Vous êtes sur la page 1sur 28

Queuing Theory Queuing Theory

unit unit - - 99
ntroduction ntroduction
The The study study of of waiting waiting lines lines is is called called queuing queuing theory theory..
In In general general aa queue queue is is formed formed when when either either units units requiring requiring
service service commonly commonly referred referred to to as as customers customers ,, wait wait for for the the
service service facility, facility, stand stand idle idle && wait wait for for the the customer customer..
Some Some customers customers wait wait when when the the total total number number of of customers customers
requiring requiring service service exceeds exceeds the the number number of of service service facilities facilities..
Some Some service service facilities facilities stand stand idle idle when when the the total total number number of of
service service facilities facilities exceeds exceeds the the number number of of customers customers requiring requiring
service service..
nput source nput source
The input source is the source from which customer comes The input source is the source from which customer comes
& joins the queue. & joins the queue.
The input source of the customer to the queuing system The input source of the customer to the queuing system
can be finite or infinite. can be finite or infinite.
source
Queue
Server 1
Server 2
Server 3
.
.
.
Sever m
Departure
from the
system
Queuing Model
Point of
arrival in
the system
Arrival characteristics Arrival characteristics
Size of the arrival population Size of the arrival population
Pattern of arrival at the queuing system Pattern of arrival at the queuing system
Behavior of arrivals. Behavior of arrivals.
Size of the arrival population Size of the arrival population
Population Population sizes sizes are are considered considered to to be be either either unlimited unlimited i i..ee
infinite infinite or or limited limited i i..ee finite finite..
It It is is considered considered to to be be infinite infinite if if the the rate rate at at which which source source
generates generates the the customers customers is is not not affected affected by by the the number number of of
customers customers in in the the service service system system..
Alternatively Alternatively source source is is considered considered finite finite if if the the rate rate is is
affected affected by by the the number number of of customers customers in in the the system system..
Pattern Arrival At the system Pattern Arrival At the system
ustomers ustomers arrive arrive at at service service facility facility according according to to some some
known known schedule schedule or or else else they they arrive arrive randomly randomly..
In In queuing queuing theory theory the the number number of of arrivals arrivals per per unit unit of of time time
can can be be estimated estimated by by probability probability distribution distribution known known as as
Poisson Poisson distribution distribution..
The The major major feature feature of of this this distribution distribution is is that that number number of of
people people arriving arriving at at aa given given time time is is not not at at all all influenced influenced by by
the the number number of of previous previous arrivals arrivals..
Arrival Arrival rate rate is is given given by by symbol symbol lamda lamda
Behavior of the arrival Behavior of the arrival
A customer joining the service , does not get the immediate A customer joining the service , does not get the immediate
service , the customer may : service , the customer may :
1. 1. stay in the system until served stay in the system until served
2. 2. move from one waiting line to another move from one waiting line to another
3. 3. estimate the waiting time & decide whenever to leave. estimate the waiting time & decide whenever to leave.
The first type of customer who stay in the system are The first type of customer who stay in the system are
called as patient customers. called as patient customers.
The other two are classified as impatient or reneging The other two are classified as impatient or reneging
customer. customer.
Queuing Length Queuing Length
The length of a line can be either limited or unlimited. The length of a line can be either limited or unlimited.
In In aa sales sales department department where where the the customers customers orders orders are are
received received ,, there there is is no no restriction restriction on on the the number number of of orders orders
that that can can come come in in so so that that aa queue queue of of any any size size can can form form..
When When there there is is oo limit limit on on its its size size the the permissible permissible queue queue is is
said said to to be be infinite infinite..
When When there there is is aa limit limit on on size size of of queue queue it it is is said said to to be be finite finite..
Service Mechanism Service Mechanism
nce the customer has entered the system for service , the nce the customer has entered the system for service , the
service system has following characteristic: service system has following characteristic:
time taken for service time taken for service
service facility service facility
service discipline of the system service discipline of the system
Service time Service time
Service time may be either constant or random. Service time may be either constant or random.
