in the system ) = P1
P ( 1 advisors are idle ) = P(two customer in the system) = P2
P ( 0 advisors are idle ) = P(three or more customers in the system )
n
1
Pn P0 ,1 n c
n!
2
1 1 3
(i)
P1 1.5 0.21 0.315 & P2 0.21 0.236
1! 2! 2
Expected number of idle advisors at any specified time is given by
= 3P0 + 2P1 + P2= 3 x 0.21 + 2 x 0.315 + 0.236 1.46 ~ 1.5
Example 14
A telephone company is planning to install telephone booths in a new airport. It has
established the policy that a person should not have to wait more than 10% of the times he
tries to use a phone. The demand for use is estimated to be Poisson with an average of 30 per
289
hour .The average phone call has an exponential distribution with a mean of 5 min. How
many phone booths should be installed?
Solution
Mean arrival rate 30 per hour
Mean service rate 1/ 5 per min (1/ 5) 60 / hr 12 per hour
In order that infinite queue may not build up, Let us assume that the company decides to install
two telephone booths
30
ie, 2.5 , 1 c 2.5
12 c
Model : M / M / C : ( / FIFO)
2 , Thus the telephone company must have at least three telephones to meet the demand for
service.
Now we have to find the number c of telephone booths such that
P(W 0) 0.10, or P(N c) 0.10
We have to find c such that
P( N c)
/ 0.10 .But this equation is not easily solvable.
c !1
Hence we proceed by trials and find out least value of c that satisfies this equation.
CASE : 1
Let ie, c 3
P( N c)
/ 0.10
c !1
(2.5)3 P0
P (W 0) 15.625 P0
2.5
3!1
3
1 1
c 1 n c 31 1
1 1 1
P0 2.5 2.5
n 3
n 0 n !
n 0 n ! 3!1 2.5
c! 1
c 3
1
2.5 (2.5)3
2
1 2.5 0.0449
2 6 0.1667
P(W 0) 15.625 0.0449 0.7022 not less than 0.10
CASE : 2
Let ie, c 4
(2.5) 4 P0
P(W 0) 4.3403P0
2.5
4!1
4
290
1
1
41 1 1 2.5 2.5 (2.5)4
2 3
2.5 2.5 1 2.5 0.0737
n 4
n 0 n ! 4!1 2.5 2 3! 24 0.375
3
P0 0.0737
P(W 0) 4.3403 0.00737 0.3199 not less than 0.10
CASE : 3
Let ie, c 5
P(W 0) 0.1304not less than 0.10
CASE : 4
Let ie, c 6
P(W 0) 0.047 0.10
Hence the number of booths to be installed is 6.
Thus an installation of six phones would give a probability of 0047 that a customer would have to
wait. Since this is less than 0.10 , it is the smallest number that will meet the company policy.
Hence 6 phone booths should be installed.
Exercise 4.3
1. A library wants to improve the service facilities in terms of the waiting time of its barrowers.
The library has two counters at present and barrowers arrive according to a Poisson distribution
with arrival rate 1 at every 6 minutes and service time follows exponential distribution with a
mean of 10 minutes. The library has relaxed its membership rules and a substantial increase in the
number of barrowers is expected. Find the number of additional counters to be provided if the
arrival rate is expected to be twice the present value and the average waiting time of the barrower
must be limited to half the present value.
Ans: P0 0.091, wq 22.75hrs, whenc 4P0 0.0213, wq 9.86 min
2. A petroleum company is considering expansion of its one unloading facility at its refinery. Due
to random variations in whether, loading delays and other factors, ships arriving at the refinery to
unload crude oil arrive at a rate of 5 ships per week. The service rate is 10 ships per week.
Assume arrivals follow a Poisson process and the service time is exponential. (i) Find the average
time a ship must wait before beginning to deliver its cargo to the refinery. (ii) If a second berth is
rented, what will be the average number of ships waiting before being unloaded? (iii) What
would be the average time a ship would wait before being unloaded with two berths? (iv) What is
the average number of idle berths at any specified time? Ans : (i)1/10 (ii)0.6 (iii)0.0333 (iv)
1/150 (v) 1.5
3. There are three billing clerks in a super market. If customers come for billing at the rate of 20
per hour in a Poisson manner and the service time for each customer is exponential with mean 5
minutes, (i) Find the probability that a customer gets service without waiting (ii) What is the
average waiting time in queue for those who actually wait? (iii) What is the average time a
customer spends in waiting for his turn and billing? (iv) If one of the clerks reads a novel in his
spare time and he takes 6 minutes to read one page, how many pages can be read in a 10 hour
day? Ans: P0 0.1727,(i)0.7002 (ii)3.75 min (iii) Lq 2.0415 & ws 6.13 min (iv)89
4. A dress shop has 3 salespersons. Assume that arrivals follow Poisson pattern with an average
of 10 min between arrivals. Also assume that any salesperson can provide the desired service for
291
any customer. If the time to provide service for a customer is exponentially distributed with a
mean of 20 min per customer, calculate E(N), E Nq , E(W), E Wq and Pn , n 0,1& 2
Ans: 2.1739,0.1739,21.739 min, 1.739 min, P0 0.1304, P1 0.2608, P2 0.2602 .
4.4 QUEUES WITH FINITE WAITING ROOMS-SINGLE SERVER QUEUEING
MODEL WITH FINITE CAPACITY [(M/M/1) : (K/FIFO)]
Introduction
For the simple queue M/M/1, the assumption is that the system can accommodate any number of
units. In this model, we assume that the system can accommodate a finite number of units (say, K)
including the one being served, if any. Customers arrive in accordance with a Poisson process
with rate , a customer will join the system whenever he finds less than K in the system and a
customer who arrives when there are K in the system leaves the system and is lost to the system.
Service time is exponential with rate . A queue with limited waiting space is known as a queue
with finite huffier.
