Operations research waiting lines (queuing theory)

Queuing theory (waiting lines)

Introduction:
Queue is a common phenomenon in our every day life. Queues are formed at the
bus-stands, traffic lights, telephone booths, enquiry windows, reservation counters, bank
and post office counters, industries, schools, hospitals, cafeterias, petrol pumps, and
theaters etc.,
Queues are formed when the current demand for a given service exceeds the
capacity of the service facility to render service. Queues can also be formed even when
service rate is higher than the arrival rate due to random pattern of arrival of customers
(i.e. when suddenly more customers come at a time for service). One can shorten queues
or prevent queues by employing additional service facilities but that may not be always
desirable. The cost of additional service facilities reduces margin of profit while
excessive waiting time results in loss of sale / loss of customers. Queuing theory gives
mathematical treatment to the waiting line problems. It helps to strike an optimum
balance between the cost associated with waiting and cost associated with prevention of
waiting.

These are two basic approaches to solve a queuing problem, mathematical

approach and simulation approach.

1
Operations research waiting lines (queuing theory)

The mathematical approach makes use of probability distribution concept to

represent arrival and service rates. The simulation approach is an iterative method where
in, an experiments based on the random samples drawn from inter arrival and service
time distribution.
Terminology:
1. Customers: The persons or objects that require certain service are called
customers.
2. Server: The person or an object or a machine that provides certain defined
service is known as server.
3. Service: The activity between server and customer is called service that consumes
some time.
4. Queue or waiting line: A systematic or a disciplined arrangement of a group of
persons or objects that wait for a service is called queue or waiting line.
5. Arrival: The process of customers coming towards service facility or server to
receive service is arrival.
6. mean arrival rate: The average number of customers arriving per unit time is
called mean arrival rate and is denoted by ‘ ’
=total No. of customers arriving / total time taken.
7. Mean inter arrival time: It is the average time gap between two consecutive
arrivals of customers. It is the inverse of mean arrival rate i.e.

8. Service rate: It is the average number of customers served per unit time and is
denoted by ‘ ’.

9. Mean service time: It is the average time taken by the server to serve a customer

and is equal to inverse of service rate. i.e.

10. Transient state and steady state:

2
Operations research waiting lines (queuing theory)

The queuing system is in either a transient state or steady state. When a system is
started, it progresses through a series of changes. Eventually, it attains stability. In the
starting position the system is greatly influenced by the starting condition and by the
amount of elapsed time. This period of rapid transition termed as transient state. After
sufficient time has passed, the system becomes essentially independent of the starting
condition and elapsed time. This stable condition is termed as a steady state.
Elements of queuing system:
Basic elements of a queuing system are
1. Input process:
Customers arrive at a service counter for service. Two important characteristics of the
input process are its size and pattern of arrival. The size represents the number of
customers that arrive from time to time for service and pattern of arrival suggests the
distribution of the arrivals.
The size of the input process is generally assumed to be infinite as this assumption
facilities analysis. The arrival may either be at a constant rate or random in
accordance with some probability distribution. Where the arrivals are completely
random, they follow Poisson distribution with mean arrival rate equal to the average
number of arrivals per unit of time.
The following information is required to study the input process.
i) arrival distribution
ii) inter-arrival distribution
iii) Mean arrival rate. This is generally represented by a Greek letter ‘ ’
(Lambda).

iv) Mean time between consecutive arrivals. This is equal to ‘ ’.

2. waiting line: The following information is required to study the waiting line
a) Waiting time: Waiting time implies that a customer spends in the queue before
being taken up for service. It equals the time that elapses between the arrival of
the customer and the commencement of the service to that customer.
b) Service time: Service time implies the time spent by the service facility to render
service to a customer. It may be either constant or variable.

3
Operations research waiting lines (queuing theory)

c) Waiting time in the system: Waiting time in the system implies the time spent by
the customer in the queue system. It equals waiting time plus service time.
d) Queue length: Queue length implies the number of customers waiting in the
queue.
e) System length: System length implies the number of customers in the queue plus
those being serviced.
3. Service discipline: Service discipline is also called queue discipline, represents
the rule by which the next customer in the waiting line is selected for service.
i) First in first out (FIFO) or first come first serve (FCFS):
It is the discipline in which the customers are served in the chronological order of
the arrivals.
E.g.: Tickets at a cinema hall, sales at a grocery shop, trains on a single line
platform etc.
ii) Last in first out (LIFO) or last come first served(LCFS):
If the service is made in opposite order of arrivals of customers, i.e. whoever
comes last is served first and first obviously goes last; it is called LIFO or
LCFS.
E.g.: Stack of plates, loading and un-loading a truck or go-down, office filing of
papers in chronological orders, wearing socks and shoes, dressing a shirt and
coat over it, packing systems etc.
iii) Service in random order (SIRO):
By this rule, the customer for service is picked up at random, irrespective of
their arrivals.
E.g.: A telephone urgent call given to a customer charged at higher price, a
separate counter for cheques at a electricity bill payment counter, priority given
at (APSRTC) reservation counter who buys a CAT (Concessional annual ticket)
card etc.

iv) pre-emptive priority rule:

