International Journal of Pure and Applied Mathematics

�������������������������
Volume 53 No. 1 2009, 133-151

ON THE ANALYSIS OF M/M/1 FEEDBACK QUEUE WITH

CATASTROPHES USING CONTINUED FRACTIONS

V. Thangaraj1
�, S. Vanitha2
1,2Ramanujan Institute for Advanced Study in Mathematics

University of Madras

Chennai, 600 005, INDIA

1e-mail: thangarajv@unom.ac.in

2e-mail: vanithasudhakar05@gmail.com

Abstract: A transient solution is obtained analytically using continued fract

ions for the system size in an M/M/1 feedback queue with the possibility of
catastrophes at the service counter. Asymptotic behavior of the probability
of the server being idle and mean system size are obtained. The steady state
probability of the system size is also presented. Some key performance meas
ures are computed. Busy period and its moments are also obtained. Further
numerical illustrations are used to discuss the system performance measures
using MAPLE software.

AMS Subject Classification: 26A33

Key Words: M/M/1 feedback queue with catastrophes, transient analys
is, steady state analysis, busy period analysis, mean system size, throughput,
continued fraction, MAPLE software

1. Introduction
In the study of queueing systems, determination of transient solution is very
much essential to analyze the behavior of the system. Transient analysis is
very useful for all queueing models to obtain optimal solutions which pave
way to control the system. Among several methods, continued fraction is one
of the techniques that is used to obtain transient solution. Even in the case
of an simple M/M/1 queue, analytical approach to obtain transient behavior

Received: April 16, 2009 2009 Academic Publications

�Correspondence author
V. Thangaraj, S. Vanitha
is very difficult. In this regard, we have obtained the transient solution for
an M/M/1 feedback queue which is subject to catastrophes by employing an
effective continued fraction technique which is very simple.

Inparticular,queues withfeedback occurinproduction systems subjectto

rework, computer networks, telecommunication systems, supermarkets, banki
ng industries, hospital management, etc. Takacs [33], in his interesting pap
er has introduced a queue with feedback. Several authors have investigated
queueingsystems subject tofeedback. To mention afew,Disney,McNickle and
Simon[11] have studied several randomprocessesthat occurinM/G/1queues
withinstantaneousBernoullifeedback. D�Avignon andDisney[10] have also
considered the same queue with state dependent feedback mechanism.

Computer networks witha virus maybe modeledbyqueueing networks with

catastrophes. Studies have been made on stochastic models for the growth
of population subject to catastrophes. These include works by Hanson and
Tuckwell [15], Pakes et al [23], Murthy [22], Trajstman [34], Brockwell et al
[3],[4],[5] and[6]. Inrecentyears,queueing systemswith catastropheshave
been studiedbyBoucherie andBoxma[2],Jain andSigman[16],Dudin and
Nishimura[12],Gelenbe andPujolle[14],Chao et al[8],Artalejo[1],Krishna
Kumar andArivudainambi[30],KrishnaKumar andPavaiMadheswari[31],
KrishnaKumar,Krishnamoorthy,PavaiMadheswari andSadiq Basha[32].

The transient analysis of queuing system demands methodologically simp

le and easily numerically implementable approach. In this content, we prop
ose simple and mathematically elegant method of obtaining transient solut
iontoM/M/1feedbackqueuewith catastrophes using continuedfractionsand
also illustrate the numerical adaptability. A systematic study of the theory
of continued fractions can be found in Jones and Thron [17]. Its applicat
ion to the study of birth and death processes was initiated by Murphy and
O�Donohoe[21]. Pearce[29],Conolly andLangaris[9],Flajolet andGuillemin
[13],Parthasarathy andLenin[24],[25],[26] and[28],Parthasarathy andSelv
araju[27] andKrishnakumar et al[32] have applied continuedfraction techn
ique to study the transient behavior of the stochastic systems.