Here the system will behave as an ordinary M/M/1 queue so long as the number in the system is
less than k. Then when it reaches the state k or the system contains k customers, no more arrival is
allowed to the system, and the number in the system cannot exceed K. From state K, only
departure is possible. For this model
n , n 1, 2,....
if n 0,1,...k
n Denote the traffic intensity /
0 if n k
The effective arrival rate to the system M/M/1/K is the expected number of customers joining
the system in unit time and equals
1 k
,
1 k 1
1 P0
k
k 1 ,
The average inter-arrival time is 1/
Theorem 4.4
Derive the steady state probability for the [(M/M/1): ( K /FIFO)] model.
Solution
From Theorem 4.1, we get
...
Pn 0 1 2 n 1 P0 n 1, 2,.. (1)
1 2 3 ... n
1
P0 2
0 12 3 .......n 1
1
n 1 1 2 3 ....... n
For this model
n , n 1, 2,....
if n 0,1,...k
n
0 if n k
292
Using (1), we get
n
Pn = p0 .........(*)
n
k
k
since the Pn =1 , using (*) P0 =1
n 0 n 0
k
n
2
k
P0 =1 P0 1 .... =1 , W .K .T a ar ...ar a
n
1 r n 1
n 0
1 r
k 1
1 1
1
P0 =1, valid for P0 ,
k 1
1 k 1
1 1
n 1
Pn = , valid for
k 1
1
when using(1)
Pn =P0
i.e P1 =P0 ; P2 =P0 ...;.......(2)
k
since the Pn =1
n 0
P0 P1 P2 ......PK 1
P0 P0 P0 ......P0 1
(k 1) P0 1
1 1
P0 , Pn ;sin ce Pn =P0
k 1 k 1
The system size probabilities are
1
,
1 k 1
P0
1 ,
k 1
n 1
if
1 k 1
Pn
1
if
k 1
Some Important characteristics for [(M/M/1): (k/FIFO)] model
Average number of customers in the system
293
(k 1) / (k 1)
k 1 k 1
Ls ,
1 / 1 1
k 1 k 1
k
Ls if
2
Average number of customers in the queue (queue length)
Lq Ls , 1 P0
Average waiting time of a customers in the system
L
Ws s
Average waiting time of a customers in the queue
Lq
Wq
Worked Examples 4.4(A)
Example 1
Draw the transition diagram for (M/M/1): (k/FIFO) model.
Example 2
Trains arrive at the yard every 15 minutes and the service time is 33 minutes. If the line
capacity of the yard is limited to 5 trains, find the probability that the yard is empty.
Solution
Yard – Single server
Trains – finite capacity
k = 5 trains
Model : M / M/1 : k / FIFO
Mean arrival time = 15 minutes
1
Mean arrival rate per minute
15
Mean service time = 33 minutes
1 1/15 33
Mean service rate per minute , 2.2
33 1/ 33 15
1 1 2.2
The probability that the yard is empty is P0 0.011
1 1 2.2
k 1 6
P0 0.011
294
Example 3
Define Effective arrival rate for (M/M/1 : N /F CFS Queueing System
(or) Explain effective arrival rate in a finite queue capacity model.
Solution
In a finite source system, the number of customers who are already in the system affects the
arrival rate. For if every potential customer is already in the queue, the arrival rate drops to zero.
Hence we introduce effective arrival rate.
or eff is the effective arrival rate and is denoted by 1 p0
P0 is the probability of no customer in the system.
Example 4
What is the effective arrival rate for (M/M/1 : 4 /F CFS queueing model
When = 2 and = 5.
Solution
Model : M / M/1 : k / FIFO
Given: = 2 , = 5, k = 4 .
2
0.4
5
1 1 0.4 0.6
P0 0.6062
1 1 0.4
k 1 41
0.9876
or eff is the effective arrival rate and is denoted by 1 p0
5 1 0.6062 1.969
Example 5
Obtain the steady state probabilities of an (M/M/1 : N /F IFO queueing model.
Solution
For a general Poisson queue we have
1
P0 =
......n 1
1 0 1
n 1 12 ........n
......n 1
Pn = 0 1 P , n 1, 2,3......
12 ........n 0
For this model
n , n 1, 2,....
if n 0,1,...k 1
n
0 if n k , k 1,...
n
Pn = p0 .........(1)
n
k k
since the Pn =1 , using (1) P0 =1
n 0 n 0
295
k 1
n 1
P0 =1 P0 =1, valid even for
k
n 0
1
1
,
1 k 1
P0
1 ,
k 1
n 1
if
1 k 1
Pn where
1
if
k 1
Worked Examples 4.4(B)
Example 1
Patients arrive at a clinic according to poisson distribution at a rate of 30 patients per hour.
The waiting room does not accommodate more than 14 patients. Examination time per
patient is exponential with mean rate of 20 per hour. (i) Find the effective arrival rate at the
clinic.(ii) What is the probability that an arriving patient will not wait? (iii) What is the
expected waiting time until a patient is discharged from the clinic?
Solution
Servers - Clinic – Single server
System capacity – Patients – Finite
Model : M / M/1 : k / FIFO
Mean arrival rate 30 per hour
Mean service rate 20 per hour
k = 1+ 14 = 15
30
1.5 / hr
20
P0 if
k 1
1 1.5 0.5 0.5
0.00076
1 1.5 1 656.93 655.93
16
P0 0.00076
(i)To find the Effective arrival rate:
1 P0 20(1 0.00076) 19.98
' 20 hrs nearly
296
(ii) Probability that an arriving patient will not wait is equal to probability of no customer in the
system. i.e. P0
P0 0.00076
(iii) Expected waiting time until a patient is discharged from the clinic
L
E Ws s (1)
Where,
(k 1)
k 1
1.5 16(1.5)16 1.5 16(656.93) 16(656.93)
Ls 3
1 1
k 1
1 1.5 1 (1.5)16
0.5 1 (656.93) 655.93
Ls 3 16.02 13.02 (2)
Sub (2) in (1)
13.02 13.02
Ws hrs 60 min Ws 39 min s
20 20
Example 2
A One Person barber shop has 6 chairs to accommodate people waiting for a haircut.