If service time is constant , it takes the same amount of If service time is constant , it takes the same amount of
time to service each customer. time to service each customer.
The symbol describing the service rate is and the units The symbol describing the service rate is and the units
are in service per unit time. are in service per unit time.
Service facilities Service facilities
umber umber of of channels channels : : AA single single channel channel phase phase case case is is where where
the the arrivals arrivals pass pass through through only only one one station station in in order order to to
complete complete the the required required service service..
E E..g g.. toll toll booth booth ,, theatre theatre ticket ticket window window
AA multiphase multiphase channel channel involves involves the the situations situations where where the the
units units to to be be served served pass pass through through aa series series of of station station in in aa
sequence sequence..
Queue
Service facility
Queue
Service
facility
Service
facility
Single channel
Single phase
Single channel
Multiple phase
Queue
Service
facility
Service
facility
Queue
Service
facility
Service
facility
Service
facility
Service
facility
Multiple channel Single
phase
Multiple channel
multiple phase
Service discipline Service discipline
Queue discipline refers to the rule by which customers in Queue discipline refers to the rule by which customers in
the line are to receive service. the line are to receive service.
Most systems use queue discipline known as the first in , Most systems use queue discipline known as the first in ,
first out rule (FIF. first out rule (FIF.
Most queues operate on the principle of first Most queues operate on the principle of first --come come -- first first
served . served .
ther ways of dealing with view can be ther ways of dealing with view can be
last in , first out ( LIF last in , first out ( LIF
#andom service #andom service
Alphabetically organized. Alphabetically organized.
Application of Queuing system Application of Queuing system
ommercial service system ommercial service system
Business industrial system Business industrial system
Social service system Social service system
Transportation service system Transportation service system
Operating characteristics of Operating characteristics of
queuing system queuing system
&tilization factor for the system &tilization factor for the system
Average number of customer in the system Average number of customer in the system
Average number of customer in queue Average number of customer in queue
Average time each customer spends in the queue Average time each customer spends in the queue
Probability that service will be idle. Probability that service will be idle.
Probability of specific number of customers in the system Probability of specific number of customers in the system
Notations & Terminologies Notations & Terminologies
otation that has been used for summarizing the main
characteristics of parallel queue has universal standardized
format:
(a/b/c : (d/e/f
A : arrival distribution
B : service time distribution
number of parallel servers
D service discipline ( FFS , LIF , SI#
E maximum number of customers allowed in system
F size of calling source
Basic Notations Basic Notations
Arrival rate( Arrival rate(AA ) ) : The average rate at which customers arrive & : The average rate at which customers arrive &
join the queue is the arrival rate. join the queue is the arrival rate.
$ervice time ( $ervice time ( ): ): The average time required to service one The average time required to service one
customer is service time. The units for the service time should be customer is service time. The units for the service time should be
the inverse of the units for arrival rate. the inverse of the units for arrival rate.
:mber :mber of of c:8tomer c:8tomer in in the the 88tem 88tem (L (L
$ $
) ): : these these are are the the
average average number number of of customer customer in in the the system system both both waiting waiting && being being
serviced serviced..
Average Average waiting waiting time time in in the the 88tem(W 88tem(W
Q Q
: : It It is is the the average average
time time each each customer customer in in the the system system both both waiting waiting && being being serviced serviced..
Average Average waiting waiting time time in in the the q:e:e q:e:e (W (W
$ $
) ): : It It is is the the average average
time time each each customer customer waits waits in in the the queue queue..
Model : (M/M/1): (FFO/ Model : (M/M/1): (FFO/++ ++
This is a single channel queuing model. This is a single channel queuing model.
The arrival time has a poison distribution & arrival rate is denoted The arrival time has a poison distribution & arrival rate is denoted
by by
Arrivals are from infinite population Arrivals are from infinite population
The queue discipline is FFS ( i.e FIF the customers are served The queue discipline is FFS ( i.e FIF the customers are served
on first come first served basis. on first come first served basis.
There is only single service station There is only single service station
The mean arrival rate is less than the mean service rate The mean arrival rate is less than the mean service rate
The waiting space available for customers in the queue is infinite. The waiting space available for customers in the queue is infinite.
