PERFORMANCE ANALYSIS USING SOME

QUEUEING MODELS
Synopsis submitted in fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
By
NEETU GUPTA
Department of Mathematics
JAYPEE INSTITUE OF INFORMATION TECHNOLOGY UNIVERSITY
A-10, SECTOR-62, NOIDA, INDIA
March 2010
1
INTRODUCTION:
Queueing theory is the mathematical study of waiting lines or queue. The theory enables
mathematical analysis of several related processes, including arriving at the queue, waiting
in the queue and being served by the server(s) at the front of the queue. The theory permits
the derivation and calculation of several performance measures including the average
waiting time in the queue or the system, the expected number waiting or receiving service
and the probability of encountering the system in certain states, such as empty, full, having
an available server or having to wait a certain time to be served.
Queueing theory is generally considered a branch of operations research because the results
are often used when making business decisions about the resources needed to provide
service. There are many valuable applications of the theory, most of which have
been well documented in the literature of probability, operations research,
management science, and industrial engineering. Some examples are traffic flow
(vehicles, aircraft, people, communications), scheduling (patients in hospitals, jobs on
machines, programs on a computer), and facility design (banks, post offices,
amusement parks, fast-food restaurants).
CHARACTERISTICS OF QUEUEING PROCESSES:
In most cases, six basic characteristics of queueing processes provide an adequate
description of a queueing system:
(1) Arrival pattern of customers,
(2) Service pattern of servers,
(3) Queue discipline,
(4) System capacity,
(5) Number of service channels,
(6) Number of service stages.
NOTATIONS OF QUEUEING MODELS:
As a shorthand for describing queueing processes, a notation has evolved, due for
the most part to Kendall (1953), which is now rather standard throughout the
queueing literature. A queueing process is described by a series of symbols and
slashes such as A/B/X/Y/Z, where A indicates in some way the interarrival- time
distribution, B the service pattern as described by the probability distribution for
service time, X the number of parallel service channels, Y the restriction on system
capacity, and Z the queue discipline. Some standard symbols for these characteristics are
presented in Table.
2

Table of Queueing Notation A/B/X/Y/Z

Characteristic Symbol Explanation
Interarrival-time distribution (A)
Service-time distribution (B)
Number of parallel servers (X)
Restriction on system capacity (Y)

Queue discipline (Z)
M
D
E
k
H
k
PH
G
1,2,,
1,2,,
FCFS
LCFS
RSS
PR
GD
Exponential
Deterministic
Erlang type k (k=1,2,..)
Mixture of k exponentials
Phase type
General
First come, first served
Last come, first served
Random selection for service
Priority
General discipline
PERFORMANCE OF QUEUEING MODELS:
Up to now the concentration has been on the physical description of the queueing
processes. Generally there are three types of system responses of interest. These are:
(1) some measure of the waiting time that a typical customer might be forced to
endure; (2) an indication of the manner in which customers may accumulate; (3) a
measure of the idle time of the servers. Since most queueing systems have
stochastic elements, these measures are often random variables and their probability
distributions, or at the very least their expected values, are desired. There are two
types of customer waiting times, the time a customer spends in the queue and the
total time a customer spends in the system (queue plus service). Depending on the
system being studied, one may be of more interest than the other. For example, if we
are studying an amusement park, it is the time waiting in the queue that makes the
customer unhappy. On the other hand, if we are dealing with machines that require
repair, then it is the total down time (queue wait plus repair time) that we wish to
keep as small as possible. Correspondingly, there are two customer accumulation
measures as well: the number of customers in the queue and the total number of
customers in the system. The former would be of interest if we desire to determine
a design for waiting space (say the number of seats to have for customers waiting
in a hair- styling salon), while the latter may be of interest for knowing how many
of our machines may be unavailable for use. Idle- service measures can include the
percentage of time any particular server may be idle, or the time the entire system
is devoid of customers. The task of the queueing analyst is generally one of the
two things. He or she is either to determine the values of appropriate measures of
effectiveness for a given process, or to design an optimal (according to some
3
criterion) system. To do the former, one must relate waiting delays, queue lengths,
and such to the given properties of the input stream and the service procedures. On
the other hand, for the design of a system the analyst might want to balance
customer waiting time against the idle time of servers according to some inherent
cost structure. If the costs of waiting and idle service can be obtained directly, they
can be used to determine the optimum number of channels to maintain and the
service rates at which to operate these channels. Also, to design the waiting facility
it is necessary to have information regarding the possible size of the queue to plan
for waiting room. There may also be a space cost which should be considered
along with customer- waiting and idle- server costs to obtain the optimal system
design. In any case, the analyst will strive to solve this problem by analytical
means; however, if these fail, he or she must resort to simulation. Ultimately, the
issue generally comes down to a trade- off of better customer service versus the
expense of providing more service capability, that is, determining the increase in
investment of service for a corresponding decrease in customer delay.
CHAPTER-1
PERFORMANCE ANALYSIS OF AN M/M/1/N QUEUE WITH
BALKING AND RENEGING:
1.1. INTRODUCTION:
In real life, many queueing situations arise in which there may be a tendency for customers
to be discouraged by a long queue. As a result, the customers either decide not to join the
queue (i.e. balk) or depart after joining the queue without getting service due to impatience
(i.e. renege). Balking and reneging are not only common phenomena in queues arising in
daily activities, but also in various machine repair models. Many practical queueing
systems especially those with balking and reneging have been widely applied to many real-
life problems, such as the situations involving impatient telephone switchboard customers,
the hospital emergency rooms handling critical patients, and the inventory systems with
storage of perishable goods [1]. In this chapter, we consider an M/M/1/N queueing
system with balking and reneging.
Queue system M/M/1/N means: arrival of queue is Poisson, departure of queue is
Exponential, number of server is one and capacity of system is N. Queueing systems with
balking, reneging or both have been studied by many researchers. Haight [2] first
considered an M/M/1 queue with balking. An M/M/1 queue with customers reneging was
also proposed by Haight [3]. The combined effects of balking and reneging in an M/M/1/N
queue have been investigated by Ancker and Gafarian [4], [5]. Abou-EI-Ata and Hariri [6]
considered the multiple servers queueing system M/M/c/N with balking and reneging.
In the next part of this chapter, we give a description of the queueing model. In part 1.3 of
this chapter, we derive the steady-state equations by the Markov process (It is a random
process which has the property that the probability of transition from a given state to any
state depend only on the present state and not on the manner in which it was reached.)
method. By writing the transition rate matrix as block matrix, we get the matrix form
solution of the steady-state probabilities and present a procedure for calculating the steady-
4
( )
( ) ( ) ( ) ( ) [ ]
( ) ( ) [ ] . ), ( 1 1
. 1 , , 2 , 1 ), ( 1 1 1
0 , 0 ) 1 (
1
1
N n N P n N P b
N n n P n b n P n n P b
n P P
N
n n
+
+ + + + +