The rest of the paper is organized as follows: In the next section, we des
cribe the mathematical model andpresent thedetailed analysis of the transient
solution using continued fraction methodology. Section 3 provides correspondi
ng steady state results and certain performance measures are given in Section

4. Section 5 investigates the busy period of our model. Finally, we give some
numerical examples to discuss the system performance measures.
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

2. The M/M/1 Feedback Queue with Catastrophes

ConsideranM/M/1queuewithinstantaneousBernoullifeedback whichissubj
ectto catastrophes. Customersarriveattheservicestation oneby oneaccordi
ng to a Poisson stream with arrival rate #(> 0). There is a single server which
provides service to all the arriving customers. Service times are independently
andidentically distributed exponential random variable withparameter �. Aft
er the completion of each service, the customer can eitherjoin at the end of the
queue with probability p or he can leave the system with probability q where
p+ q = 1. The customer both newly arrived and those that are fed back are
servedinthe orderin which theyjointhetail of the originalqueue. Wedo not
distinguish between the regular arrival and feedback arrival. The customers
are served according to the first come, first served rule. Apart from arrival
and service processes, the catastrophes also occur at the service facilities as a
Poisson process with rate . Whenever a catastrophe occurs at the system, all
the available customers are destroyed immediately, the server get inactivated
momentarily and theserveris ready forservice only whenanew arrival occurs.

Now we obtain a system of difference differential equations satisfied by

the M/M/1 feedback queue which is subject to catastrophes and obtain the
transient solution of the queueing system.

Theorem 1. The system of Chapman-Kolmogorov forward differential

difference equations to describe M/M/1 feedback queue with catastrophes is
given by (2.1) and (2.2) with initial condition (2.3) and the transient solution
is given by

#.
n
.
t p

# 2
(2 #�q(t -u))

-(#+�q+)(t-u)
In

Pn(t)= nP0(u)e du, n =1,2,��� ,

�q (t -u)

where

(n +1) In+1(2 #�q t)

-(#+�q+)t

P0(t)= n+1
(�q)n e
(#�q) 2
t

n=0
t p

1
.

(n +1) In+1(2 #�q u)

-(#+�q+)u

+ n+1
(�q)n e du.

(#�q) 2

n=0

Proof. We model this queueing system as a continuous time Markov chain

(CTMC).Let{X(t): t 2 R+}be the number ofcustomersin the system at time

t. Let Pn(t)= P(X(t)= n), n =0,1,... be the state probabilities that there
V. Thangaraj, S. Vanitha
are n customers in the system at time t. Based on the above assumptions and
bytheChapman-Kolmogorovforwarddifferential-difference equations, the state
probabilities Pn(t),n =0,1,... forthequeueing system underinvestigation are

P0'
(t)= -#P0(t)+�qP1(t)+(1-P0(t)), (2.1)

P
'
(t)= #Pn-1(t)-(#+ �q+ )Pn(t)+�qPn+1(t),n =1,2,...,(2.2)

where #,�,q and are described above.

Withoutloss ofgenerality, assumethatinitially thereis no customerinthe

system, i.e.

P0(0)=1. (2.3)

In the sequel, for any function f(�), letf#(z)denote its Laplace transform.
By taking Laplace transform, the above system of equations is transformed to
the following system of simultaneous equations.

##

(z + #+ )P0
(z) =1+ + �qP 1
(z), (2.4)

###

(z + #+ �q+ )P (z)= #P n-1(z)+�qP n =1,2,... . (2.5)

nn+1(z),

After some algebraic adjustments, equations (2.4) and (2.5) respectively

reduce to

1+

# z
P0
(z)= , (2.6)

�qP1
(z)

(z + #+ )-
#

P0
(z)

Pn(z) #

= ,n =1,2,... . (2.7)

Pn-1(z) �qP

n+1(z)

(z + #+ �q+)-

P(z)

Now using(2.7)iterativelyin(2.6), we expressP0

#(z)as a continuedfraction

1+

# z
P0
(z)= . (2.8)

#�q

(z + #+)-

#�q

(z + #+ �q+ )-

(z + #+ �q+ )-���

The above equation is written as

1+

# z
P0
(z)= , (2.9)

(z + #+ )-#(z)
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

where

#�q

#(z) = (2.10)

#�q

(z + #+ �q+)-

(z + #+ �q+ )-���
#�q

= .