Assume that the customers who arrive when all the 6 chairs are full leave without entering
the barber shop. Customers arrive at the rate of 3 per hour and spend an average of 15
minutes in the barber’s chair. Compute (1) P0 (2) Lq (3) P7 (4)Ws
Solution
Server –One barber - Single server
System capacity - finite capacity
Model : M / M/1 : k / FIFO
k= 6+1=7
Mean arrival rate 3/ hour
1 1
Mean Service rate / min 60 per hour 4 / hour
15 15
3
0.75
4
i) To find Po :
1 1 0.75
P0 0.2778
1 1 0.75
k 1 7 1
ii) Average number of customers in the queue :
Lq Ls , Where,
(k 1) 0.75 (7 1) 0.75 (8) 0.1001
k 1 7 1
Ls 3 2.11
1 1
k 1
1 0.75 1 0.75
7 1
1 0.1001
Ls 2.11
1 P0 4(1 0.2778) 0.7222
Lq 2.11 0.722 1.3878
297
Lq 1.388
iii) To find P7 :
P( N 7) P7 7 P0 0.75 0.2778 0.037 3.7%
7
P7 3.7%
Ls 1
iii) Average waiting time in the system E Ws 2.11 0.7304hrs
2.888
E Ws 0.7304 60 min s 43.8 min s
Example 3
Trains arrive at the yard every 15 minutes and the service time is 33 minutes. If the line
capacity of the yard is limited to 5 trains, find the probability that the yard is empty and the
average number of trains in the system, given that the inter arrival time and service time
are following exponential distribution.
Solution
Servers - One yard – Single Server
System capacity – Trains – finite capacity
Model : M / M/1 : k / FIFO
k =5
Mean arrival time = 15 mins
1
Mean arrival rate / min
15
Mean service time = 33 mins
1
Mean service rate / min
33
33
2.2
15
(i) Probability that the yard is empty
1 (2.2) 1.2
P0 0.0107
k 1
1 (2.2) 6
112.38
(ii) Average number of trains in the system:
(k 1)
k 1
Ls
1 1
k 1
2.2 (5 1) 2.2
51
680.2794
1.833 1.833 6.0534 4.2204
1 2.2 1 2.2
51
112.38
L s 4.22
Example 4
Obtain the mean number of customers in the system, average number of customers in the
queue for M/M/1/K queue.
Solution
We Know
298
1
, valid for
n
Pn =
N 1
1
when using(1)
Pn =P0 i.e P1 =P0 ; P2 =P0 ...;.......(2)
,
k
since the Pn =1 P0 P1 P2 ......PN 1
n 0
P0 P0 P0 ......P0 1
(k 1) P0 1
1 1
P0 , Pn ;sin ce Pn =P0
k 1 k 1
n
if
k 1
Pn
1
if
k 1
Average number of customer in the system (Ls):
1
n 1 k
k
k
d
k 1 k 1
Ls n Pn x , where x
n
n
n0 n 0 n 0 dx
1 1
(1 x) x d
k 1
. 1 x x 2 ... x k
1 x dx
(1 x) x d 1 x k 1 1 x k 1
2
k
1 x k 1 dx 1 x
. 1 x x ... x
1 x
(1 x) x (1 x) (k 1) x (1 x ) x (k 1) x (k 1) x 1 x
k k 1 k k 1 k 1
1 x k 1
(1 x) 2 1 x 1 xk 1
x 1 (k 1) x k kx k 1 x 1 (k 1) x k kx k 1 x k 1 x k 1
1 x 1 xk 1 1 x 1 xk 1
x 1 x k 1 (k 1) x k k 1 x k 1 x 1 x k 1 k 11 x x k 1
1 x 1 xk 1 1 x 1 x k 1
x
k 1 x k 1
1 x 1 x k 1
299
k 1
k 1
Ls
, if
k 1
1 1
k
n 1 1 k (k 1)
and Ls [1 2 ... k ]
n0 k 1 k 1 k 1 2
k
Ls , if
2
Average number of customers in the queue (Lq):
Lq E ( N q ) E ( N 1)
k k k
( n 1) Pn nPn Pn
n 1 n 1 n 1
k
k
nPn Pn P0
n 0 n 0
E ( N ) (1 P0 )
Lq Ls 1 P0
As per Little’s formula,
This is true when the average arrival rate is λ throughout. Now, 1 P0 because the average
arrival rate is λ as long as there is a vacancy in the queue and it is zero when the system is full.
Hence we define the overall effective arrival rate, denoted by is λ or is λeff, as
1 P0 or 1 P0
Then E(Nq) becomes,
Lq Ls which is modified Little’s formula.
Ls Lq
E Ws and E Wq
Example 5
The local one attender barber shop can accommodate a maximum of 5 people at a time
(4 waiting and 1 getting hair cut). Customers arrive according to a Poisson distribution with
mean 5 per hour. The barber cuts at an average rate of 4 per hour. (i) What is the expected
number of customers waiting for a haircut? (ii) How much time can a customer expect to
spend in the barber shop?
Solution
One person barber shop – single server
System Capacity – finite capacity
k=5
Model : M / M/1 : k / FIFO
Mean arrival rate λ = 5/hr
Mean service rate µ = 4/hr
300
5
1.25
4
1 1.25 0.25
P0 0.0888
1 1.25
k 1 6
2.815
P0 0.0888
(i) Expected number of customers waiting for a hair-cut
Lq Ls
where, 1 P0 4(1 0.0888) 3.6448
(k 1) 6 1.25
k 1
6
1.25 6(3.8147)
Ls 5 5 8.1317 3.1317
1 1 1 1.25 1 1.25
k 1 6
1 3.8147
3.6448
Lq 3.1317 2.22
4
L q = 2.22 customers
ii) Customer can expect to spend time in the barber shop
1 3.1317
E Ws = Ls 0.8592 hrs 0.8592 60 51.54 min
3.6448
Example 6
Customers arrive at a one window drive in a bank according to a Poisson distribution with a
mean 10 per hour. Service time per customer is exponential with mean 5 minutes. The space
in front of window, including that for the serviced can accommodate a maximum of 3 cars.
Others cars can wait outside this space. (i) What is the probability that an arriving customer
can drive directly to the space in front of the window? (ii) What is the probability that an
arriving customer will have to wait outside the indicated space? (iii) How long is an arriving
customer expected to wait before being served?