state probabilities. In part 1.4 of this chapter, we give some performance measures of the
system. Based on the performance analysis, we formulate a cost model to determine the
optimal service rate. Some numerical examples are presented to demonstrate how the
various parameters of the model influence the behavior of the system. Conclusion is given
in last of this chapter.
1.2. SYSTEM MODEL
In this chapter, we consider an M/M/1/N queueing system with balking and reneging. The
assumptions of the system model are as follows:
1. Customers arrive at the system one by one according to a Poisson process with rate .
On arrival a customer either decides to join the queue with probability b
n
or balk with
probability 1- b
n
when n customers are ahead of him (n=0,1,N-1), where N is the
maximum numbers of the customers in the system, and
0 b
n-1
b
n
<1, 1 n N-1,
b
0
=1, and b
n
= 0, n N.
2. After joining the queue each customer will wait a certain length of time T for service to
begin. If it has not begun by then, he will get impatient and leave the queue without getting
service. This time T is a random variable whose density function is given by
d(t) = e
-t
, t 0, >0,
where is the rate of time T. Since the arrival and departure of the impatient customers
without service are independent, the average reneging rate of the customer can be given by
(n-i). Hence, the function of customers average reneging rate is given by
r(n) = (n-i), i n N, i= 0,1
r(n) = 0, n>N.
3. The customers are served on a first-come, first served (FCFS) discipline. Once service
commences it always proceeds to completion. The service times are assumed to be
distributed according to an exponential distribution with density function as follows:
s(t) = e
-t
, t 0, > 0, where is the service rate.
1.3. STEADY STATE PROBABILITY
In this part, we derive the steady-state probabilities by the Markov process method. Let
P(n) be the probability that there are n customers in the system when the server is on
available. Applying the Markov process theory, we obtain the following set of steady-state
equations.
5
( ) ( ) ( )
( ) ( )
( )
( ) ( ) ) 4 . 1 ( 1 . .
) 3 . 1 ( ) 1 ( . .
) 2 . 1 (
) 1 . 1 ( 1
1
1
1
1

N
n
N
n
n
N
n
N
n
q
n P n R R
n P b R B
n nP N E
n P n N E

( )
( ) ( ) ( ) [ ]
( ) ( ) ( ) [ ]
( ) ( )
.
3 ), 3 ( 2 2
2 ), 2 ( 3 2 1
1 ), 1 ( 2 0
0 , 0 ) 1 (
2
2 1
1
+
+ + + +
+ + +

n P P b
n P b P P b
n P b P P
n P P




For N=3 steady state difference equations are:
On solving these we get:
P(0) = k
P(1) = k /
P(2) =
2
b
1
k / (+)
P(3) =
3
b
1
b
2
k / (+)(+2)
Using
P(0)+P(1)+P(2)+P(3) = 1 ,
We get
k = [1+ ( / ) + ((
2
b
1
)

/ (+)) + ((
3
b
1
b
2
)

/ (+)(+2))]
(-1)
In general for N = n
P(n) = (
n
b
1
b
2
...b
n-1
k)

/ ((+)(+2). (+(n-1)))
And using
P(0)+P(1)+P(2)+P(3)+P(4)++P(n) = 1 ,
We get
k = [1+ ( / ) + ((
2
b
1
)

/ (+)) + ((
3
b
1
b
2
)

/ (+)(+2))+ +(
n
b
1
b
2
..b
n-1
)

/
((+)(+2). (+(n-1)))]
(-1)
1.4. PERFORMANCE MEASURES AND COST MODEL
In this part, we give some performance measures of the system. Based on these
performance measures, we develop a cost model to determine the optimal service rate.
1.4.1. PERFORMANCE MEASURES
Using the steady-state probability presented in part 1.3 of this chapter, we can obtain some
performance measures of the system, such as the busy probability of the server P
B
, the
expected number of the waiting customers E(N
q
) and the expected number of the customers
in the system E(N) as follows:



6






L.R. = B.R. + R.R. (1.5)

1.4.2. COST MODEL
In this subpart, we develop an expected cost model, in which service rate is the control
variable. Our objective is to control the service rate to minimize the systems total average
cost per unit. Let
C
1
cost per unit time when the server is busy,
C
2
cost per unit time when a customer joins in the queue and waits for service,
C
3
cost per unit time when a customer balks or reneges.
Using the definitions of each cost element listed above, the total expected cost function per
unit time is given by
F() = C
1
P
B
+ C
2
E(N
q
) + C
3
L.R.,
where E(N
q
), E(N), B.R., R.R. and L.R. are given in Eqs. (1.1) (1.5) .The first item is the
cost incurred by the server. The second item C
2
E(N
q
) is the cost incurred by the customers
waiting. The last item C
3
L.R. is the cost incurred by the customer loss.
1.4.3. NUMERICAL RESULTS
In this subpart, we present some numerical examples to demonstrate how the various
parameters of the model influence the optimal service rate *, the optimal expected cost of
the system F(*) and other performance measures of the system. We fix the maximum
number of customers in the system N = 3, the probability b
n
= 1/(n + 1) and the cost
elements C
1
= 15, C
2
= 12, C
3
= 18.
Table 1.1.The case for = 0.1
0.4 0.5 0.6 0.7 0.8 0.9
* 0.9468 0.9089 0.8754 0.8444 0.8155 0.7882
F(*) 17.2026 18.4446 19.9224 21.6150 23.4906 25.5186
E(N
q
) 0.0658 0.1039 0.1495 0.2016 0.2592 0.3211
E(N) 0.4047 0.5180 0.6327 0.7481 0.8627 0.9757
L.R. 0.0785 0.1221 0.1738 0.2331 0.2989 0.3703
7

Graph between and * , E(Nq) ,E(N), L.R.

0
0.2
0.4
0.6
0.8
1
1.2
0 0.2 0.4 0.6 0.8 1

*(Opimal srevice rate)
E(Nq): Length of queue
E(N): Length of queueing
system
L.R.(Average rate of
customer loss)
Graph between arrival rate and expected optimal
cost function
0
5
10
15
20
25
30
0 0.2 0.4 0.6 0.8 1


F
(

*
)
Table1.2. The case for = 0.5
0.05 0.1 0.2 0.3 0.4 0.5
* 0.9652 0.9089 0.8010 0.6993 0.6041 0.5160
F (*) 18.2430 18.4572 18.9180 19.4208 19.9734 20.5152
E(N
q
) 0.1006 0.1039 0.1105 0.1173 0.1239 0.1302
E(N) 0.5002 0.5180 0.5550 0.5956 0.6830 0.6825
L.R. 0.1131 0.1228 0.1440 0.1674 0.1937 0.2196
First, we select the rate of the waiting time = 0.1 and change values of arrival rate of
customers . The numerical results are summarized in Table 1.1. This shows that: (i) the
optimal service rate decreases with the increasing of and its minimum expected cost
F(*) increases with the increasing of ; (ii) the expected number of the waiting customers
E(N
q
), the expected number of customers in the system E(N) and the average rate of
customer loss L.R. all increase with the increasing of . This is because the number of the
customers in the system increases with the increasing of . Thus, E(N
q
) and L.R. all
increase which result in the increasing of the optimal cost.
Finally, we select = 0.5, and change values of . The numerical results are summarized in
Table 1.2. This shows that: (i) the optimal service rate * decreases with the increasing of
, and its minimum expected cost F(*) increases with the increasing of ; (ii) E(N
q
),
E(N) and L.R. all increase with the increasing of . Thus E(N
q
) and L.R. all increases
which result in the increasing of the optimal cost. The following graph is showing the
performance of the model at = 0.1 and = 0.5. The below Figure 1.1 and 1.2 are for =
0.1 and these are showing the performance of the model. Similarly, Figure 1.3 and 1.4 are
for = 0.5 and these are showing the performance of the model.

Fig. 1.1
8
Graph between and * , E(N
q
) ,E(N), L.R.
0
0.2
0.4
0.6
0.8
1
1.2
0 0.1 0.2 0.3 0.4 0.5 0.6

*(Opimal srevice rate)

E(Nq): Length of queue
E(N): Length of queueing
system
L.R.(Average rate of
customer loss)
Graph between and F(*)
18
18.5
19
19.5
20
20.5
21
0 0.1 0.2 0.3 0.4 0.5 0.6

F
(

*
)