Model : (M/M/1): (SRO/ Model : (M/M/1): (SRO/++ ++
This This model model is is essentially essentially same same as as model model II except except that that
service service discipline discipline follows follows SI# SI# rule rule ( ( service service in in random random
order order
As As the the probability probability of of nn customer customer in in the the system system does does not not
depend depend upon upon any any queue queue discipline discipline it it may may be be concluded concluded that that
SI# SI# rule rule case case is is same same as as MDEL MDEL II
MODEL MODEL : (M/M/1): (FCFS/N/ : (M/M/1): (FCFS/N/+ +
This model is different from model I in respect to the This model is different from model I in respect to the
capacity of the system. capacity of the system.
Here the capacity of the system is limited to customers Here the capacity of the system is limited to customers
only. only.
Therefore difference equations derived in model as long as Therefore difference equations derived in model as long as
n < n <
Model V : (M /E Model V : (M /E
KK
/1) : /1) :
(FCFS / (FCFS /++ ++
The The queues queues in in which which arrivals arrivals && or or departures departures may may not not
follow follow poison poison distribution distribution are are called called as as non non poison poison
distribution distribution..
This This model model consists consists of of kk identical identical exponential exponential service service
phases phases in in series series each each with with mean mean service service time time of of 11/k /k ..
AA new new service service does does not not start start until until all all kk phases phases have have been been
completed completed..
AA service service station station in in which which customer customer passes passes through through kk
phases phases shall shall have have the the service service time time following following Erlang Erlang
distribution distribution..
Characteristics of Erland Characteristics of Erland
Distribution Distribution
The Erlang distribution is useful for defining the service The Erlang distribution is useful for defining the service
time of a customer who receives k identical repetitions of a time of a customer who receives k identical repetitions of a
particular service. particular service.
The probability function of Erlang distribution is given by The probability function of Erlang distribution is given by
(k (k
mm
(k (k- - 1 1
Where m is the expected number of customers completing Where m is the expected number of customers completing
service per unit of time service per unit of time
In case of constant distribution of customer interaarival In case of constant distribution of customer interaarival
times k tend to times k tend to ++
Limitation of Single channel Limitation of Single channel
Queuing Model Queuing Model
The The waiting waiting space space may may be be limited limited ..
The The arrival arrival rate rate is is dependent dependent .. Potential Potential customers customers are are
discouraged discouraged from from being being entering entering the the queue queue if if they they observe observe
along along line line at at the the time time they they arrive arrive..
Another Another limitation limitation that that arrival arrival process process is is not not stationary stationary
The The population population of of customers customers may may be be finite finite ..
Multi Channel Queuing Models Multi Channel Queuing Models
In In aa multiple multiple queuing queuing system system two two or or more more server server are are
available available to to handle handle the the customer customer who who arrive arrive for for service service..
By By introducing introducing number number of of servicing servicing units units the the length length of of the the
queue queue && the the waiting waiting time time are are reduced reduced..
Model v: (M/M/S) Model v: (M/M/S) : (FCFS / : (FCFS /++ ++
This a multichannel model in which parallel service stations This a multichannel model in which parallel service stations
are serving a single queue. are serving a single queue.
All service stations provide same service All service stations provide same service
ustomers are served on first come first basis. ustomers are served on first come first basis.
Arrival rate has poison distribution with mean rate as A Arrival rate has poison distribution with mean rate as A
Service stations have exponential service times with mean Service stations have exponential service times with mean
1/ customers entering in queue & served by all service 1/ customers entering in queue & served by all service
stations stations
Queueing Model in series Queueing Model in series
In In real real life life situations situations where where the the arriving arriving customer customer passes passes
through through successive successive number number of of service service channels channels in in order order to to
complete complete his his service service are are known known as as queues queues in in series series or or
Queues Queues in in Tandem Tandem..
Queues Queues in in series series with with possibility possibility of of queuing queuing..
Queues Queues in in series series without without possibility possibility of of Queueing Queueing

Vous aimerez peut-être aussi