Solution
Servers – one window drive – single server
System capacity – space in front of the window - Finite
Model : M / M/1 : k / FIFO
k=3
Mean arrival rate 10 / hr
Mean service time = 5 min
1
Mean service rate 1/ 5 per min 60 / hr 12 / hr
5
10
0.83
12
i) Probability that an arriving customer can drive directly to the space in front of the window
0.83 0.17
P0 k 1
4
0.327
0.83 0.52
ii)Probability that an arriving customer will have to wait outside the indicated space
1 P0 1 0.327 0.673
301
iii) Expected waiting time of a customer before being served
1 P0 12 1 0.327 8.076
(k 1) (4) 0.83
k 1
4
0.83 1.898
Ls 4.882 1.27
1 1 1 0.83 1 0.83
k 1 4
0.5254
8.076
Lq Ls 1.27 0.597
12
Lq 0.597
E Wq 0.0739hr
8.076
Example 7
In a railway marshaling yard, goods trains arrive at a rate of 30 trains per day. Assuming
that the inter-arrival time follows an exponential distribution and the service time
distribution is also exponential with an average 36 minutes. If the yard can admit 9 trains at
a time, Calculate the following: (i) The Yard is empty (ii) Average Queue Size
Solution
Server - one yard –-- single channel
System capacity – Capacity of the yard is 9 –-- finite capacities
Model : M / M/1 : k / FIFO
Given: trains arrive at a rate of 30 trains per day (changing into minutes)
30 30 1
Mean arrival rate / hr / min /min
24 24 60 48
1
Mean service rate per min
36
1/ 48
N = 9, 0.75
1/ 36
i) The probability that the yard is empty:
1 0.75 0.25
P0 k 1
0.265
1 0.75
10
0.9437
P0 0.265 0.27
1
1 P0 1 0.27 0.023
36
ii) Average queue length:
Lq Ls
(k 1)
k 1
(9 1) 0.75
9 1
0.75 0.4505
where, Ls 3 2.52
1 1
k 1
1 0.75 1 (0.75)91 0.9437
Ls 2.52
0.023
Lq Ls 2.52 1.7
1/ 36
Example 8
302
A car park contains 5 cars. The arrival of cars is Poisson at a mean rate of 10 per hour. The
length of time each car spends in the car park has negative exponential distribution with mean
of 2 min. How many cars are in the car park on an average and what is the probability of a
newly arriving customer finding the car park full and leaving to park his car elsewhere.
Solution
Server - One car park – single server
System capacity Cars – finite capacity
Model : M / M/1 : k / FIFO
Mean arrival rate λ = 10/hr
Mean service time = 2 mins
Mean service rate (µ) = 1/2 per min= 1/2 x 60 per hr=30/hr
k=5
10 1
30 3
1 1 (1/ 3)
P0 0.668
1 k 1 1 1/ 36
P0 0.668
(i)Number of cars in the park
6 1/ 3
6
(k 1) k 1 1/ 3 0.0082
Ls 0.5 0.49
1 1 k 1
1 (1/ 3) 1 1/ 3 6
0.6686
Ls 0.49
(ii)If the car park is full, then the newly arrived car has to go away
5
1
P[ Newcar giving away ] P[ N 5] , P5 P0 0.668 0.0027
5
3
P5 0.0027
Example 9
A small mail - order business has one telephone line and a facility for call waiting for
additional customers. Orders arrive at the rate of one per minute and each order requires 2
minutes and 30 seconds to take down the particulars. What is the expected number of calls
waiting in the queue? What is the mean waiting time in the queue?
Solution
Servers - Telephone line
System capacity – waiting facility - Finite
Model : M / M/1 : k / FIFO
Mean arrival rate 1 per minute
Mean service time 2 minutes and 30 seconds =2.5 min
1
/ min 0.4 / min
2.5
1
k = 1+ 2= 3, 2.5
0.4
303
1 2.5 1.5 1.5
P0 k 1
0.0394
1 2.5 1 39.06 38.06
4
P0 0.0394
1 P0 0.4(1 0.0394) 0.3842
(i)To find the expected number of calls waiting in the queue:
(k 1)
k 1
2.5 4(2.5) 4
Lq Ls where, Ls
1 1
k 1
1 2.5 1 (2.5) 4
4(39.06) 156.25
1.66 1.66 2.45
1 39.06 38.06
Ls 2.45
0.3842
Lq Ls 2.45 1.4895
0.4
(ii)To find the mean waiting time in the queue:
Lq 1.4895
E Wq 3.8769 min s
0.3842
Example 10
In a car wash service facility, cars arrive for service according to a Poisson distribution with a
mean 5 per hour. The time for washing and cleaning each car varies but is found to follow an
exponential distribution with mean 10 min per car. The facility cannot handle more than one car at
a time and has a total of 5 parking spaces. (i) What is the probability that an arriving car will get
service immediately? (ii) Find the effective arrival rate. (iii) Find the expected number of parking
spaces occupied.
Solution
Server - One car wash service – single server
System capacity - Parking – finite capacity
Model : M / M/1 : k / FIFO
k=5
Mean arrival rate λ = 5/hr
Mean service time = 10 mins
Mean service rate µ = 1/10per min=(1/10) x 60 per hr = 6/hr
5
6
i) Probability that an arriving car will get service immediately
5
1
6 0.167 0.2511
P0
k 1 6
5 0.665
1
6
Effective arrival rate
1 P0 6 1 0.25 4.5
ii) Expected number of parking spaces occupied
304
6
5 5
6
(k 1)
k 1
6 6
6
Lq Ls Where, Ls
1 1
k 1
1
5 5
6 1 6
6(56 ) 93750
5 6 6 5 5 3.0212 1.9788 5 2.7098 2.29
6 5 31031
Ls 2.29
4.5
Lq 2.29 1.54
6
The expected number of parking space occupied = 1.54
Example 11
A derive in car wash facility with single server has a capacity to hold on its grounds a maximum
of 10 cars (including the one in wash). The company has found its arrivals to be poisson with a
mean rate of 20 cars/hour and its service time to be exponential with a mean of 12 min. What is
the average number of cars lost to the firm every 10 hour day as a result of its capacity limitations?