Fig. 1.2
Fig. 1.3
9

Fig. 1.4
1.5. CONCLUSION
In this chapter, we considered an M/M/1/N queueing system with balking and reneging.
We developed the equations of the steady state probabilities and derived the matrix form
solution of the steady-state probabilities. We also gave some performance measures of the
system, and formulated a cost model to determine the optimal service rate. Although the
function of the cost is too complicated to derive the explicit expression of the optimal
service rate, the performance measures and the optimal service rate can be numerically
evaluated by the formula in Section 1.4.2. Some numerical examples were presented to
demonstrate how the various parameters of the model influence the behavior of the system
as given above.
CHAPTER-2
PERFORMANCE ANALYSIS OF AN M/M/c/N QUEUE WITH
BALKING AND RENEGING:
2.1. INTRODUCTION:
We have found performance of M/M/1/N queueing system with balking and reneging. In
M/M/1/N model no. of server is one. In this chapter, we consider an M/M/c/N queueing
system with balking and reneging. In M/M/c/N model no. of server is c and each server
has an independently and identically distributed exponential service time distribution.
Queueing systems with balking, reneging or both have been studied by many researchers.
Haight [2] first considered an M/M/1 queue with balking. An M/M/1 queue with customers
reneging was also proposed by Haight [3]. The combined effects of balking and reneging in
an M/M/1/N queue have been investigated by Ancker and Gafarian [4], [5]. Abou-EI-Ata
and Hariri [6] considered the multiple servers queueing system M/M/c/N with balking and
10
reneging. Queue system M/M/c/N means: arrival of queue is Poisson, departure of queue is
Exponential, number of server is c and capacity of system is N.
In the next part of this chapter, we give a description of the queueing model. In part 2.3 of
this chapter, we derive the steady-state equations by the Markov process method. By
writing the transition rate matrix as block matrix, we get the matrix form solution of the
steady-state probabilities and present a procedure for calculating the steady-state
probabilities. In part 2.4 of this chapter, we give some performance measures of the system.
Based on the performance analysis, we formulate a cost model to determine the optimal
service rate. Some numerical examples are presented to demonstrate how the various
parameters of the model influence the behavior of the system. Conclusion is given in part
2.5 of this chapter.
2.2. SYSTEM MODEL
In this chapter, we consider an M/M/c/N queueing system with balking and reneging. The
assumptions of the system model are as follows:
1. Customers arrive at the system one by one according to a Poisson process with rate .
On arrival a customer either Rdecides to join the queue with probability b
n
or balk with
probability 1- b
n
when n customers are ahead of him (n=0,1,N-1), where N is the
maximum numbers of the customers in the system, and
0 b
n-1
b
n
<1, 1 n N-1,
b
0
=1, and b
n
= 0, n N.
2. After joining the queue each customer will wait a certain length of time T for service to
begin. If it has not begun by then, he will get impatient and leave the queue without getting
service. This time T is a random variable whose density function is given by
d(t) = e
-t
, t 0, >0,
where is the rate of time T. Let i denote the number of severs being busy and n represent
the number of customers in the system. If n is less than or equal to i, the customers will get
service instantly upon arrival to the server, and the phenomenon of reneging will not occur.
If n is greater than i, then there are ni customers who have to wait in the queue. Since the
arrival and the departure of the impatient customers without service are independent, the
average reneging rate in this state is given by (ni) .
3. The customers are served on a first-come, first served (FCFS) discipline. Once service
commences it always proceeds to completion. The service times are assumed to be
distributed according to an exponential distribution with density function as follows:
s(t) = e
-t
, t 0, > 0, where is the service rate.
2.3. STEADY-STATE PROBABILITY
In this part, we derive the steady-state probabilities by the Markov process method. Let
P(n) be the probability that there are n customers in the system. Applying the Markov
process theory, we obtain the following set of steady-state equations.
P(1) = P(0) , n = 0,

P(n-1) + (n+1) P(n+1) = ( +n) P(n) , 1 n c-1,
11
(c+ )

P(c+1 ) +

P(c-1) = (c+ b
c
) P (c), n = c,

b
n-1
P(n-1)

+ [c +(n+1-c) ]P(n+1) = [c + b
n
+(n-c) ]P(n) c < n < N,
b
N-1
P(N-1) =

[c

+(N-c) ]P(N) n = N.
For N=4 and c=2 steady state difference equations are:
P(1) = P(0), n = 0,
P(0)+2 P(2) = (+) P(1), n = 1,

P(1)+(2+) P(3) = ( b
2
+2) P(2) , n = 2,
b
2
P(2) + (2+2)P(4) = ( b
3
+2+ ) P(3), n = 3,
b
3
P(3) = (2+2)P(4), n = 4.

On solving these we get
P(0) = k
P(1) = k /
P(2) =
2
k / 2
2

P(3) =
3
b
2
k /2
2
(2+)
P(4) =
4
b
2
b
3
k /2
2
(2+) (2+2)
Using P(0)+P(1)+P(2)+P(3)+P(4)=1 ,
We get
k = [1+( / ) + (
2

/ 2
2
) + (
3
b
2
/ 2
2
)

+ (
4
b
2
b
3
/2
2
(2+) (2+2)) ]
(-1)
For N = n
P(n) = (
n
b
2
b
3..
b
(n-1)
k) /(2
2
(2+) (2+2).. (2+(n-2)))
Using P(0)+P(1)+P(2)+P(3)+P(4)............+P(n)=1 ,
We get
k=[1+( / ) + (
2

/ 2
2
) + (
3
b
2
/ 2
2
)

+ (
4
b
2
b
3
/2
2
(2+) (2+2)).........+ (
n
b
2
b
3..
b
(n-
1)
) /(2
2
(2+) (2+2).. (2+(n-2)))]
(-1)
12
( ) ( )
( )
( )
( )
( ). 5 . 2 . . . . .
4 . 2 ) ( ) 1 ( .
3 . 2 ) ( ) 1 ( . .
2 . 2 ) ( ) (
1 . 2 ) ( ) 1 (
1
1
1
1
R R R B R L
n P n R R
n P b R B
n nP N E
n P n N E
N
n
N
n
n
N
n
N
n
q
+

