Beruflich Dokumente
Kultur Dokumente
UNDER CBCSS
2012-2015
By
Candidate Code
ANITHA ALEXANDER
ANUJA JOPPAN
22012126006
22012126003
MERIN.M
RESHMI.J.S
:22012126030
Subject Code
22012126026
:
Exam code:
Subject:
Mathematics
DEPARTMENT OF MATHEMATICS
ST.GREGORIOS COLLEGE
KOTTARAKKARA
MARCH -2015
A STUDY ON QUEEING THEORY
CERTIFICATE
I here bycertify that this dissertation is a bonafide record of work carried out by
Anitha Alexander , AnujaJoppan , Merin.M ,Reshmi.J.S inMathematics
(CBCSS 2012-2015) of this college under our supervision in partial
fulfilmentthe requirements for B.Sc Degreein Mathematics of University of
Kerala.
ACKNOWLEDGEMENT
Place:Kottarakara
ANITHA ALEXANDER
Date
ANUJA JOPPAN
MERIN.M
RESHMI.J.S
DECLARATION
Anitha Alexander,Anujajoppan,
Merin.M, Reshmi.J.Sand that no part thereof has been presented for any other
Degree.
ANITHA ALEXANDER
22012126003
ANUJA JOPPAN
22012126006
MERIN.M
22012126026
RESHMI.J.S
22012126030
CONTENTS
1. INTRODUCTION
2. PRELIMINARIES
3. QUEUE MODELS
4. M|M|1:(/FIFO) MODEL
5. BIBLIOGRAPHY
INTRODUCTION
Gauss referred to mathematics as the Queen of the Science. The word
corresponding to science means a field of knowledge.
Queues or waiting lines are very common in every life we quite often face the
problem of long queues for a bus, a movie ticket and for various other situations.
Long queues are generally seen infront of railway booking offices, post offices and
bank counters particularly in large cities similarly we also find automobiles waiting
at service station, ships waiting for berths, airplanes waiting for landing and patients
waiting for doctors. Queues are thus, a very common phenomenon of modern
civilized life.
Here we discuss about the important models and examples of queuing theory.
The dissertion divided into three chapters. In chapter one, we collect some basic
definitions in queuing theory which we needed in the project.
The second chapter deals with queue models, definitions classification and
some important examples of queuing theory.
The third chapter deals with a queue model M/M/l :( /FIFO). Here we derive
equations and important examples related to this topic.
Chapter-1
PRELIMINARIES
DEFINITION1.1
Random Experiment
An experiment whose outcome cannot be predicted is called random
experiment
Eg: 1.1
Tossing a coin is a random experiment whose outcomes are Head and Tail
which could not be predicted
Eg: 1.2
Throwing of a die is a similar case whose possible outcomes are 1, 2, 3, 4, 5 or 6
DEFINITION1.2
Sample space
The set of outcomes of a random experiment is known as sample space(s)
In tossing of a coin s= {Head, Tail}
In tossing of a die s= {1, 2, 3, 4, 5, 6}
Definition:1.3
A STUDY ON QUEEING THEORY
Events
In a random experiment the subsets of sample space is known as event or
cases. In such students null set is called impossible events and same set is called
sure event. Single ten subsets are called elementary events and others are composte
event.
Eg: 1.3
In a bag with 7 blue and 5 green balls the prohability of selecting a green ball
is 5/12 and that of a blue ball is 7/12
Definition 1.4
Random variable
If the numerical values assumed by a variable are the result of some chance
factors, so that a particular value cannot be exactly predicted in advance the variable
is then called a random variable. It is also called chance variable or stochastic
variable.
Eg: 1.4
If a coin is tossed thrice,let the random variable X: number of Heads. The X
can take the values 0, 1, 2 and 3
Definition: 1.5
Descrete and continuous random variable
Adescrete random variable is one which can assume only isolated values
such as natural numbers
Eg : 1.4 is such a case
A continuous randomvariablein one which can assume any value with in an
interval ie, all values of continuous scale. For example,
(i)
The weights of a group of individuals
(ii) The height of a giving of individuals are all continuous
Definition : 1.6
Probability distribution
A STUDY ON QUEEING THEORY
P(X)
P1
X2
P2
Xn
Pn
Pi=1 , then
i=1
Is called probability distribution for X and 1 spells out how a total probability of 1 is
distributed over several values of a random variable.