Solution
Servers - Single server
System capacity – cars - Finite
Model : M / M/1 : k / FIFO
Mean arrival rate: 20 / hr
Mean service time = 12 min
: 1/12 per min 1/12 60 per hr 5 / hr
λ 20
ρ 4
μ 5
1 14 3
p0
1 4194304
k 1
1411
P0 0.00000072
is the effective arrival rate
1 P0 5
Number of cars lost per hour = λ λ! =20 – 5 =15
Number of cars lost per day = 15 x 10 = 150.
Example 12
Consider a single server queueing system with Poisson input, exponential service times.
Suppose the mean arrival rate is 3 calling units per hour, the expected service is 0.25 hours
and the maximum permissible number calling units in the system is two. Find
i) the steady state probability distribution Pn n 0 of the number of calling units in the
system and in the Queue. (ii) Average waiting time in the system and in the queue
(iii) Calculate the expected number in the system.
Solution:
Server – Single server
System capacity - Calling units – Finite capacity
305
Model: M / M/1 : k / FIFO
k=2
Mean arrival rate 3 per hour
Mean service time = 0.25 hour
1
Mean Service rate per hour 4 per hour
0.25
3
0.75
4
Pn n P0
1 1 0.75 0.25
P0 0.433
1 1 0.75
k 1 3
0.578
Pn 0.75 0.433
n
Average number of calling units in the system
(k 1)
k 1
0.75 (2 1)(0.75) 21 0.75 (3)(0.75)3
Ls 3 2.189 0.81
1 1
k 1
1 0.75 1 (0.75) 21 1 0.75 1 (0.75)3
Ls 0.81
Average number of calling units in the queue
Lq Ls
1 P0 4(1 0.433) 2.27
2.27
2.27 0.81 0.2425
4
Lq 0.2425
Average waiting time in the system
L 1
E Ws s 0.81 0.36 hrs
2.27
= (0.36)(60) = 21.6 mins
E Ws 21.6 mins
Average waiting time in the queue
Lq 1
Wq 0.2425 0.11 hrs = (0.11)(60) = 6.6 mins
2.27
E Wq 6.6 mins
Exercise 4.4
1. At railway station, only one train is handled at a time. The railway yard is sufficient only for
two trains to wait while the other is given signal to leave the station. Trains arrive at the station
at an average rate of 6 per hour. Assuming Poisson arrivals and exponential service
distribution, find the probabilities for the number of trains in the system. Also find the average
waiting time of a new train coming in to the yard. If the handling rate is doubled, how will the
above results get modified?
306
Ans: Case(1) Pn 1 / 4,n 1, 2, 3, Ls 3 / 2, Ws 20min ,
Case (2) P0 8 / 15, P1 4 / 15, P2 2 / 15, P3 1 / 15, L s 0.733, Ws 7.85min
2. In a lorry-wash service facility, lorries arrive for service according to Poisson distribution with
mean 5 per hour. The time for washing and cleaning each lorry has exponential distribution
with mean 10 minutes per lorry. The facility cannot handle more than one lorry at a time and
has a total of 5 parking spaces. (i) find the effective arrival rate (ii) What is the probability that
an arriving lorry will get service immediately upon arrival? (iii) Find the expected number of
parking spaces occupied.
Ans: λ 4.602, P0 0.232,Ls 2.29,Lq 1.53
3. If for a period of 2 hours in the day(8 to 10 am) trains arrive at the yard every 20 minutes but
the service time is 36 minutes, then calculate for this period (i) The probability that the yard is
empty (ii) The average number of trains in the system, on the assumption that the line capacity
of the yard is limited to 4 trains only. Ans: P0 0.04, Ls 3trains
4.5 QUEUES WITH FINITE WAITING ROOMS MULTI- SERVER QUEUEING MODEL
WITH FINITE CAPACITY [(M/M/C) : (k/FIFO)]
Introduction
Consider a c-server model with Poisson input and exponential service time such that when all the
c-channels are busy an arrival leaves the system without waiting for service. This is called a (c-
channel) loss system and was first investigated by Erlang. This is a birth-and-death queueing
model with arrival and service rates are defined as
( k n ) , 0 n k n , 0 n k
n , n
0, n k s , n k
Theorem 4.5
Derive the steady state probability for the [(M/M/C): (k/FIFO)] model.
Solution
From Theorem 4.1, we get
...
Pn 0 1 2 n 1 P0 n 1, 2,.. (A)
12 3 ... n
1
P0 B
0 12 3 .......n 1
1
n 1 1 2 3 ....... n
For the model [(M/M/C):(k/FIFO)]
, n 0,1, 2,...k 1
n , and
0, n k
n , 0 n c
n
s , n c
From (A), we deduce
For 0 n c
307
n 1
i n 1
( k i )
Pn P0 P0
i 0 i 1 i 0 (i 1)
n
k (k 1)(k 2)...(k n 1)
P0
n!
n
k
Pn P0 , n 0,1,..., c 1
n
For c n k
n 1
n 1
( k i )
Pn i P0 P0
i 0 i 1 i 0 (i 1)
n
k! 1
Pn P0
(k n)! c !c n c
Using Normalizing condition
k
P
n0
n 1
We deduce
1
c 1 k n k k!
n
1
P0
n 0 n n c (k n)! c !c
n c
Some Important characteristics for [(M/M/C): ( K /FIFO)] model
Average number of customers in the queue (queue length)
c
P (1 k c ) (k c) k c 1
Lq 0
c ! (1 ) 2 1
Average waiting time of a customers in the system
' c 1
Ls Lq ' c (c n) Pn
n 0
Average waiting time of a customers in the system
L
Ws s (By Little’s formula)
Average waiting time of a customers in the queue
Lq
Wq (By Little’s formula)
Worked Examples 4.5(A)
Example 1
Draw the state transition diagram for the model (M/M/c):(N/FIFO)
Solution
308
Example 2
For (M/M/C): (N/FIFO) model, write down the formula for
a) Average number of Customers in the queue
b) Average waiting time in the system
Solution
Average number of customers in the queue (queue length) is given by
c
P0 (1 k c ) (k c) k c 1
Lq
c ! (1 ) 2 1
Average waiting time of a customers in the system is given by
'
Ls Lq
Worked Examples 4.5(B)
Example 1
Derive the formula for (i)Average number Lq of customers in the queue. (ii) Average
waiting time of a customer in the queue (iii) And also derive Ws, and Ls as well as in the
system for M/M/c : k /F IFO.