(z + #+ �q+)-#(z)
It is clear that #(z)satisfies the quadratic equation

#2(z)-(z + #+ �q+ )#(z)+#�q =0

the roots of which are

! � !2
-4#�q

#(z),
(z)= ,

2
where ! = z +#+�q+. It is seen that
(z)is a uniquereal root within[0,1)
and 0 # z< 1. Hence we consider only
(z)for further discussion.

Substituting
(z)for #(z)in(2.9), weget after some algebraic calculations

1+

# z
P0
(z)=
".
.
(2.11)
! - !2
-4#�q

(z + #+ )-

! - !2
-4#�q

1+

z 2#�q
=
".
.
(2.12)

! - !2
-4#�q

1-�q

2#�q

p!n+1

! - !2
-4#�q

=(�q)n

2#�q

n=0

p!n+1

! - !2
-4#�q
+(�q)n . (2.13)
z 2#�q

n=0
Oninversion,(2.13) yieldsthefollowing explicit expressionfor P0(t).

(n +1) -(#+�q+)t In+1(2 #�q t)

P0(t)= n+1
(�q)n e
n=0
(#�q) 2
t

t p

1
.

(n +1) In+1(2 #�q u)

-(#+�q+)u

+ n+1
(�q)n e du, (2.14)
u

(#�q) 2

n=0

0
V. Thangaraj, S. Vanitha
where In(�)is the modified Bessel function of the first kind of order n.

Now, we obtain the remainingtransientprobabilitiesinterms ofP0(t). From

equation(2.7), for n =1,2,��� we get the continued fraction as

Pn(z)

Pn-1(z)

= . (2.15)

#�q

(z + #+ �q+ )-

#�q

(z + #+ �q+ )-

(z + #+ �q+ )-���

By similar arguments, as before, the above equation is written as,

P (z) #

=
#.
#
,for n =1,2,..., (2.16)

Pn-1(z) ! + !2
-4#�q
2

which yields after some calculations,

.
n

! - !2
-4#�q

#
P (z)= P0
# (z), for n =1,2,... (2.17)

2�q

where ! = z + #+ �q+ .

On inversion, we get, for n =1,2,...

#.
n
.
p

# (2 #�q(t -u))

-(#+�q+)(t-u)
In

Pn(t)= nP0(u)e du. (2.18)

�q (t -u)

Thusequations(2.14) and(2.18) completelydetermineall thestatedependent

probabilities Pn(t) n =0,1,2,... of the system size which completestheproof
of the theorem.

Remark. When = 0 and q =1, then equation(2.18) coincides exactly

withKrishnaKumar et al[32] by taking # =0 in that paper.

Theorem 2. If > 0, then the asymptotic behavior of the probability of

the server being idle is given by

".
#k

1(#+ �q+ )- (#+ �q+ )2

-4#�q

P0(t)!

(#+ )k+1
2
k=0

as t !1. (2.19)
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

Proof. We have,

1+

# z
P0
(z)= , (2.20)

#�q

(z + #+ )- #.
.
! - !2
-4#�q

! -

where ! = z + #+ �q+ .

Now taking the limits as z ! 0, the above equation reduces to

.

lim zP0(z)= . (2.21)

z!0
2#�q

(#+ )- .
(#+ �q+ )+ (#+ �q+ )2
-4#�q
Multiplying numerator and denominator of the right hand side of above equat
ion by

(#+ �q+ )- (#+ �q+ )2

-4#�q
4#�q
and after some simple algebraic simplification, equation (2.21)reduces to

".
#k

1(#+ �q+ )- (#+ �q+ )2

-4#�q

P0
(z)# ,
z (#+ )k+1
2

k=0

as z ! 0. (2.22)

By using theTauberianTheorem,the result (2.19) followsfrom(2.22).

Theorem 3. If > 0, then, the asymptotic behavior of the mean system

size m(t)is given by

(#-�q)2�q

m(t)! +
h.

2(#+ )- (#+ �q+ )- (#+ �q+ )2

-4#�q

as, t !1. (2.23)

Proof. Let us consider equations (2.1) and (2.2) with initial condition
(2.3).