2.4. PERFORMANCE MEASURES AND COST MODEL

In this part, we give some performance measures of the system. Based on these
performance measures, we develop a cost model to determine the optimal service rate.
2.4.1. PERFORMANCE MEASURES
Using the steady-state probability presented in the part 3 of this chapter, we can obtain
some performance measures of the system, such as the busy probability of the server P
B
,
the expected number of the waiting customers E(N
q
) and the expected number of the
customers in the system E(N) as follows:





2.4.2. COST MODEL
In this subpart of this chapter, we develop an expected cost model, in which service rate
is the control variable. Our objective is to control the service rate to minimize the systems
total average cost per unit. Let
C
1
cost per unit time when the server is busy,
C
2
cost per unit time when a customer joins in the queue and waits for service,
C
3
cost per unit time when a customer balks or reneges.
Using the definitions of each cost element listed above, the total expected cost function per
unit time is given by
F() = C
1
P
B
+ C
2
E(N
q
) + C
3
L.R.
where E(N
q
), E(N), B.R., R.R. and L.R. are given in Eqs. (2.1) (2.5) .The first item is the
cost incurred by the server. The second item C
2
E(N
q
) is the cost incurred by the customers
waiting. The last item C
3
L.R. is the cost incurred by the customer loss.
2.4.3. NUMERICAL RESULTS
In this subpart of this chapter, we present some numerical examples to demonstrate how
the various parameters of the model influence the optimal service rate *, the optimal
expected cost of the system F(*) and other performance measures of the system. We fix
13

Graph between and * , E(Nq) ,E(N), L.R.

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2 0.4 0.6 0.8 1

*(Opimal srevice rate)

E(Nq): Length of queue
E(N): Length of queueing
system
L.R.(Average rate of
customer loss
the maximum number of customers in the system N = 3, the probability b
n
= 1/(n + 1) and
the cost elements C
1
= 15, C
2
= 12, C
3
= 18.
Table 2.1.The case for = 0.1
0.4 0.5 0.6 0.7 0.8 0.9
* 0.1301 0.1549 0.1929 0.2479 0.3060 0.3647
F(*) 21.3084 22.6902 23.466 23.6616 23.8824 24.1074
E(N
q
) 0.2296 0.2615 0.2630 0.2403 0.2215 0.2074
E(N) 0.6193 0.6985 0.7000 0.6406 0.5913 0.5546
L.R. 0.1974 0.2529 0.2950 0.3210 0.3458 0.3712
Table2.2. The case for = 0.4
0.05 0.1 0.2 0.3 0.4 0.5
* 0.1559 0.1301 0.0901 0.0581 0.0352 0.0212
F(*) 24.198 21.3084 22.3338 23.1978 23.8278 24.1056
E(N
q
) 0.2154 0.2296 0.2402 0.2414 0.2312 0.2122
E(N) 0.5759 0.6193 0.6591 0.6715 0.6505 0.6023
L.R. 0.3674 0.1974 0.2473 0.2945 0.3363 0.3644
First, we select the rate of the waiting time = 0.1 and change values of arrival rate of
customer .The numerical results are summarized in Table 2.1. This shows that: (i) the
optimal service rate increases with the increasing of and its minimum expected cost
F(*) increases with the increasing of ; (ii) the expected number of the waiting customers
E(N
q
), the expected number of customers in the system E(N) first increase with the
increasing of and then decreases with the increases of ; (iii) average rate of customer
loss L.R. increase with the increasing of this is because the number of the customers in
the system increases with the increasing of . Thus, F(*) increases with the increasing of
.
Finally, we select = 0.4, and change values of . The numerical results are summarized
in Table 2.2. This shows that: (i) the optimal service rate * decreases with the increasing
of , and its minimum expected cost F(*) increases with the increasing of ; (ii) E(N
q
),
the expected number of customers in the system E(N) first increase with the increasing of
and then decreases with the increases of . Thus F(*) first decrease with the increasing
of and then increases with the increases of . The following graph is showing the
performance of the model at = 0.1 and = 0.4. The below Figure 2.1 and 2.2 are for =
0.1 and these are showing the performance of the model. Similarly, Figure 2.3 and 2.4 are
for = 0.4 and these are showing the performance of the model.
14
Graph between and F(*)
21
21.5
22
22.5
23
23.5
24
24.5
0 0.2 0.4 0.6 0.8 1

F
(

*
)
Graph between and F(*)
21
21.5
22
22.5
23
23.5
24
24.5
0 0.1 0.2 0.3 0.4 0.5 0.6

F
(

*
)
Graph between and * , E(Nq) ,E(N), L.R.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.1 0.2 0.3 0.4 0.5 0.6

*(Opimal srevice rate)
E(Nq): Length of queue
E(N): Length of queueing
system
E(N): Length of queueing
system
Fig. 2.1
Fig. 2.2
Fig. 2.3
15
Fig. 2.4
2.5. CONCLUSION
In this chapter, we considered an M/M/c/N queueing system with balking and reneging.
We developed the equations of the steady state probabilities and derived the matrix form
solution of the steady-state probabilities. We also gave some performance measures of the
system, and formulated a cost model to determine the optimal service rate. Some numerical
examples were presented to demonstrate how the various parameters of the model
influence the behavior of the system. Although the function of the cost is too complicated
to derive the explicit expression of the optimal service rate, the performance measures and
the optimal service rate can be numerically evaluated by the formula in Section 2.4.2. Some
numerical examples were presented to demonstrate how the various parameters of the
model influence the behavior of the system as given above.