Solution
W.k.t for M/M/C : k /F IFO.
1 n
P0 for n c
n!
n
1
Pn n c P0 for c n k
c!c
0 nk
1
1
n
c 1 1 n 1 c k n c
Pn P0 for c n k & P0
c!c n c n 0 n ! c! n c c
i) The average queue length is given by
n nc c
k k
1 P k (n c)
Lq (n c) Pn (n c) n c P0 0 n c
n c n c c!c c! n s c
c n c c n c
P k
(n c) P k
0
c!
n c 0 (n c)
ns c c! nc c
309
c
P k
(n c)
n v
0 put n c x
c! n c
P0 1 k c
c c k c k c 1
P k c
0 ( x)
x
c! c! 1 2 1
x 0
c
P (1 k c ) (k c) k c 1
Lq 0
c ! (1 ) 2 1
Average waiting time of customers in the system is given by
'
Ls Lq
c 1
' c (c n) Pn
n 0
Average waiting time of a customers in the system is given by
L
Ws s (By Little’s formula)
Average waiting time of a customers in the queue is given by
Lq
Wq (By Little’s formula)
Example 2
A 2 person barber shop has 5 chairs to accommodate waiting customers. Potential
customers, who arrive when all 5 chairs are full leave without entering barber shop.
Customers arrive at the average rate of 4 per hour and spend an average of 12min in the
barber’s chain compute P0, P1, P7, E(Nq) and E(W).
Solution
Servers - 2 person barber shop – multiple server
System capacity – 5 chairs – Finite capacity
Model : M / M / C : K / FIFO
Mean arrival rate λ 4 per hour,
Mean service rate μ 5 per hour,
k=2+5=7
i) To find P0
1 1
c 1 1 n 1 c k n c 1 1 4 n 1 4 2 7 2 n 2
P0 =
n 0 n ! c! n c c n 0 n! 5 2 5 n 2 5
1 1
4 8 2 2 2 2 3 2 4 2 5 9 8 1 (0.4)7
= 1 1 = =0.4287
5 25 5 5 5 5 5 5 25 1 0.4
P0 0.4287
ii) To find P1 and P7
n
1
Pn P0 for c n k
c!c n c
310
n
1
P1 P0
1!
4
P1 0.4287 0.3430
5
P1 0.3430
7
1 4
P7 72
0.4287
2 2 5
P7 = 0.0014
iii) To find E ( N q )
c
P (1 k c ) (k c) k c 1
E ( N q ) Lq 0 ,
c ! (1 ) 2 1
λ
(0.4287) (0.8) 2
0.4 1 (0.4)5 5 0.6 (0.4)5
Where ρ 2
μs 2 0.6
Lq 0.15 customer
iv) To find E (W )
s 1 1
E ( N ) E ( N q ) s ( s n)Pn 0.1462 2 (2 n)Pn
n 0 n 0
2.1462 (2 P0 1 P1 ) 2.1462 (2 0.4287 1 0.3430)
=0.95 customer
1
E(w) ' E(N)
λ
s 1
λ μ s (s n)Pn
'
Where
n 0
4[2 (2 0.4287 1 0.3430)] =3.1984
0.9458
E (W ) 0.2957h or 17.7 min
3.1984
Example 3
A barber shop has 2 barbers and 3 chairs for waiting customers. Assume that customers
arrive in Poisson fashion at a rate of 5 per hour and that each barber services customers
according to exponential distribution with mean of 15 min. Further, if a customer arrives
and there are no empty chairs in the shop, he will leave. Find the steady state probabilities.
What is the probability that the shop is empty? What is the expected numbers of customers
in the shop?
Solution
Server - Two barbers--multiple channel
System capacity - 3 chairs -finite capacity
Model : M / M / C : K / FIFO
Mean arrival rate 5 per hr
Mean service time = 15 mins
311
60
Mean service rate 4 per hr
15
C= 2 and k = 2+3= 5
5
4
(i)Staedy state probability :
n
1
Pn P0 for s n k
s !s ns
ii) Probability that the shop is empty(P0) :
1 1
P0
C 1 1 n 1 C k n C 1 1 5 1 5 5 5
n 2 n2
n 0 n ! C ! n C C n 0
n! 4 2! 4 n2 4 x 2
1 1
5 25 5 5 5
2 3
5 25 5 5 5
2 3
1 1
1 1
4 8 8 8 8 4 32 8 8 8
1 1 1
1 1.25 0.781251 0.625 0.390625 0.244 1 1.25 1.765 4.015
P0 0.249
k
1
Therefore Pn x P0
c !c k c
5
1 5
P5 x 0.249
2!23 4
P5 0.0475
iii) Expected numbers of customers in the shop:Ls
C 1
Ls =L q L q C C n pn
n 0
C
Lq P0 1 k C k C 1 k C
2
c !1
where 0.625
C
0.625
Lq 0.249 1.25 X 1 0.6253 31 0.625 0.6253
2
2
2!1 0.625
0.625
Lq 0.389 X 1 0.244 3 0.375 0.244 Lq 0.389 X 1.07
0.28125
Lq 0.416
Therefore:
312
c 1
Ls =L q c c n pn .
n 0
1
Ls =0.416 2 2 n pn .
n 0
Ls =0.416 2 2 p0 p1 where p1 p
0
Ls =0.416 2 2 0.249 0.31125
Ls =0.416 2 2 0.249 0.31125
Ls =0.416 2 0.80925
Ls =1.607 customers
Example 4
A group of engineers has 2 terminals available to aid in their calculations. The average
computing job requires 20 min of terminal time and each engineer requires some
computation about once every half an hour. Assume that these are distributed according to
an exponential distribution. If there are 6 engineers in the group, Find (i) The expected
number of engineers waiting to use one of the terminals and in the computing center. (ii) The
total time lost per day.
