0 Introducilol

Waiting
for service is par of our daily life. We wait at hotels, we

aD at the railway reservation

at th counter, and line-up' for service
up
que

and
the waiting phenomenon is not an experience limitedto
th
banks
in
in
human
Da beings alone. Jobs wait to be processed on a machine. Aeroplanes
ore given permission to land at an airport. Cars wait
befor-
le in air
Ve cannot avoid waiting without incurring inordinate
traffic signals. We
thei
at
delay. In tact, we can hope to reduce the adverse impact
exense or
to acceptable limits.
of waiting
The study of queues deals with quantifying the phenomenon of
watting in lines using representative measures of performance, such as
average queue length, average waiting time in queue and average facility
utilisation.

4.1 MARKOVIAN MODELSs

4.1 () The queue discipline

It is a rule according to which the customers are selected for
service when a queue has been formed.

4.1 (i) Queueing Problem

(a) Probability distribution of queue length.

(b) Probability distribution of waiting time of customers.
The busy period distribution
 Probability and
Queueing Theon
4.2
models
Solutions for Queueing
4.1 (ii)
state equations governing
(a) To obtain the system of steady the queue,
for finding out the probabilitv diet.
(b) To solve these equations
of queue length.
function for waiting time distri
(c)To obtain probabilty density
distribution,
distribution.
(c) To find the busy period
()To derive formula for Ls, Lq, (L/Lq>0), Ws, Wo (W/W 1)

( To obtain the probability of arrival during the service time of anany

customer.

4.1 (iv) Kendall's Notation for Representing Queuing Models

D.G. Kendll (1953) and later A. Lee (1966) introduced useful
notation for queuing models. The complete notation can be expressed
as

(a/b/c): (dVe)
a Inter-arrival distribution

b Service-time distribution
C Number of channels (or) servers
d System capacity
Queue disipline
The following conventional codes are generally used to replace
the symbols a, b and e:

Symbols for a and b

M Markovian (Poisson) arrival or departure distribution
(or exponential interarrival or service time distribuuo
E Erlangian or gamma interarrival service
or
u
distribution with parameter k,
GI=general independent arrival distribution,
 general departure distribution,
determine interarrival or service times.
determine

jmtasjore
jmbos
jor

first come, first served (or) FIFO,

FCFS

LCFS
last come, first served,
SIRO0 = service in random order,
GD general service discipline.

four important queueing systems are

The
The

(M/M/1) (/FIFO)
i)(M/M/s) (/FIFO)
Git) (M/M/1): (k/FIFO)
(v)(M/M/s) (k/FIFO)
Note: Mean, average, expected are same meaning.

41 (Terminology :

arrival
1
1 Mean
Mean arrival rate
rac =

Mean arrival time

Mean service rate =

Mean service time

1 P traffic intensity/ utilisation factor / busy period.

2 Po P[No customer in the system / idle]

3 Pn Pn customers in the system], [P (N=n) = Pnl

4 Average no. of customers in the system

Average no. of customers in the queue/Average length

of the queue.
6. Ws
s Average waiting time of a customer in the system
1. Wa
q Average waiting time of a customer in the queue.
8 wLw Average no. of customers in the non-empty queuc
9
P[N>k P[the no. of customers in the system exceeds k]
 Model
Model II
(M/M/1) : (co/FIFOD)
= and p, = , ( <u) (M/M/1) : (k/FIFO)

1
2 Po1 - p 2. 1-P if A*
Po1F 1
Po+1 if l =u

BPP" Po 3 Pa p"Po, if A * j

Pn if i =u
4. 4 L= -k+1)p*+1
s 1p 1-pk*
I~pk+1

'

|W =

where

|1'= # (1 - Po) =a (1- Pn)

Lw 1-P
6. P (N> K) = pk*1

Vs (w) =
(u -

) e--4)w
|8.