Eg: 1.7
In tossing of 2 coins at a time, the random variable X=no: of heads then its
probabilities is
X
0
P(X)
2/4
P(X)
Definition : 1.8
A STUDY ON QUEEING THEORY
10
Theoretical Distributions
Frequency distribution can be classified under two heads
(i)
Observed
(ii) Expected (Theoretical)
The first case is based on actual observation and experimentation. If certain
hypothesis is assumed, it is sometimes possible to derive mathematically what the
frequency distribution of certain universe should be such distribution are called
theoretical distribution
There are many type of theoretical distributions but we shall consider only two
which are taken into consider here
(i)
Poissons Distribution
(ii) Exponential Distribution
Definition: 1.9
Poissons Distribution
It is a discrete distribution; random variable X is said to follow the
poissons distribution if its probability density function is,
- x
Definition; 1.10
Exponential Distribution
A STUDY ON QUEEING THEORY
11
F(x) =
-x
, x0
0, otherwise
12
Chapter-2
Queue Models
Meaning of queue or waiting line
Ordinarly, the line that forms in front of servicefacilities is called a
queue or a waiting line. A queue, thus involves arriving customers who wait to be
serviced at the facility which provide the serviced at they want to have the word
queue refers to waiting line. The idea about a queue may be expressed as,
Customers
Arrived
possible waiting
line or queue
getting
serviced
departure of
customers
But for queuing theory purpose it may be remembered that the queue need not
be a physical line of customers it may be a dispersed list of persons.
Eg: waiting list for a berth on a train or for a trunk telephone call. Such lists are as
real as physical queues.
The following are some of the instances where we generally come across
waiting line problem.
13
Arrivals
Customer in
Waiting for
Service
Input/arrival process
Units/units
In service
Service mechanism
Or patterns/ facility
14
Units/unit
Departing
From system
output
15
Queuing models are those where a facility performs a service. Most elementary
queuing models assume that the input/arrivals and output/ departures follow a birth
and death process. Any queuing model is characterised by situations where both
arrivals and departures take place simultaneously. Depending upon the nature of
input and service facilities, there can be a number of queuing models. Queuing
models can be broadly classified into three categories.
Queuing Models
Probabilistic
Deterministic
Mixed
Either of the
arrival and service
is unknown random
variable andother
known and fixed
16
Classification of Queues
A queuing model is specified completly by the following six main
characteristic
(1) Input or arrival (inter arrival) distribution
(2) Output or departure (service) distribution
(3) Service channels
(4) Service discipline
(5) Maximum number of customers allowed in the system
(6)Calling source or population
(i) Input process
Balking
Size of arrivals
Finite population
Reneging
Customer behaviour
collusion
Jockeying
Infinite population
constant
Random
Input implies the mode of arrival of customers at the service facility. The
number of customers emanated from finite or infinite source typically customers
arrive at the system randomly singly or in batches. The input process is characterised
by the nature of the arrivals, capacity of the system and behaviour of the system.
A STUDY ON QUEEING THEORY
17
(A)
The size of customers arriving for servicing depends on the nature of the
population which can be finite or infinite. From practical view point a
population is considered to be finite, if the probability of an arrival is greatly
changed when one member of the population is already receving service
The period between the arrival of the individual customers may be
constant or scattered in some fashion .most queuing models assume that the
same inter arrival time distribution applies for all customers throughout the
period of study. The most convenient way is to designate some random
variables corresponding to the times between arrivals. In genera lthe arrivals
follow poisson distribution when the total number of arrivals during any
given time interval is independent of the number of arrivals that have already
prior to the beginning of the time interval
(B)
In many system the capacity of the space where the arrivals have to wait
before taken into service is limited. In such cases where the length of waiting
line crocess a certain limit no further units/arrivals are permitted to enter the
system till some waiting space becomes vacant such queue systems are known
as systems with finite capacity and considerably affects the arrival pattern of
the system
Eg : a doctor may give appointment to fixed number of patients each day.
(C)
The human behaviour and the facilities of servicing in any system are
important factors for the development of queuing problem. The behaviour of
the customers behaviour can be classified in follow categories
(i)Balking : A customer may not like to join the queue being it very long and
be maynot like to wait
(ii) Reneging : He may leave the que due to impatience after joining it
(iii)Collusion : several customers may collaborate ang only one of them may
stand in the queue
(iv)Jockeying : If there are number of queues then one may leave one queue to
join another.
18
the completion time of service. Service mechanism of any system mainly determined
by :
Service machanism
Single channel
Multi channel
FIFO/ FCFS FIFO/LCFS SIRO
Priority
queue
Multi phase
(A) Service facility design : The facilities at the service station can be divided in two
main categories
(i) Single channel (ii) Multi channel
(i) single channel queue : there may be only one counter for servicing and as such
only one unit can be served at a time . the next unit can be taken into service. When
the servicing operations on the previous unit are completed. The single channel queue
can be divided in two types.