Solution
Servers - Two terminals i.e multiple server
System capacity – 6 Engineers i.e finite
Model : M / M / C : K / FIFO
Mean Arrival rate 2 per hr
Mean Service rate 3per hr
Here C (number of service channel) = 2
Capacity k = 6
1
P0
C 1 1 n
1 k
C n C
n 0 n ! C ! n C C
1
P0
1 1 2 1 2 6 2
n 2 n 2
n! 3
n 0 2! 3 n 2 3(2)
1
P0
2 1 2 2
1 1 1 1
2 3 4
1 1
3 2 3 3 3 3 3
1
P0
5 2 121
3 9 81
P0 0.5003
313
i) The expected number of engineers waiting to use one of the terminals and in the
computing centre.
C
1
Lq P0 1 k C k C 1 k C , Here
2
C !1 C 3
1
2
2
1 6 2 1 1
62
0.5003 x 3
2
1 6 2 1
3 1 3 3 3
2!1
3
1
2
2
1 4 2 1
4
2 3 1
2
8
0.5003 x 3 2 1 4 0.5003 x 1
3 2 3 3 3 3 8 81 243
2
3
Lq 0.0796
ii) The total time lost per day.
Lq
Wq
c 1 1
' c (c n) Pn 3 2 (2 n) Pn 3 2 2P0 P1
n 0 n 0
1
Pn P0 , n s
n!
12
P1 0.5 0.335
1! 3
c 1
' c (c n) Pn
n 0
3 2 2 0.5 0.335 3(0.665) 1.995
1.995
Lq 0.079
Wq 0.0396 Wq 0.0396
1.995
Example 5
At a port there are six unloading berths and 4 unloading crews. When all the berths are full,
arriving ships are diverted to an overflow facility 20 kms down the river. Tankers arrive
according to a Poisson process with a mean of 1 every 2 hour. It takes for an unloading crew,
on the average, 10 hours to unload a tanker, the unloading time following an exponential
distribution. Find (i) How many tankers are at the port on the average? (ii) How long does a
tanker spend at the port on the average? (iii) What is the average arrival rate at the
overflow facility?
Solution
Servers – 4 unloading crews – multiple server
System capacity – 6 unloading births – finite capacity
Model : M / M / C : K / FIFO
314
1
Mean arrival rate per hr
2
1
Mean arrival rate per hr c 4, k 6
10
1
2 5
5 , 1.25
1 c 4
10
1 1
c 1 1 n 1 c k n c 3 1 1 6
P0 (5)n (5) 4 (1.25) n 4
n 0 n ! c ! n c c n 0 n ! 4! n4
1
1 1
1 5 52 53
2! 3!
54
4!
2
1 1.25 1.25
1
25 125 625
2.25 1.5625 18.5 20.8333 99.2838
1
6
2 6 24
1
[138.6171] P0 0.007
(i) Total number of tankers at the port:
c
P (1 k c ) (k c) k c 1
Lq 0
c ! (1 ) 2 1
0.007 4 1 (1.25) 2 (2)(1.25)3
(5) (1.25) 2
4!
(1 1.25) 1 1.25
625 1.25 2
0.007 2
(1 1.5625) 1.9531
24 (0.25) 0.25
0.182311 15.625
0.1823 4.375
Lq 0.7975
(ii) The Time spent by a tanker at the port
= Average waiting time by a tanker at the port
Ls
Ws
'
Ls Lq
c 1
' c (c n) Pn
n 0
' 3
4 (4 n)Pn 4 4 P0 3P1 2 P2 P3
n 0
1
2
1
3
4 4 P0 3 P0 2 P0 P0
2! 3!
315
1 125
4 P0 4 3 5 52 53 4 0.007 44 4 0.007 64.833
6 6
'
3.5462
Ls 0.7975 3.5462
Ls 4.3437
'
But 3.5462
1
' 3.5462 ' 3.5462
10
4.3437
Ws 10 12.2448 hrs
3.5462
Ws 12 hrs
(iii) Average arrival at the over flow capacity
Since when there are more than 6 tankers in the port it overflows
n
1
Pn P0 for c n k
c!c n c
n= 6 ,c= 4 , 4 6 6
6
1 1 6 109.375
P ( N 6) n c P0 5 0.007
c !c 4!4 2
16 24
Average arrival at the overflow capacity=average arrival at the port *{P (N=6)]
1
0.2848
2
14.24 per hr
Example 6
Let there be an automobile inspection situation with three inspection stalls. Assume that cars
wait in such a way that when a stall becomes vacant, the car at the head of the line pulls up to
it. The station can accommodate almost 4 cars waiting at one time. The arrival pattern is
Poisson with a mean of one car every minute during the peak hours. The service time is
exponential with a mean of 6 min. Find the average number of customers in the system
during the peak hours, the average waiting time and the average number per hour that
cannot enter the station because of fully capacity.
Solution
Servers – Inspection stalls – multiple server
System capacity – Cars – finite capacity
Model : M / M / C : K / FIFO
Given Mean arrival rate = 1/ min
Mean service rate= 1/ 6 per min
C = 3, k = 4+3 = 7
Traffic intensity / c 2
316
1 1
c 1 1 n 1 c k n c 31 1 1 7 n 3
P0 6 6 2
n 3
n 0 n ! c ! n c c n 0 n ! 3! n 3
1 6 18 36(1 2 4 8 16) 1141 0.00088
1 1
(i)The Average number of customers in the queue is
c
P (1 k c ) (k c) k c 1
Lq 0 3.09 cars
c ! (1 ) 2 1
(ii)The average number of customers in the system
1 n
P0 for n c
n!
n
1
Pn n c P0 for c n k
c!c
0 nk
P1 (6)(0.00088) 0.0053
P2 (18)(0.00088) 0.0158
'
Ls Lq
c 1
' c (c n) Pn
n 0
' 2
3 (3 n)Pn 3 3P0 2P1 P2 3 0.0202 2.9798
n 0
Ls 6.06 cars
(iii)The waiting time in the system is
L
E Ws s 12.3min
(iv)Expected number of cars that cannot enter cannot enter the station is
7
1
P7 60 7 3 P0 60 0.5069 60 30.4 cars / hr
3!c
Example 7
55. A car servicing station has 2 bays where service can be offered simultaneously. Because
of space limitations, only 4 cars are accepted for servicing. The arrival pattern is Poisson
with 12 cars per day. The service times in both the bays is exponentially distributed with
8 cars per day. Find (i)The average number of cars in the service station. (ii) The average
number of cars waiting for service and (iii) The average waiting time a car spends in the
system.
