UNIT-V

QUEUEING THOERY AND SIMULATION

Queue
Waiting line is called queue.
Eg: Queue id formed in ticket booking counter, bank, ATM, ration shop etc.
Characteristics of queueing system or elements of a queueing system(Structure of a Queueing
system)
A queueing system is specified completely by seven main elements.
1. Input or arrival(inter-arrival) distribution
2. Output or departure(service) distribution.
3. Service channels
4. Service discipline
5. Customer’s behavior
6. System capacity
7. Calling source or population

1). Input Pattern (or) Arrival Pattern

The input pattern represents the manner in which customer arrives the service and join
the queueing system. Customers arrive one at a time in the queueing system. The difference
between consecutive arrivals ins inter arrival time. The average number of customers arrived per
unit time is called arrival rate. Arrival rate is denoted by 𝜆. Usually, Poisson distribution is used
for arrival rate.
2) Service Pattern
The service Pattern is represents the manner in which customer served in the queueing
system. Service time is the time a customer spent at each service station. The average time a
customer is served per unit time is called service rate. Service rate is denoted by 𝜇. Service rate
is reciprocal of service time. Usually, exponential distribution is used for service rate.
3)Service channels
The queueing system may have a single service channel or more than one queue,
arranged in parallel or in series. The system may have 1 server or more than one server.
4) Queue Discipline
Some of the commonly used disciplines are:
First IN First Out (FIFO) Eg: Ration shop, Bank counters.
First In Last Out(FILO) Eg: CD stack, Godown
Server IN Random Order (SIRO) Eg: Lucky corner game.
Priority Queue Eg: In Voting booth, VIP’s are given first priority.
In hospitals, Emergency cases are treated first.
5) Customer Behaviour
a)Balking
Customer refusing to enter queueing system because the queue is too long
b) Jockeying
when there are parallel queues, the customer jumps from one queue to another
with shortest design to reduce the waiting time is said to be jockey.
c)Reneging
the customer who leaves the queue without receiving service because of too much
waiting, it is said to be reneging.
6) System Capacity
The maximum number of customers in the queueing system can be either finite or
infinite, in some queueing models, only limited customers or units are allowed in the system.
When the limiting value is reached, no further customers are allowed to enter the queueing
system.
STATE OF A SYSTEM
There are 3 states is queueing system
(i)Transient state
(ii)Steady state
(iii)Explosive state
Transient state:
A queueing system is said to be transient when tits characteristics arte depend on time.
Steady state:
A queueing system is said to be steady state when its characteristics are independent of
time.
Explosive state:
If the arrival rate of the customer is greater than the service rate, then the queue length
will go on increasing with time. This state is said to be explosive state.
KENDALL’S NOTATION FOR REPRESENTING QUEUEING MODEL
(OR)SYMMETRIC REPRESERNTATION:
Symbolic representation of a queueing model. A queue is generally represented by,
(𝑎 ∕ 𝑏 ∕ 𝑐 ): (𝑑 ⁄𝑒) is called Kendall’s notation.
𝑎 ⟶ type of distribution of the number of arrivals per unitr time.
𝑏 ⟶ type of distribution of service time.
𝑐 ⟶ number of servers
𝑑 ⟶capacity of system
𝑒 ⟶ queue discipline
In Kendall’s notation, a and b usually take one of the following symbols:
M: for Markovian or Exponential distribution
G: for arbitrary or general distribution
D: for fixed or deterministic distribution

The four important Queueing models described in Kendall’s notation are:

1) (𝑀 ∕ 𝑀 ∕ 1) ∶ (∞⁄𝐹𝐼𝐹𝑂)
2) (𝑀 ∕ 𝑀 ∕ 𝑠) ∶ (∞⁄𝐹𝐼𝐹𝑂)
3) (𝑀 ∕ 𝑀 ∕ 1) ∶ (𝑘⁄𝐹𝐼𝐹𝑂)
4) (𝑀 ∕ 𝑀 ∕ 𝑠) ∶ (𝑘⁄𝐹𝐼𝐹𝑂)

MODEL-1
( Single Server : Infinite capacity) (𝑴 ∕ 𝑴 ∕ 𝟏) ∶ (∞⁄𝑭𝑰𝑭𝑶)

(i)𝑠 = 1, 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 = ∞

Probability that there are n customers in the system

𝜆 𝑛 𝜆
𝑃𝑛 = ( ) [1 − ] , 𝑛 = 0,1,2,3, … ….
𝜇 𝜇
𝜆
Average number of (or) Expected number of customers in the system, 𝐿𝑠 =
𝜇−𝜆

𝜆2
Average number of (or) Expected number of customers in the queue, 𝐿𝑞 =
𝜇(𝜇−𝜆)
 1
Average waiting time (or) Expected time a customer spent in the system, 𝑊𝑠 =
𝜇−𝜆

Average waiting time (or) Expected time a customer spent in the queue,
𝜆
𝑊𝑞 =
𝜇(𝜇−𝜆)

Probability of waiting time of customer in the system exceeds 𝑘,

𝜆 𝑘+1
𝑃(𝑁 > 𝑘) = ( )
𝜇

Model: II : 𝑴 ∕ 𝑴 ∕ 𝒔 ∶ ∞⁄𝑭𝑪𝑭𝑺 (𝑭𝑰𝑭𝑶)