(a) Single phase
In a single phase queue the whole service operations are completed in one
stage
(a) Single channel single phase queue
Input
Arrivals for
Service
Queue
Arrivals
Waiting for
Service
Service
One item
Being
served
19
Output
Departure
(b) Single channel multi phasequeue :- Here the unit taken for service has to pass
through many stage before the unit goes out of all servicing channel. All the
phases of service arranged in a ordered sequence
Single channel K phases arranged in series
Queue
Service I
(ii) Multi channel :- Due to rush of customers. Management may decide to provide a
number of service counters so that queue length may not become unreasonably large
and organisation may not loose customers due to long queue. But too many counters
may result in long idle time of counters due to shortage of customers.
(a) Multi channel Queue discipline with single phase:-
Input
Queue
Service
Queue
Service
Queue
Service
Output
Departure
20
Queue
Service
Queue
Input
Queue
Service
Queue
Service
Service
output
Queue
Service
B] queue/ service Discipline : Queue Discipline identifies the order in which arrivals
in the system are taken into service. The Queue discipline does not always take into
account the order of arrivals. Various methods are available to solve queuing
problems under different queuing but most of these introduce complications in the
analysis. The most common discipline in First in First out( FIFO) or first come first
served. Here the customers are serviced strictly in the order of their journey the
system.
eg: Queues at booking stations
The Last Come First Served[LCFS] or Last In First Out[LIFO] system is one
where the item arriving last are first go into service.
eg: In big stores the items arriving last are issued first.
Similarly in elevators passenger entering last may stand near the gate and thus
may leave first.
Service In Random Order[SIRO] rule implies that arrivals are taken in service
randomly irrespective of the order of their arrivals in the system. Here the server
chooses one of the customers to offer service at random.
21
22
eg: Variation in service facility with regard to number of service channels, service
rate and the number of service facilities will lead to a new queuing system.
Queuing models measure the effect of uncertainity on the behaviour of the Queuing
system
The solution of a queing problem consist in selective the best compramise for
the function which can be controled.
23
Chapter-3
M|M|1:(|FIFO) Model
M|M|1:(|FIFO) ie, a system with poisson input, exponential waiting time and
poisson output. With single channel. Queue capacity of system being infinite with
first out mode.
Here first M in the notation stands for poisson input
2nd M in the notation stands for poisson output.
1 in the notation stands for number of channel.
in the notation stands for infinite capacity.
FIFO in the notation stands for First in First out.
Situations
There are so many queue which follows this model some of such situations
are;
(i) Arrivals at a telephone both with a given time between two arrivals.
(ii)Arrivals of train in the platform with FIFO
(iii) Arrivals in booking situation such as railway, bus station etc
(iv) Arrivals of patients in the clinic.
Let there be n units in the system including the one in service at any time.
Ie, the queue length is n at time t with probability Pn(t). since the output
distributin is poisson, the chance that a unit arrives in the system during the interval
(t,t+t) is
n t=0(t)2
24
Similarly, the chance that any unit leaves the system during this interval will be n
t+0(t)2, because the output distribution is also poisson. The chance that more than
one unit arrive or leaves the system during the interval (t, t+t) will be of order
0(t)2. Events with probabilities involving (t)2 cannot happen.
Let the system be in a state En at time (t+t) with probability pn(t+t). Then
there can be following three possibilities the interval (t,t+t).
(i) System is in the state En-1 at time t and one unit arrives in the interval t and no
units leaves during this period.
Probability for this event=
(probability that there are n-1 units at time t) x probability (onr unit
arrives during the time t)
= Pn-1(t)+ n-1t x (1-n-1t)
(1)
(ii) System is in the state En at time t and no unit arrives and no unit leaves the
system during the interval t
Probability of this event=
Probability (there are n units at time t) x
Probabiliy (no unit arrives in time t) x
Probability (no unit leaves in time t)
= Pn(t) (1-nt) x (1-nt)
(2)
(iii) the system is in the state En+1 at time t and no unit arrives in time t, but
one unit leaves the system after the competition of service.