9.
F ( >t) =e - ) t
10. 1
W
4.8
Probability and eing Theory
Model IV
Model II
(M/M/s) : (k/FIFO)
(M/M/s) : (o0/FIFO)
AnA, V n ( <sp)

S
S
2.
2. Po
Po
S -1

s(1-p) | |
3.
P 0sn<a05n <s Pon nss
1
Po Pn s!-s
Po
s! - s
s <n sk
n 2s
P 0, n>k
4. u)s*1
e d - e(1-p
SS! (-P (1-P
Ls Lg k-s)pk-s+1]

1-P Po
P [ > 0 ] = P[N2s
5.

s!(1-p)o
.
E (W/Ws >0 EWa
P[W>0]
1
=

s-1

where
PoPo P S-1

P[N2s + 1] =

s!(1p)
'=us26-n)Pa|
n=0

Probability that there will

be some one waiting

8. P
 4.9
iQUeing T h e o r y

Characteristics of infinite capacity, single

server
4 1a ) C h a

model I (M|M|): (o | FIFO),

poisson queue
when An
= and un= (A <). [A.U. N/D 2003, 2011]

Tet N denotes the number of customers in the queueing system

number of customers in the queue
and the n u m b e
is (N-1)
the

Mean arrival rate

Mean service rate

P traffic intensity= u
Po 1-p Pn = p " (1-p)

1. Average number of customers in the system (L:

L 2
n=0
nPn =2 nph (1-p), 0<p<1
n=0

(1-P) 2 np"= (1-p) lp +2p+3p°+.

n=0

=
(1-P)P[1 +2p + 3p +...]
- (1-P)P(1-p) =
1-P
2. Average number of customers in the queue (

i.e., queue length)

Ls- by using Little's formula.

Lq
3. average waiting time of a customer in the system (W):

. W = Ls by using Little's formula.

4. The average waiting time of a customer in the queue (W):

W Lg by using Little's formula.

5 Average number of customers in non-empty queues (Lw

Lw= Lq
P(n> 1)
E x a

Arrivalsals at a telephone booth are considered to be poisson with an

oe time of 12 min. between one arrival and the next. The length
aver.

nhone call is
assumed to be distributed exponentially with mean

ofa
4 min.

the average number of persons waiting in the system.

Find
[A.U N/D 2010]

at the booth will

What is the probability that a person arriving
(i)
have to wait in the queue ?
more than 10 min.
What is the probability that it will take him
(i) call ?
to wait for the phone and complete his
altogether
be in use.
Estimate the fraction of the day when the phone will
(iv)
booth, when
department will install a second
() The telephone for atleast
has to wait on the average
convinced that an arrival
flow of arrivals should
how much the
3 min for phone. By
a second
booth ?
increase in order to justify
time-
of the queue, that forms from
(vi) What is the average length
to time ? [A.U. N/D 2004

booth >
single server
Solution: Given: Telephone
> infinite capacity
Arrivals at a telephone booth
model.
(»/FIF0)
The given problem is (M/M/1):
(or) (M/M/1): (»/FCFS)

[ FIFO=FCFS]
= 12 minutes
Given Mean arrival time
minute
per
Mean arrival rate () =

12

= 4 minutes
Mean service time
4.18

per minute
Mean service rate (u) =

- /12 12)
number of persons waiting in the system
) Average

Ls 1-P
1-3
Probability that the person
arriving in the booth has
(i)
to wait in the queue

P[W>0]= 1-P[W=0]
= 1-P[no customer in the system

=1-Po 1-[1-P =p -}
A person takes more than
10 minutes to wait and
(ii)
complete this call

P[W>1 = e - l ) t

e"l10 6/3)= 0.1889

P[W> 10
=
=

= P [phone is busy]
(iv) P[phone in use]
= 1 -P [phone is idle]
= 1 -Po = 1-[1-P

(V) The second phone will be installed if E [Wa]>3

H - 3

is the required
arrival ralc
ie.) If 3 where Ar
 (i.e.,) If AR -AR)

AR+R
3
16

3
16
3
28

3 1 1
Hence the arrival rate should increase by 42 Per minute
28 12

4
(vi) The average length of the queue 1
4 12

)--15
Exampl

customers
A
eDermarket
Superma,
has a single cashier. During peak hours,
att a
arrive a rate of 20 per hour. The average number of customers
that can be processed by the cashier is 24 per hour. Calculate:
shatcan
that the cashier. is idle.
d The probability
number of customers in the Queueing system.
T h e average
time a customer spends in the system.
ii) The average

average number of customters in the queue.

iv) The
waiting for
time a customer spends in the queue
()The average
[A.U Trichy N/D 2010
service.