(Infinite capacity, Multiserver Queueing model)
𝜆
1. Traffic intensity 𝜌 =
𝑠𝜇
2. Average no.of customers in the system
𝜆
𝐿 𝑠 = 𝐿𝑞 +
𝜇
3. Average no.of customers in the queue
1 𝜆 𝑠+1 1
𝐿𝑞 = ( ) 𝑃
𝑠. 𝑠! 𝜇 (1 − 𝜌)2 0
𝐿
4. Average waiting time in the system, 𝑊𝑠 = 𝑠.
𝜆
𝐿𝑞
5. Average waiting time in the queue, 𝑊𝑞 =
𝜆
6. Prob. that there is no customer in the system

𝑠−1 −1
𝑛 𝑠
1 𝜆 1 𝜆 1
𝑃0 = [∑ ( ) + ( ) ]
𝑛! 𝜇 𝑠! 𝜇 1 − 𝜌
𝑛=0
1 𝜆 𝑠 1
7. Prob. that an arrival has to wait 𝑃[𝑁 ≥ 𝑠] = ( ) . 𝑃0
𝑠! 𝜇 1−𝜌
8. Prob. of a server busy = 𝜌
9. Prob. Of a server idle= 1 − 𝜌
 10. Average no. of customers (in non-empty queue) who have to actually
𝜌
wait.𝐿𝑤 = .
1−𝜌
1
11. Mean waiting time in the queue for those who actually wait .
𝑠𝜇−𝜆
12. Prob. that an arrival enters the service without waiting = 1 − 𝑃(𝑁 ≥ 𝑠).
13. Prob. that the waiting time of a customer in the system exceeds t.

𝜆 𝑠
𝜆
−𝜇𝑡(𝑠−1−𝜇)
( ) [1 − 𝑒 ] 𝑃0
𝜇
𝑃[𝑊𝑠 > 𝑡] = 𝑒 −𝜇𝑡 [1 + 𝜆 𝜆
]
𝑠! (1 − ) (𝑠 − 1 − )
𝜇𝑠 𝜇

Model: III : 𝑴⁄𝑴⁄𝟏 ∶ 𝒌⁄𝑭𝑪𝑭𝑺 (𝑭𝑰𝑭𝑶)

(Finite capacity, single server Queueing model)
𝜆
1. Traffic intensity 𝜌 =
𝜇
2. Prob that there are “n” customers in the system
(1 − 𝜌)𝜌 𝑛
𝑖𝑓 𝜆 ≠ 𝜇
1 − 𝜌 𝑘+1
𝑃𝑛 =
1
{ 𝑘+1 𝑖𝑓 𝜆 = 𝜇
3. Effective arrival rate 𝜆′ = 𝜇(1 − 𝑃0 )
1−𝜌
4. Idle time 𝑃0 = 𝑘+1
𝑖𝑓 𝜆 ≠ 𝜇
1−𝜌
5. Average no.of customers in the system
𝜌 (𝑘 + 1)𝜌 𝑘+1
𝐿𝑠 = − , 𝜆≠𝜇
1−𝜌 1 − 𝜌 𝑘+1
𝑘
𝐿𝑠 = , 𝜆=𝜇
2
6. Average no. of customers in the queue
𝜆′
𝐿𝑞 = 𝐿𝑠 − 𝑤ℎ𝑒𝑟𝑒 𝜆′ = 𝜇(1 − 𝑃0)
𝜇

𝐿𝑠
7. Average waiting time in the system𝑊𝑠 =
𝜆′
 𝐿𝑞
8. Average waiting time in the queue 𝑊𝑞 =
𝜆′
9. Prob. that the customer turned away 𝑃(𝑁 = 𝑘 ) = 𝜌 𝑘 ∗ 𝑃0

Model: IV : 𝑴⁄𝑴⁄𝒔 ∶ 𝒌⁄𝑭𝑪𝑭𝑺 (𝑭𝑰𝑭𝑶)

(Finite capacity, Multi server Queueing model)
𝜆
1. Traffic intensity 𝜌 =
𝑠𝜇
2. Effective arrival rate 𝜆 = 𝜇[𝑠 − ∑𝑠−1
′
𝑛=0(𝑠 − 𝑛) 𝑃𝑛 ]
3. Prob that there are “n” customers in the system
1 𝜆 𝑛
( ) 𝑃0 𝑓𝑜𝑟 𝑛 ≤ 𝑠
𝑛! 𝜇
𝑃𝑛 = 1 𝜆 𝑛
( ) 𝑃0 𝑓𝑜𝑟 𝑠 < 𝑛 ≤ 𝑘
𝑠! 𝑠 𝑛−𝑠 𝜇
{ 0 𝑓𝑜𝑟 𝑛 > 𝑘

−1
1 𝜆 𝑛 1 𝜆 𝑠 𝜆 𝑛−𝑠
4. 𝑃0 = [∑𝑠−1
𝑛=0 𝑛! (𝜇) + ( ) 𝑘
∑𝑛=𝑠 ( ) ]
𝑠! 𝜇 𝑠𝜇
5. Average queue length (or) Average no. of customers in the queue
𝜌 𝜆 𝑠
𝐿𝑞 = 2
( ) [1 − 𝜌 𝑘−𝑠 − (𝑘 − 𝑠)(1 − 𝜌)𝜌 𝑘−𝑠 ]𝑃0
𝑠! (1 − 𝜌) 𝜇

6. Average no.of customers in the system

𝜆′
𝐿 𝑠 = 𝐿𝑞 +
𝜇

𝐿𝑠
7. Average waiting time in the system𝑊𝑠 =
𝜆′
𝐿𝑞
8. Average waiting time in the queue 𝑊𝑞 =
𝜆′