CHAPTER-3
PERFORMANCE ANALYSIS OF AN M/M/1/k QUEUE WITH NON-
PREEMPTIVE PRIORITY:
3.1. INTRODUCTION
In chapter 1 and 2 we discussed the case of balking and reneging on the model M/M/1/N
and M/M/c/N. Now we take criterion of priority, obviously a considerable portion of real-
life queueing situations contains priority considerations. In chapter 3 and 4 we will discuss
the case of different priorities. The priority discipline followed may be either non-
preemptive or preemptive in nature. When the priority discipline is non-preemptive in
nature, a job in service is allowed to complete its service normally even if a job of higher
priority enters the queue while its service is going on. In the preemptive case, the service to
the ongoing job will be preempted by the new arrival of higher priority. If the priority
discipline is preemptive resume, then service to the interrupted job, when it restarts,
continues from the point at which the service was interrupted. For the preemptive non
resume case, service already provided to the interrupted job is forgotten and its service is
16
started again from the beginning. Note that there may be loss of work in the preemptive
non-resume priority case. Such loss of work will not happen in the case of the other two
priorities. Since the service times are assumed to be exponentially distributed, they will
satisfy the memory-less property and that, therefore, the results will be the same both for
the preemptive resume and preemptive non-resume cases. We consider here queues with
preemptive priority disciplines. The approach suggested may be applied to both single
server and multi-server queues. Moreover, queues with both finite and infinite buffers may
be analysed using the suggested approach.
The solution of queueing system is given by L. Kleinrock [7] in 1975.The idea about the
specially push out priority queue discipline given by D. P. Heyman [8] in 1985.After that
the performance evaluation of queueing system shown by H. Takagi [9] in 1991 and
mixing time and loss priorities in a single server queue is presented by A. Gravey and G.
Hebuterne [10] in1991.
Many papers of queueing model on priority is given by many authors time to time. The
idea of finite buffer on two class queue with different schemes is shown by X. Cheng and
Akyildiz [11] in 1992. In year 2006 Takagi [12] described explicit delay distribution in first
come first served M/M/m/k and M/M/m/K/n queue and a mixed loss delay system. We
have also worked on performance analysis of queueing model [13], [14].
In this chapter we define the way in which such a queue may actually be analysed by
solving its balance equations. As usual with any M/M/1/k queue, the basic approach to its
solution is to define the system-states appropriately so that a state transition diagram of the
queue may be drawn. Using the exponential (memory-less) distributions of the inter-arrival
and service times, the probability flows in the state transition diagram are identified. The
balance equations are then obtained by choosing proper closed boundaries and equating the
flow across each such boundary. These balance equations are then solved along with the
normalisation condition (i.e. sum of the probabilities of all states equal to unity) to obtain
the state probabilities. Once these have been obtained, the performance parameters of
interest may be suitably calculated. In the following, we consider the case of non-
preemptive priority. For each case, we first give the analytical approach for the M/M/1 case
and then show how it may be used to handle the M/M/1/K queue and the M/M/n/K queue.
Queue system M/M/1/k means: arrival of queue is Poisson, departure of queue is
Exponential, number of server is one and capacity of system is k.
3.2. ASSUMPTIONS
In this chapter, we consider an M/M/1/K queueing system with non-preemptive priority.
The assumptions of the system model are as follows:
1. The system at any instant of time is one of the two modes of operation.
2. All distribution are negative exponential, ie, all transition rates are constants.
3. The probability of two or more random events occurring simultaneously is negligible.
3.3. NOTATIONS
i mode of system operation, i = 1, 2
N maximum number of customers in the system
17
1 1 , 0 , 1 1 1 , 0 , 2
1 2 , 1 , 0 2 2 , 1 , 1
2 2 , 1 , 0 2 2 , 2 , 0
2 1 , 0 , 1 1 1 , 1 , 1
2 2 , 1 , 1 1 1 , 0 , 2 1 0 1 2 1 1 , 0 , 1
2 2 , 2 , 0 2 0 2 2 1 2 , 1 , 0
1 1 , 0 , 1 2 2 , 1 , 0 2 1 0
) (
) (
) (







p p
p p
p p
p p
p p p p
p p p
p p p

+ + + +
+ + +
+ +
1
2
1 , 1 , 1
1
1
1 , 0 , 1
1 , 0 , 2

k
p
k
p
k p

( )
( )
( )
( )
2 2 1 1 2
2
2
1
2 , 1 , 1
2 2 1
2
1 2
2
2
2
1
2 , 2 , 0
2 2 1
2
1
2 1
2
1
0
2 2 1
2
1
2
2
1
2 , 1 , 0
) (








+ +

+ +

+ +
+

+ +

k
p
k
p
k
p
k
p
n current number of customers in the system

i
arrival rate in mode i

i
service rate in mode i
L mean number in the system
W mean waiting time in the system
L
q
mean number in the queue
W
q
mean waiting time in the queue
3.4. SYSTEM MODEL

We first consider the 2- priority M/M/1 queue with infinite buffers operating with a non-
preemptive priority discipline. Since all the flows are known, the equilibrium probabilities
of each of the states may be found by writing and solving the corresponding balance
equations with the normalisation condition. As an example, consider the 2- priority
M/M/1/2 queue with non-preemptive priority.