25
Probability for this event = Probability (there are n+1 units at time t) x(Probabiliy
that no unit arrives in time t) x(Probability that
one unit leaves in time t)
= Pn+1(t) (1-n+1t) (n+1t)
(3)
Naturally all these 3 cases are mutually exclusive. Hence the probability that
the system is in state En at time t+t will be equal to the sum of the probabilities in
these 3 cases and Hence
Pn(t+t)=Pn-1(t)n-1t(1-nt)+Pn(t)(1-nt)(1-nt)+ Pn-1(t)( 1-nt) n+1t
= Pn-1(t)n-1t+ Pn(t)- Pn(t)[ n+n] t+ Pn+1(t) n+t[neglecting terms of
2
order (t) ]
Pn(t+t)- Pn(t)= Pn-1(t) n-1t-(n+n) Pn(t)t+ Pn-1(t) n+1t
ie,
P n ( t+ t ) P n (t)
t
ie, lim
P n ( t+ t ) P n (t)
=Pn(t)=Pn-1(t)n-1 -(n+n) Pn(t)+ Pn+1(t) n+1 ----(4)
t
26
Now,
P0(t+t)= [(Probability that no units arrives int) x
(Probabiliy that there was no unit at time t)] +
[(Probability that there was one unit at time t ) x
Probability that one unit leaves but no unit arrives
in the timet)]
= P0(t) (1-0t) + P1(t) (1-1t) -1 t
= P0(t) -0 P0(t) t+ P1(t)1 t+ terms of 0(t)2
P10(t) = -0 P0(t)+ P1(t) 1...........................................(5)
Now, when the system reaches the steady state then we have,
t
lim
P1n(t)= 0
(6)
Also let n = , -n =
Hense under steady state of the system ,(4) and (5) reduce to
Pn-1-(+) Pn+ Pn+1 =0 .............................................(7)
P10 = - P0+ P1 =0
And : P1=(/) P0 = P0, where = (/) ..............................(8)
Now, from (7), Pn+1 = (+)Pn- Pn+1
Ie, Pn+1 = (/+1 ) Pn - / Pn-1= (+1) Pn-Pn+1
Putting n=-1,2,3 successively in the above relation
27
Now, we have
Pn+1= (+1) Pn-Pn-1= (+1) nPo-n-1 P0
=(+1) nPo-nPo
= nPo(+1-1)
=n+1 P0
But
Pn
n=0
=1 ;
n=0
nPo = 1
or, Po[1++2+............]=1
1
28
29
Meanarrivalrate
Meanservicer
The unit of is trlang. Any queuing system can settle down in a steady state only
when <1 ie,
<
can also be written as
1 /
1 /
Meanservicetime
Meaninterarrivaltime
If >1, then the queue length will go in increasing and will tend towards infinity
with time.
To show that average number of units in a M|M|1 system is equal to (1-)
E(n)= Average queue length in the system
n n
=
P
n=0
P0 n
n=0
30
..........................................(11)
31
1/=
6024
30
= 48 minites
1/= 36 minites
ie, =(/) = 0.75
The mean queue size,
E(n) =
0.75
= 300.25
=
3trains
32
E(L) =
n=0
Pn -
n=0
n=1
E(n)-
P
n=1
P
1 n=1
n+ P0[since E(n) =
= 1 1+(1) (as
2
1
P
n=0
=1)
........................................(12)
Example . 2 :
Let an average 96 patients per 24 hour day require the service of an emergancy clinic.
Also an average, a patient requires 10 minites of active attention. Assume that the
facility can handle only one emergancy at a time. Suppose that it costs the clinic Rs.
100 per patient treated to obtain an average servicing time of 10 minutes and that
each minute od decrease in this average time would cost Rs. 10 per patient treated.
How much have to be budgeted by the clinic to decrease the average size of the
queue from one third patient to half a patient?
33
Solution:
Here we have =
96
24
= 4 patients/ hour
= 3
1
2
To show that average length of the waiting line with the condition that it is always
greater than zero is (1/1-)
E(L/L>0) =
E( L)
= P(n>1)
E (L)
P(0)
n>1 ]
= 1
P
n=2
34
2
= 1
( 2 P 0+ + )
1
3P 0
1
2
P0
1
P0
E[L|L>0] = 1
.........................................(13)
But E(n2)=
n
n=0
Pn
n
n=0
= P0[12P+22P2+32P2+................... terms]
= P0[12+222+322+........]
Let
= 1+22P+32P3+.............................
35
. dp
= +22+32+........
= [1+2+32.....]
= (1-)-2
Differentiating w.r.t, we get
1
(1)2
2
(1)3
(1)+2 P
= (1)3
(1+ )
(1)3
(1+ P)
(1+ )
(1)2
+
2
(1)
V (n) =
(1+ )
(1)2
2
- (1)2
2
(1) ...........................(14)
36
= /
Traffic intensity.
P0 =(1- )
Pn = n P0
E(n) = 1
E(L) = 1
line.
1
E(L/L>0) = 1
V(n) = (1)2
P(w>0) =
P(w/w>0) = (1-) e-w(1-)dw
P(v)dv = (1-) e-w(1-) dw :
system.
E(w) = (1)
37
1
E(v) = (1)
1
E(w/w>o) = = (1)
BIBLIOGRAPHY
38