Let P(z,t)= Pn(t)z. From the above assumptions, it is clear that the
n=0

probabilitygenerating function P(z,t)satisfiesthepartialdifferential equation

@P(z,t) �q 1

= #z + -(#+ �q+ ) P(z,t)+�q 1- P0(t)+. (2.24)

@t z z
V. Thangaraj, S. Vanitha
Figure 1: The graph between #0
and #.
G1: � =5, =3,q =0.5 G2: � =5, =3,q =0.6
G3: � =8, =3,q =0.5 G4: � =5, =4,q =0.5

.
@P(z,t)
The mean system size is m(t)= nPn(t)= |z=1.

@z

n=1

Differentiating(2.24) with respectto z at z =1, we get

dm(t)

+ m(t)= #-�q[1-P0(t)]. (2.25)

dt

Solving the above differential equation for m(t) with m(0) = nPn(t) =0,
n=1

we obtain,

# -t)-
�q -(t-u)

m(t)=(1-e (1-e -t)+�q P0(u)e du. (2.26)

If m #(z)is the Laplace transform of m(t), then from above equation we get
# �q �q

m # (z)= - + P0
(z). (2.27)

z(z + ) z(z + ) z +
If > 0,thenfrom(2.27) and(2.11) we obtainthe required result(2.23).
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

Figure 2: The graph between #0

and .
G1: # =4,� =5,q =0.5 G2: # =4,� =5,q =0.4
G3: # =6,� =5,q =0.5 G4: # =4,� =6,q =0.5

3. Steady State Analysis

In this section, we shall derive the steady state probability distribution for our
queueing model.

Theorem 4. For > 0, the steady state distribution {#n;n # 0}of the
M/M/1 feedback queue with catastrophe corresponds to

#0
= (1-#), (3.1)
#n = (1-#)#n n =1,2,... (3.2)

where

(#+ �q+ )- #2
+ �q2
+ 2
+2# +2�q -2#�q

# = . (3.3)

2�q
V. Thangaraj, S. Vanitha
Figure 3: The graph between E(X)and .
G1: # =4,� =5,q =0.8 G2: # =4,� =5,q =0.6
G3: # =3,� =5,q =0.8 G4: # =4,� =6,q =0.8

Proof.

1+

# z
P0
(z)=
".
#
, (3.4)
! - !2
-4#�q

(z + #+)-

2
where ! = z + #+ �q+ .
Now multiplying(3.4) by z on both sides and taking the limit as z ! 0 we
get,

lim zP 0
(z)=#0
=
".
#
.

z!0

(#+ �q+ )- (#+ �q+ )2

-4#�q

( + #)-

2
After some algebraic manipulation, the above expression reduces to

-(#+ )+�q+(#+ �q+)2

-4#�q

#0
= ,

2�q
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

Figure 4: The graph between E(X)and #.

G1: � =5, =3,q =0.8 G2: � =5, =3,q =0.4
G3: � =6, =3,q =0.5 G4: � =5, =3,q =0.5

Figure 5: The graph between U and �.

G1: # =5, =3,q =0.5 G2: # =5, =3,q =0.4
G3: # =6, =3,q =0.4 G4: # =5, =2,q =0.4
V. Thangaraj, S. Vanitha
Figure 6: The graph between U and .
G1: # =4,� =5,q =0.5 G2: # =5,� =5,q =0.6
G3: # =7,� =5,q =0.5 G4: # =4,� =4,q =0.5

which simplifies to

(#+ �q+ )- #2
+ �q2
+2
+2# +2�q -2#�q

#0
=1- .

2�q

Now for n # 1, multiplying(2.17) by z on both sides and taking the limit

as z ! 0, we obtain,

.
!n

! - !2
-4#�q

##

lim zP (z)= lim zP 0

(z),

z!0
2�q z!0

#n = lim zP (z)

z!0

".
#n

(#+ �q+ )- #2
+ �q2
+ 2
+2# +2�q -2#�q
=#0
.

2�q

Thus equations(3.1)-(3.3) provide the steady-state distributionfor the syst

em size. Obviously, the steady statedistribution existsif and onlyif #< 1.

Remark. When = 0, # = �q < 1, which is the well known steady state

conditionfor an M/M/1feedbackqueue(see[33]).
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

Figure 7: The graph between U and #.