3.5. STEADY-STATE PROBABILITY
In this part of this chapter, we derive the steady-state probabilities by the Markov process
method. Let P(n) be the probability that there are n customers in the system. Applying the
Markov process theory, we obtain the following set of steady-state equations.

On solving these equation we get;

18
1
1 , 0 , 2 2 , 1 , 1 1 , 1 , 1 1 , 0 , 1 2 , 2 , 0 2 , 1 , 0 0
+ + + + + + p p p p p p p
( ) ( ) ( )
( )
1
2 2 1 1 2
2
2
1
2 2 1
2
1 2
2
2
2
1
2 2 1
2
1
2 1
2
1
2 2 1
2
1
2
2
1
1
2
1
1
) (
1

1
]
1

+ +
+
+ +
+
+ +
+
+
+ +
+ + +

k
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0.1 0.2 0.3 0.4 0.5 0.6 0.7
L(Length of
queueing
system)
Lq(Length of
queue)
W(Waiting time
of queueing
system)
Wq(Waiting time
of queue)
Using :

We get:
3.5.1. NUMERICAL RESULTS
In this subpart of this chapter, we present some numerical examples to demonstrate how
the various parameters of the model influence the performance measures of the system. We
fix the maximum number of customers in the system N = 2.
Table 3.1.The case for
1
/
1
= 0.8

2
/
2
0.1 0.2 0.3 0.4 0.5 0.6 0.7
L 0.9206 0.9877 1.0469 1.1213 1.1473 1.1952 1.2369
L
q
0.2954 0.3275 0.3574 0.3926 0.4107 0.4362 0.4593
W 4.1849 4.1157 4.0266 4.0050 3.8240 3.7352 3.6380
W
q
1.3430 1.3645 1.3747 1.4022 1.3689 1.3629 1.3509
First, we select the
1
/
1
= 0.8 and change values of arrival rate of customer
2
/
2
.The
numerical results are summarized in Table 3.1. This shows that: (i) the length of queueing
system increases with the increasing of
2
/
2
. (ii) The length of the queue increases with the
increasing of
2
/
2
. (iii) The Waiting time in queueing system decreases with increasing of

2
/
2.
(iv) The Waiting time in queue first increases with increases of
2
/
2
and then
decreases
.
The following graph is showing the performance of the model at
1
/
1
= 0.8.
Graph between
2
/
2
and L , L
q
,W, W
q
19

2
/
2
3.6. CONCLUSION
In this chapter, we considered an M/M/1/k queueing system with non-preemptive priority.
We developed the equations of the steady state probabilities and derived the matrix form
solution of the steady-state probabilities. We also gave some performance measures of the
system. Some numerical examples were presented to demonstrate how the various
parameters of the model influence the behavior of the system as given above.
CHAPTER-4
PERFORMANCE ANALYSIS OF AN M/M/1/k QUEUE WITH
PREEMPTIVE PRIORITY:
4.1. ASSUMPTIONS
In this chapter, we consider an M/M/1/K queueing system with preemptive priority. The
assumptions of the system model are as follows:
1. The system at any instant of time is one of the two modes of operation.
2. All distribution are negative exponential, i.e., all transition rates are constants.
3. The probability of two or more random events occurring simultaneously is negligible.
4.2. NOTATIONS

i mode of system operation, i = 1, 2
N maximum number of customers in the system
n current number of customers in the system
i arrival rate in mode i
20
2
2
2
1
2
2 2
1
0 , 2
2
2
1
2
2
1
1 , 0
2
1
2
1
0
1
1
0 , 1
0 , 2

k
p
k
p
k
p
k
p
k p

1
2 , 0 0 , 2 1 , 0 0 , 1 0
+ + + + p p p p p
1
2
2
1
2
2
2
2
1
2
2
1
2
2
1
2
1
2
1
1
1
1

1
]
1

+ + + +

k
1 0 , 1 1 0 , 2
2 1 , 0 2 2 , 0
1 0 1 0 , 2 1 1 0 , 1
2 2 , 0 2 0 2 2 1 , 0
1 0 , 1 2 1 , 0 2 1 0
) (
) (
) (





p p
p p
p p p
p p p
p p p

+ +
+ +
+ +
i service rate in mode i
L mean number in the system
W mean waiting time in the system
Lq mean number in the queue
Wq mean waiting time in the queue
4.3. SYSTEM MODEL
We first consider the 2- priority M/M/1 queue with infinite buffers operating with a
preemptive priority discipline.
Since all the flows are known, the equilibrium probabilities of each of the states may be
found by writing and solving the corresponding balance equations with the normalisation
condition. As an example, consider the 2- priority M/M/1/2 queue with preemptive priority.
4.4. STEADY-STATE PROBABLITY
In this part of this chapter, we derive the steady-state probabilities by the Markov process
method. Let P(n) be the probability that there are n customers in the system. Applying the
Markov process theory, we obtain the following set of steady-state equations.