Solution:

Given Single cashier > single server

Customers infinite capacity

model.
. The given problem is
(M/M/1): (o/FIFO)
(or) (M/M/1): (c/FCFS)

(2) = 20 per hour

Mean arrival rate
24 per hour
(u)
=

Mean service rate

20
P 24
idle:
the cashier is
() The probability that
Po = 1-P
Po
= 1 - 24

1-

= 0.1667
 in the system:
(ii) The average number of customers

Ls 1 -PP

20/24
1-2424
20/24
4/24

)- 5
(ii) The average time a customer spends in the system :

Ws
-

-h hour =
15 minutes

iv) The average number of customers waiting in the queue

Lg Ls-P
Lq

s-

100 = 4.1667
24

(v) The average time a customer spends in the queue :

Wa
(100
24 n0ur

= 12.5 minutes
Example 4.1a(13)|

arrival rates of telephone calls at telephone booth are according

The
Poisson distribution with an average time of 12 minutes between
erival of two consecutive calls. The length of telephone call is assumned
arr

abe exponentially distributed with mean 4 minutes.

Determine the probability that the person arriving at the booth
will have to wait.

(i) Find the average queue length that is formed from time to time.

(ii) The telephone company will install the second booth when
convinced that an arrival would expect to wait atleast 5 minutes
for the phone. Find the increase in flows of arrivals which will
justify the second booth.

(iv) What is the probability that an arrival will have to wait for
more than 15 minutes before the phone is free?
[AU NID 2006, A.U. CBT ND 2011]

Solution

Given Telephone booth single server

Telephone calls infinite capacity
The given problem is (M/M/1): (co/FIFO) model.

(or) (M/M/1): («/FCFS)

Given Mean arrival time = 12 minutes

Mean arrival rate () 1

=

12 per.minute
Mean service time = 4 minutes

Mean service rate (u) =

per minute

P =V12
1/4
 The probability that person
arriving at the booth will have to
(i)
wait is given by
0.3333
P(N 1)
=
p-

The average queue length that is formed from time to time

(ii) is
given by

1
1/3
1/3
1/3
-(9-
La
(ii) The telephone company will install a second
booth is W>5, i.e.,

R 5
Wqu(4-AR)
where AR if the required arrival rate. Thus, we have

AR
T> 5
or Ag R|
1..,
16

AR36
Hence, the arrival rate should increase
by 1 8 Per minute
1
to jusity the second telephone booth.
(iv) The probability that an arrival will have to wait for more than
15 minutes is given by

e - 2 ) 15
= e2.= 0.0821
 4.40

MULTISERVER
QUEUES eueing Deon
4.1. (b)
infinite multiple server
capacity, rver
Characteristics of poissson
net

: (o/PIFO) model, when 2 quee

model II (M/M/s)

1. Values of Po and Pn: [We use C or

s
For the Poisson queue system, Pn 15 given by

Ag 1 2 An Po, n 2 1
Pn .1)

where
-+ 1M2 n 2)
There is a single server, n 4,n. But there are . serven
working independently of each other. If there be less than s eetoe
(ie.) if n <s, only *' of the 's servers will be busy and theustoth
omer,
idle and hence the mean service rate will be nu.
others

If n 2s, all the s servers will be busy and hence the mean sernig
rate= S.
n, 0 Sn <s
Hence
n = su, nzs . (3)

Using (3) in (1) and (2) we have

Pn 14.24.34... n Po, if s n <s

= 1 2 ) Po if 0 Sn <s

and P = P
a.2.. (-1)Hs4.su... (n -s+ 1)
- 1 ! '(Gyn-s++ Ko

S!shs
P Po if as n
 2P = 1
is given by
Now
Now
Po n=0

1 Po 1
+2
ns s!sn-s u

Po= 1
n 0 / 2 !
=s4s

1
n
+ 1 Po 1

1 Po= 1

n
S -1
1
. Po
S

customers in the queue or average queue

2 The average number of

length,
00

Lq 2 (n -

s) Pn
n=s

co
1 Po
n-s) s!ns

n=s

Put n-S =X
StX

Po
x=0 s!
 X=0
x=0

(/u) 3* Po
p(1-P)=
ss!(1-p)

The average number of customers

in the system by Little's formula
3.
is

4. The average time a custmer has to spend in the system by Little's

formula is

5. The average time a customer has to spend in the queue by Little's

formula

W
The probability that an arrival has to wait = the probability
6.
that there are s or more customers in the system.