G1: � =6, =2,q =0.5 G2: � =6, =2,q =0.6
G3: � =2, =2,q =0.5 G4: � =6, =5,q =0.5

4. Moments under Steady State

Inthis section, we obtain the moments connected withthe steady state obtained
earlierinthispaper. Thesteady state moments are oftengood approximations
to the transient counterpart expressions even when time t is of moderate size.

Theorem 5. For > 0, the steady state probability generating function

#(s)is given by
1-#

#(z)= . (4.1)

1-#z
The mean, the variance of the system size namely E(Q),V ar(Q)and the mean
of the queue size E(Qq)are obtained as

E(Q)= , (4.2)

1-#
#

Var(Q)= , (4.3)

(1-#)2

#2
and E(Qq)= , (4.4)

(1-#)
V. Thangaraj, S. Vanitha
where # is given by equation (3.3).

Proof. Multiplying equations(3.1) and(3.2) by sand summing over n,

after simplesimplification, wederive(4.1). Theresults(4.2) and(4.3) follow
directly from(4.1) ondifferentiation with respectto s and setting s =1.

Similarly,thehigherfactorial moments canbe obtainedfrom(4.1) by succ

essive differentiation with respect to s and evaluating at s =1.

Now E(Qq)= .
(n -1)#n = #2

1-#.

n=1

Remark. Themean numberof customersinthesystem E(X)includes the

down time when there are no customers in the system.

P(Server is busy) = #n = #, (4.5)

n=1

andP(Server is idle) = #0
=(1-#), (4.6)

where # is asgivenin(3.3).

Theorem 6. For > 0, under steady state, the system throughput, U is

given as

U = �q #, (4.7)

where # is given by equation (3.3).

Proof. The system throughput, U, is the rate at which customers exit the
queue whenever there are one or more customers in the system. With the exit
rate �q, we obtain

U =[1-#0]�q = �q #, (4.8)

where wehave used equation(3.1).

Remark. The number of customers that were ejected per unit time is

#
n#n = E(X)= . (4.9)

1-#

n=1

Therefore, the fraction of customers who entered the queue and were ejected
when the server failed due to catastrophe is

E(X) #

= . (4.10)

##(1-#)

Theprobability that anarriving customerwill completeservice, conditioned on

the arrival during the working state of the server is given by

U �q

= #. (4.11)

##
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

5. Busy Period Analysis

A busy period is defined as the interval of time commencing at the instant 0
when a customer arrives at an empty counter and terminating at the instant
when the server becomes free for the first time. Let the length of the interval
whichis a random variablebe T and {N#(t)}be the stochasticprocessdenoting
the number of customers present at the instant t during the busy period.

We have {N#(0)=1}and the duration of the busy period is the first t for
which {N#(t)=0}. Let qn(t)= P{N#(t)= n|N#(0)=1},n =0,1,... be the
zero-avoiding state probabilities. Then q1(0)=1,qn(0)=0,n =1,2,....

Now qn(t)will satisfy the following difference differential equations.

q1'
(t)= -(#+ �q+ )q1(t)+�q q2(t), (5.1)

q
'
(t)= #qn-1(t)-(#+ �q+ )qn(t)+�q qn+1(t),n =2,3,.... (5.2)

We solve the above equations using continued fractions methodology. By

similar arguments as in Section 2, the solution for the above equations is given
by

#n/2

#n p

-(#+ �q+ )t

qn(t)= eIn(2t #�q),n =1,2,....

�q #t

Conditioning on the number of customers present at instant t, all of which

completetheir servicesin(t,t + dt), we have

b(t)dt ' P{t # T<t+ dt}

= P{t # T<t+ dt|N # (t)= j}P{N # (t)= j}

j=1

= P{t # T<t+ dt|N # (t)=1}P{N # (t)=1}

+ P{t # T<t+dt|N # (t)= j}P{N # (t)= j}.

j=2
The firsttermimpliesthatthereisonly onecustomer(attheinstant t)whose
serviceis completedbetween(t,t+dt), theprobabilityofthis eventbeing �qdt+
o(dt). The second term implies service completion of two or more customers in
(t,t + dt) and the probability of this event is o(dt). Thus taking the limit as
dt ! 0,

##1/2

# 1

-(#+�q+)t

b(t) =[�qq1(t)]= �q eI1(2t #�q). (5.3)

�q #t
V. Thangaraj, S. Vanitha
The Laplace transform of T is given by

b # (z)= L{b(t)}= �qL{q1(t)}= �qq 1

(z)

(z + #+ �q+ )- (z + #+ �q+ )2
-4#�q

= .