On solving these equation we get;



Using
We get

21
Graph between
2
/
2
and L , Lq , W, Wq
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 0.2 0.4 0.6 0.8 1

2
/
2
L(Length of the system)
Lq(Length of the queue)
W(Waiting time of system)
Wq(Waiting time of queue)
4.4.1. NUMERICAL RESULTS
In this subpart, we present some numerical examples to demonstrate how the various
parameters of the model influence the performance measures of the system. We fix the
maximum number of customers in the system N = 2.
Table 4.1.The case for
1
/
1
= 0.8

2
/
2
L L
q
W W
q
0.1000 0.8067 0.2543 3.9125 1.5596
0.2000 0.8656 0.2537 3.6087 1.4571
0.3000 0.9044 0.3667 3.4784 1.4101
0.4000 0.9331 0.2666 3.3325 0.9521
0.5000 0.9653 0.2789 3.2176 0.9521
0.6000 0.9998 0.2940 3.1240
0.9189
0.7000 1.0081 0.2113 2.9647 0.9155
0.8000 1.0721 0.3298 2.9582 0.9161
First, we select the
1
/
1
= 0.8 and change values of arrival rate of customer
2
/
2
. The
numerical results are summarized in Table 4.1. This shows that: (i) the length of queueing
system increases with the increasing of
2
/
2
. (ii) The length of the queue increases upto

2
/
2
from 0.1 to 0.3, it decreases at 0.4, again increases from 0.5 to 0.6, again decrease at
0.7 and then increases. (iii) The Waiting time in queueing system decreases with increasing
of
2
/
2.
(iv)The Waiting time in queue again decreases with the increasing of
2
/
2.
The
following graph is showing the performance of the model at
1
/
1
= 0.8.
4.5. CONCLUSION
In this chapter, we considered an M/M/1/k queueing system with priority. We developed
the equations of the steady state probabilities and derived the matrix form solution of the
steady-state probabilities. We also gave some performance measures of the system. Some
numerical examples were presented to demonstrate how the various parameters of the
model influence the behavior of the system as given above.
22
References
1. Robert E., Reneging phenomenon of single channel queues, Mathematics of
Operations Research 4(2) (1979), pp. 162-178.
2. Haight F. A., Queueing with balking, Biometrika 44 (1957), pp. 360-369.
3. Haight F. A., Queueing with reneging, Metrika 2 (1959), pp. 186-197.
4. Ancker C. J., Gafarian A., Some queueing problems with balking and reneging: I,
Operations Research 11 (1963), pp. 88-100.
5. Ancker C. J., Gafarian A., Some queueing problems with balking and reneging:
II, Operations Research 11 (1963), pp. 928-937.
6. Abou El-Ata M. O., Hariri A. M. A., The M/M/C/N queue with balking and
reneging, Computers and Operations Research 19 (1992), pp. 713-716.
7. Kleinrock L., Queueing Systems, Volume 1:Theory, Wiely, New York, 1975.
23
8. Heyman D. P., The push-out priority queue discipline, Operations Research,
vol.33, no. 2 (1985), pp. 397-403.
9. Takagi H., Queueing Analysis: A Foundation of Performance Evaluation, Vol.2,
Finite System, North-Holland, Amsterdam, 1991.
10. Gravey A., Boyer P. and Hebuterne G., Tagging versus strict rate enforcement in
ATM networks, Proceedings of IEEE GLOBECOM 1991, pp. 271-275.
11. Cheng X., Akyildiz I. F., A finite buffer two class queue with different scheduling
and push-out schemes, Proceedings of IEEE INFOCOM 1992, pp. 231-241.
12. Takagi H., Explicit delay distribution in first-come first-served M/M/m/K and
M/M/m/K/n queue and a mixed loss-delay system, Proceedings of Asia-Pacific
Symposium on Queueing Theory and Its Applications to Telecommunication
Networks, Korea, 2006.
13. Gupta N., Mishra G.D., Choubey A., Performance analysis of an queueing model
M/M/c/N with balking and reneging, International Journal of Computer
Mathematical Sciences and Applications, Vol 2, No. 4, October-December 2008,
pp. 335-339.
14. Gupta N., Mishra G.D., Choubey A., Performance analysis of An M/M/1/K
Queue with Non-Preemptive Priority, International Journal of Mathematical
Sciences and Engineering Applications, Vol 3, No. 2 (2009) pp. 191-197.
15. Gupta N., Mishra G.D., Choubey A., Performance analysis of queueing model
M/M/I/N with balking and reneging, International Journal of Pure Applied
Mathematical Sciences, Vol. LXX, No. 1-2, September 2009 pp. 59-65.
16. Gupta N., Mishra G.D., Choubey A., Performance Analysis of An M/M/1/K
Queue with Preemptive Priority, International journal of Business Research, Vol.
July, 2009 pp. 50-57.
Publications of Authors
The four Research papers of Authors have published in International Journals, which are as
follows:
1. Gupta N., Mishra G.D., Choubey A., Performance analysis of an queueing
model M/M/c/N with balking and reneging, International Journal of Computer
Mathematical Sciences and Applications, Vol 2, No. 4, October-December
2008, pp. 335-339.
2. Gupta N., Mishra G.D., Choubey A., Performance analysis of An M/M/1/K
Queue with Non-Preemptive Priority, International Journal of Mathematical
Sciences and Engineering Applications, Vol 3, No. 2 (2009), pp. 191-197.
24
3. Gupta N., Mishra G.D., Choubey A., Performance analysis of queueing
model M/M/I/N with balking and reneging, International Journal of Pure
Applied Mathematical Sciences, Vol. LXX, No. 1-2, September 2009, pp. 59-
65.
4. Gupta N., Mishra G.D., Choubey A., Performance Analysis of An M/M/1/K
Queue with Preemptive Priority, International journal of Business Research,
Vol. July, 2009, pp. 50-57.



25