(i.e.) P (W,> 0) P (N 2 Ss)

-
Po
n

Po
s! (1-

7. The probability that an arrival enters the service without waitingS

=
1-P (an arrival has to wait)

=1
Po
s!1 S
QUeUeny
4.43
The mean
waiting time in the
queue for those who
E (Wa) actually wait
E (WWs >0) =

P(W,> 0)
s (1-
S.s! (1- Po
S
Po
1
1
us (1-
S

9 The probability that there will be sameone

waiting
=
P (N 2 s+1)

P n =s+1 n
Pa-P s
(N =s)

Po -

s(1 s!

s! 1 Po
s!
S
The average number of customers (in non-empty queues), who
have to actually wait,
E
(N/N21) = E (N P (NG21)
S+1
! (1 LS
1
S .S P
1- S

1
S Po
 Example 4.1.b(4)

A supermarket has 2 girls attending to sales at the counters. If tha

service time for each customer is exponential with mean 4 min and
if people arrive in poisson fashion at the rate of 10 per hour,

What is the probability that a customer has to wait for service?

(ii) What is the expected % of idle time for each girl ?

(ii)
iii) Tf the customer has to wait in the queue, what is the expected
length of his waiting time ? [A.U N/D 2008, M/J 2012]
Solution

Given girls multiple server

people infinite capacity
The given problem is (M/M/s) : (o/FIFO) model.

Given s = 2,

Mean arrival rate () =

10 per hour

=
10 x per minute

perper minute

Mean service time = 4 minutes

Mean service rate (u) =

per minute

1/6)
-
(1/4) - ---« 0.67

#---a
1-p 1- LS
-1--o67
=
 -1

Po s! (1 (1)

Tak -no n (0.67)"= 1+0.67 = 1.67

(0.67)
Take
Takes(1-P)- 2! (0.67)
=
0.335

(1) Po =
(1.67 +0.355)= 0.5
P[a customer has to wait] = P[N2 2
()

W.K.T (A/Po
PNs] =- P

2 = (0.67
2 0.67 (0.5)
=
0.168

i ) The fraction of time when the girls are busy =

S

the fraction of time when the girls are idle = 1 - = 2/3

the expected % of idle time for each girl = x 100 = 67

1
(ii) E (W/W, > 0)=

= 3 min.

4 2-
6
 4 pumps. The service
times follow the
A petrol pump station has
with a mean of 6
min. and carsarrive f or
distribution
exponential
rate of 30 cars per hour,
service in a poisson process at the

arrival would have to wait in

What is the probability that
an
(i)
line?
in the system
the
Find average waiting time, average time spent
(ii)
and the average number of cars
in the system.
idle on an average ?
(ii) For what % of time would a pump be

A.U. CBT A/M 20111

[A.U Tvli M/J 2010, A.U Trichy N/D 2010,
[A.U Trichy M/J 2011]

Solution
Given Petrol pumps Multiple server
Cars Infinite capacity
The given problem is (M/M/s) : (oo/FIFO) model
Given Mean arrival rate ( ) = 30 per hour

Mean service time = 6 minutes

Mean service rate (4) per minute

= x60 per hour = 10 per hour

S= 4

303
10
30
Pu (4(10)
S (910
-

0.75
1-P 1-A 1-= 025

fo n=0 ! (1)
s!1
o - 1 (3 1+: =
13

/u =
13.5
Take s! (1-P) (41) (0.25)
Po [13+ 13.5] = [26.5] = 0.0377

Plan arrival has to wait] =

P[W> 01
S

= PIN2S]S (1-
Po
34
P(N2 4) (0.25) (O.0377)
=
0.509
(4!)
(Gi) (a) To find W
Wa L (2)
Lg 1 aus+1
Ss! (1-p? o . (3)

(3) Lq
1 (3) (0.0377) = 1.527
(4)(4!) (0.25)
(0.25)
(2) W =

(1.527) 0.0509 h =

=
0.0509 x 60 min =
3.05 min
(b) To find Ws

Ws =
W+ (3.05+6) per min =
9.05 min
(c) To find Ls

Ls=Lg + =
1.527 +3= 4.527 =
4.53 cars
Q1) The fraction of
time when the pumps are busy =
traffic intensity
S 4
the fraction of time
when the pumps are idle
the required % 25% =
=
1-
12. Define Little's formula. [A.U A/M 2004]
(or) For (M/M/1) : (oo/FIFO) model, write the Little's formula.
[A.U CBT M/J 2010] [A.U A/M 2003]
(or) Write down the Little's formulas that hold good for the
infinite capacity Poisson queue models.
(or) Write the relations among Ls , Lq , Ws and Wq
Solution Aliter
1
) W - -

i) Wa =

W- i) LgL,-P
(ii) Ls =
à Ws Gii) W =

iL
Giv) Lg 1 1
=

Wa