Solution
317
Servers – car service station – multiple server
System capacity – Cars – finite capacity
Model : M / M / C : K / FIFO
Arrival rate λ=12/day and Service rate μ=8/day , c =2, k =4
1
c 1 1 n 1 c k n c 1.5 1 2
1
P0 1 1.5 1 0.75 0.75
2
n 0 n ! c! n c 1 2
c
P0 0.1960 , , Lq P0 1 k c k c 1 k c
2
c c!1
0.75
Lq =0.1960 1.5 2
1- 0.75 -2 0.25 0.75 =0.4134 car.
2 2 2
2 0.25
s 1 1
Ls =Lq s s n pn 0.4134 2 2 n pn 2.4134 2 2 p0 p1 1.73cars.
n 0 n0
Ls s 1
Ws = where λ s s n p n 8 2 2p0 p1 10.512
λ n 0
1.73
Ws 0.1646 day
10.512
Average time that a car has to spend in the system = 0.16
Exercise 4.5
1. An automobile inspection station has 3 inspection stalls. Assume that cars wait in such a way
that when a stall becomes vacant, the car at the head of the line pulls up to it. The station can
accommodate at most 4 cars waiting (7 in the station) at one time. The arrival pattern is
Poisson with a mean of 1 car every minute during the peak hours. The service time is
exponential with a mean of 6 minutes. Find the average number of customers in the system
during peak hours, the average waiting time and the average number of cars per hour that
cannot enter the house because of full capacity. Ans:
P0 0.000876, Lq 3.0905cars, P7 0.5046, Ls .063, Ws 12.24min, 30.28
2. A car servicing station has two bays where service can be offered simultaneously. Due to
space limitation, only four cars are accepted for servicing. The arrival pattern is Poisson with a
mean rate of one car every minute during peak hours. The service time is exponential with
mean 6 minutes. Find (i)The average number of cars in the service station (ii)The number of
cars in the queue. (iii)The average waiting time in the system.
Ans: P0 0.00415, 3.5355,1.5687, 0.8983min
318
You might also like
- Soal Bahasa Inggris Kelas 9 SMP/MTs - Report Text100% (10)Soal Bahasa Inggris Kelas 9 SMP/MTs - Report Text2 pages
- Lecture 5 - QUEUING MODEL-Discrete ModelNo ratings yetLecture 5 - QUEUING MODEL-Discrete Model13 pages
- Lecture 1 Queuing Theory (Elements and Characteristics)No ratings yetLecture 1 Queuing Theory (Elements and Characteristics)34 pages
- Lecture Note Sta412 - Operations Research IINo ratings yetLecture Note Sta412 - Operations Research II41 pages
- Unit 4 - OR and Supply Chain - WWW - Rgpvnotes.inNo ratings yetUnit 4 - OR and Supply Chain - WWW - Rgpvnotes.in20 pages
- Queueing Theory Basics: Processing DelayNo ratings yetQueueing Theory Basics: Processing Delay25 pages
- Analysis of A Queuing System in An Organization (A Case Study of First Bank PLC, Nigeria)No ratings yetAnalysis of A Queuing System in An Organization (A Case Study of First Bank PLC, Nigeria)11 pages
- Queueing Models: Discrete-Event System SimulationNo ratings yetQueueing Models: Discrete-Event System Simulation50 pages
- Multiple Models Approach in Automation: Takagi-Sugeno Fuzzy SystemsFrom EverandMultiple Models Approach in Automation: Takagi-Sugeno Fuzzy SystemsNo ratings yet
- Telecommunications Traffic : Technical and Business ConsiderationsFrom EverandTelecommunications Traffic : Technical and Business ConsiderationsNo ratings yet
- BA4105 Legal Aspects of Business All Units 2 Mark Question and Answer100% (1)BA4105 Legal Aspects of Business All Units 2 Mark Question and Answer20 pages
- A240CX-BD CD DD Flameproof Coil Solenoid Valves PDFNo ratings yetA240CX-BD CD DD Flameproof Coil Solenoid Valves PDF1 page
- IX Science Ch-12 Solutions (Improvement in Food Resources)No ratings yetIX Science Ch-12 Solutions (Improvement in Food Resources)6 pages
- The Soil Underfoot - Infinite Possibilities For A Finite Resource (Gnv64)100% (2)The Soil Underfoot - Infinite Possibilities For A Finite Resource (Gnv64)462 pages
- The Policy Challenges For Green Economy and Sustainable Economic DevelopmentNo ratings yetThe Policy Challenges For Green Economy and Sustainable Economic Development13 pages
- Estimate TRR SDF Roads & Drains 18-8-2023No ratings yetEstimate TRR SDF Roads & Drains 18-8-2023750 pages
- Submittal Data: End Suction Stainless Steel PumpsNo ratings yetSubmittal Data: End Suction Stainless Steel Pumps3 pages
- Health Waves: ※ 본 전국연합학력평가는 17개 시도 교육청 주관으로 시행되며, 해당 자료는 Ebsi에서만 제공됩니다. 무단 전재 및 재배포는 금지됩니다No ratings yetHealth Waves: ※ 본 전국연합학력평가는 17개 시도 교육청 주관으로 시행되며, 해당 자료는 Ebsi에서만 제공됩니다. 무단 전재 및 재배포는 금지됩니다6 pages
- Инструкция Panasonic KX-TCD150FXC (77 страницы)No ratings yetИнструкция Panasonic KX-TCD150FXC (77 страницы)3 pages
- Eric Parquet and Qun Lin - Microwave-Assisted Wolff-Kishner Reduction ReactionNo ratings yetEric Parquet and Qun Lin - Microwave-Assisted Wolff-Kishner Reduction Reaction1 page
- 001 Unit 3 Disasters and Triage Management FinalNo ratings yet001 Unit 3 Disasters and Triage Management Final45 pages