2#

Moments of the Busy Period

E(T)= - b # (z)|

z =0

dz

1 #+ �q+ - (#+ �q+ )2

-4#�q

=
.
, (5.4)
2# (#+ �q+)2
-4#�q

d 2�q

E(T2)= - b # (z)|= , (5.5)

z =0 3

dz 2

[(#+ �q+ )2
-4#�q]

##2

8�q#2
-!1
(#+ �q+ )2
-!1

V(T)= 3
, (5.6)
4#2
[(#+�q+ )2
-4#�q]2
where !1
=(#+ �q+ )2
-4#�q.

6. Numerical Illustrations
In the Sections 3 and 4, we have obtained explicit expressions for #n,n =
0,1,��� , the steady state probabilities of the system size, E(X), the mean
number of customersin the system, U, the system throughput. In this section,
we present some numerical illustrations to show the effect of the catastrophe
parameter and the arrival rate # on #0
E(X)and U.

In Figures 1 and 2, we plot the behavior of the probability #0

of no cust
omers in the system as a function of # and respectively. Obviously, #0
decreases as the arrival rate # increases and increases with the catastrophe rate
.

Figure3highlightstheeffect of thecatastrophe rate on the mean number

of customers E(X)in the system. ItisfoundthatE(X)is adecreasingfunction
of . Figure 4 illustrates the behavior of E(X)as a function of the arrival rate
#. It is found that E(X)is an increasing function of #.
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

Figures 5, 6 and 7 show the nature of the throughput U. It is found that U

is an increasing function of the service rate � as shown in Figure 5. In Figure
6, it is seen that U is an decreasing function of catastrophe parameter . From
Figure 7, one observes that U increases as arrival rate # increases.

Acknowledgements

ThesecondauthorthanksUniversitygrantscommissionforits financial support

during the preparation of this paper.

References

[1] J.R.Artalejo,A versatileapproachforwork removalinqueueing networks,

J. Oper. Res., 126, No. 2(2000), 233-249.
[2] R.J.Boucherie,O.J.Boxma,TheworkloadintheM/G/1queuewith work
removal, Probab. Eng. Inform. Sci., 10 (1996), 261-277.
[3] P.J. Brockwell, J. Gani, S.I. Resnick, Birth, immigration and catastrophe
process, Adv. Appl. Prob., 14 (1982), 709-731.
[4] P.J. Brockwell, J. Gani, S.I. Resnick, Catastrophe processes with continuo
us state space, Austral. J. Statist., 25 (1983).
[5] P.J. Brockwell, The extinction time of a birth, death and catastrophe proc
ess and of a related diffusion model, Adv. Appl. Prob., 17 (1985), 42-52.
[6] P.J. Brockwell, The extinction time of a general birth, death process with
catastrophes, J. Appl. Prob., 23 (1986), 851-858.
[7] X. Chao, A queueing network model with catastrophes and product form
solution, O.R. Letters, 18 (1995), 75-79.
[8] X. Chao, M. Miyazawa, M. Pinedo, Queueing Networks, Customers, Sign
als and Product Form Solutions,JohnWiley andSons,NewYork(1999).
[9] B.W. Conolly, C. Langaris, Solution of the stochastic birth emigration
model using continued fractions, Math. Scientist., 17 (1992), 12-127.
[10] G.R. D�Avignon, R.L. Disney, Single server queues with state dependent
feedback, INFOR, 14 (1976), 71-85.
V. Thangaraj, S. Vanitha
[11] R.L. Disney, D.C. Mc. Nickle, B. Simon, The M/G/1 queue with instant
aneous bernoulli feedback, Nav. Res. Log. Quat., 27 (1980), 633-644.
[12] A.N. Dudin, S. Nishimura, A BMAP/SM/1 queueing system with Markov
ian arrival input of disasters, J. Appl. Prob., 36 (1999), 868-881.
[13] P. Flajolet, F. Guillemin, The formal theory of birth and death processes,
lattice path combinatorics, and continued fractions, Adv. Appl. Prob., 32
(2000), 750-778.
[14] P. Gelenbe, K. Pujolle, Introduction to Queueing Networks, 2-nd Edition,
JohnWiley andSons,NewYork,(1998).
[15] F.B. Hanson, H.C. Tuckwell, Logistic growth with random density indep
endent disasters, Theoret. Popn. Biol., 19 (1981), 1-18.
[16] G.Jain,K.Sigman,APollaczek-KhichineformulaforM/G/1queueswith
disasters, J. Appl. Prob., 33 (1996), 1191-1200.
[17] W.B. Jones, W.J. Thron, Continued Fractions: Analytic Theory and App
lications, Encyclopedia of Mathematics and its Applications, II, Addison
Wesley, Newyork,(1980).
[18] W. Lederman, G.E. Reuter, Spectral theory for the differential equations
of simple birth and death process, Phil. Trans. Roy. Soc., 246 (1956),
321-369.
[19] L. Lorentzen, H. Waadeland, Continued Fractions with Applications,
North-Holland, Amsterdam(1992).
[20] J. Medhi, Stochastic Processes,IIEdition,Wiley andSons(1994).
[21] J.A. Murphy, M.R. O�Donohoe, Properties of continued fractions with app
lications in Markov processes, J. Inst. Maths. Appl., 11 (1975), 345-354.
[22] D.N.P.Murthy,A modelforpopulationextinction, Appl. Math. Modelling,
5 (1981), 227-230.
[23] A.G. Pakes, A.C. Trajstman, P.J. Brockwell, A stochastic model for a
replicating population subjected to mass emigration due to population
pressure, Math. Biosci., 45 (1979), 137-157.
[24] P.R. Parthasarathy, R.B. Lenin, On the exact transient solution of finite
birth and death processes with specific quadratic rates, Math. Scientist.,
22 (1997), 92-105.
ON THE ANALYSIS OF M/M/1 FEEDBACKQUEUE WITH...

[25] P.R. Parthasarathy, R.B. Lenin, On the numerical solution of transient

probabilities of quadratic birth and death process, J. Diff. Eqns. Appl., 4
(1998), 365-379.
[26] P.R. Parthasarathy, R.B. Lenin, An inverse problem in birth and death
processes, Comput. Math. Appl., 38 (1999), 33-40.
[27] P.R. Parthasarathy, N. Selvaraju, Transient analysis of a queue where pot
ential customers are discouraged by queue length, Math. Probl. Eng., 7
(2001), 433-454.
[28] P.R.Parthasarathy,K.V.Vijayashree,R.B.Lenin,AnM/M/1driven fluid
queue, continuedfraction approach,QueueingSystems, 42(2002),189-199.
[29] C.E.M. Pearce, Extended continued fractions, recurrence relations and
two-dimensional Markov processes Adv. Appl. Prob., 21 (1989), 357-375.
[30] B. Krishna Kumar, D. Arivudainambi, Transient solution of an M/M/1
queue with catastrophes, Computers and Mathematics with Applications,
40 (2000), 1233-1240.
[31] B.KrishnaKumar,S.PavaiMadheswari,TransientbehavioroftheM/M/2
queue with catastrophes, Statistica, anno LXII, 1 (2002).
[32] B. Krishna Kumar, A. Krishnamoorthy, S. Pavai Madheswari, S. Sadiq
Basha, Transient analysis of a single server queue with catastrophes, failu
res and repairs, Queueing Systems, 56 (2007), 133-141.
[33] L. Takacs, A Single server queue with feedback, The Bell System Tech.
Journal, 42 (1963), 134-149.
[34] A.C. Trajstman, A bounded growth population subjected to emigrations
due to population pressure, J. Appl. Prob., 18 (1981), 571-582.
152