4
Operations research waiting lines (queuing theory)

Under this rule, highest priority is given to certain customers irrespective of

their arrival and costs. The customer is allowed enter into the service
immediately after entering into queue system, even if another (lower priority)
customer is already in service. i.e. the lower priority customer who is in service
will be interrupted (pre-empted) to facilitate the special customer.
E.g.: An emergency case arriving at a doctor’s clinic who is attending to a
regular out-patient. The doctor will stop his service to the regular patient and
immediately rushes to emergency case.
A minister of VIP coming to receive a service at counter is given highest
priority.
v) Non pre-emptive priority rule:
This is also a rule by priority to the special customer but the priorities will not
emptive the current service. The service to the special customer starts
immediately after the completion of current service.
E.g.: A medical representative will be given appointment immediately after the
current service to an out patient at a doctor’s clinic.
*UN satisfied customer: A customer who is not satisfied by the service by the
quantity is said to be un-satisfied.
*dissatisfied customer: A customer who is not satisfied by the quality of service is
said to be dissatisfied.
4. customer behaviour:
Customer behaviour represents the reactions of typical customers when
confronted with a queue before the service facility. Typical tendencies observed
among some customers are:
i) A new customer may not join the queue because of his / her reluctance to
wait. This trait referred to as balking.
ii) Only few of the new customers may join the queue and hence some of them
may demand service on their behalf as well as on behalf of some other customers.
This action of their is called collusion:

5
Operations research waiting lines (queuing theory)

iii) A customer may join the queue wait up to certain time and then leave the
queue system without getting service. This action of the customer is called
reneging.
iv) A customer may keep on switching from one queue to another. This happens
when there are more than one service counter. This tendency so observed is called
Jockeying:
5. Service facility:
Service facility represents servers, which render service. To analyze a service
facility, both number of servers and arrangement of servers need to considered.
The service facility can be arranged in any one of the following types:
a) single queue: single server
Under this arrangement there is one queue and one server in the service facility.

b) single queue: multiple servers ( series arrangement)

Under this arrangement, there is one queue but several servers in the service
facility. The servers are placed in a series arrangement.
Each server renders a specialized service and the customers to be serviced needs
to pass each of these servers before leaving the system.

6
Operations research waiting lines (queuing theory)

c) Single queue: multiple servers (parallel arrangement):

Under this arrangement, there is one queue but more servers in the service facility
are placed in a parallel arrangement. The customers wait in the queue until one of
the servers free to take them in for service.

d) Multiple queue: multiple servers:

Under this arrangement, there are more than one queue and more than one server
in the service facility. A customer on arrival can join any one of the queues and
may be required to pass either one server or more servers depending upon their
arrangement in the service facility.

7
Operations research waiting lines (queuing theory)

6. System output:
System out put refers to the rate at which customers are rendered service (i.e. rate
at which customers leave the queue system after service). System output is
dictated by the service time required by the facility to render service and
arrangement of service facility.
The Traffic Intensity:
The traffic intensity is the ratio of the average arrival rate to the average service rate.
That is

i) when there is just one server=

ii) when there are multiple servers (say C) =

Traffic intensity signifies the likelihood and the extent of queuing. For example,
traffic intensity ( ) of
i) Less than unity ( <1) implies that customer will be serviced faster than
their average arrival rate.
ii) Equal to unity ( =1) implies that customers will be arriving as fast as
they can be serviced and
iii) Greater than unity ( >1) implies that customers will be arriving faster
than they can be serviced and as such the queue will go on building up.

Queuing models: Kendall-Lee Notations:

8
Operations research waiting lines (queuing theory)

D.G. Kendall in 1953 has noticed three of them in the form (a / b / c) and later A.M. Lee
in 1966 added two more in the form (d / e) to describe a queue model. These are known
as Kendall-Lee notations in the standard format as (a / b / c): (d / e).

Classification of queuing systems:

Model I:(M/M/1):(
1. Mean Number of customers in the system:

L s= = --------------------------- (1)

Since for

There fore equation (1) can be written as

2. Mean number of customers in the queue (Mean queue length):

9
Operations research waiting lines (queuing theory)

Lq= = =

3. Mean length of non-empty queue:

(L: L>0) =

4. Mean waiting time that a customer spends in the system (waiting time + service
time):

Ws=

5. Mean waiting time that a customer spends in the queue:

Wq=

7. The probability that the number of customers in the queue being served is
greater than k is:

P(n>k)=

8. probability of waiting time in the queue : P(W>w):

9. Probability of waiting time in the system P (W>w):

10. To find the variance of queue length:

10
Operations research waiting lines (queuing theory)

11. To find probability of arrivals during the service time of any given customer:

= where r = number of arrivals during service time.

12. prob. [queue size

= 1-

Problems on model- I
1. Customers arrive at a booking office window, being manned by a single individual at
a rate of 25 per hour. Time required to serve a customer has exponential distribution
with a mean of 120 seconds. Find the mean waiting time of a customer in the queue.
Solution:
We have = 25 customers per hour

Customers / hour

And

11
Operations research waiting lines (queuing theory)

Therefore, mean waiting time of a customer in the queue:

Hours or 10 minutes

2. A repair shop attended by a single machine has an average of four customers an hour
who bring small appliances for repair. The mechanic inspects them for defects and
quite often can fix them right away or otherwise render a diagnosis. This takes him
six minutes on the average. Arrivals are Poisson and service time has the exponential
distribution. You are required to
a) Find the proportion of time during which the shop is empty
b) Find the probability of finding at least 1 customer in the shop
c) What is the average (mean) number of customers in the system?
d) Find the average (mean) time spent, including service.
Solution:
=4 customers per hour

Customers per hour

The traffic intensity is:

a) proportion of time during which the shop is empty:

P0=
b) Probability of at least 1 customer in the shop: P(W>0) = 1-P0=1-0.6=0.4

c) average (mean) number of customers in the system: Ls=

d) The average (mean) time spent, including service: W s= or 10

minutes.
3. The belt snapping for conveyors in an open cast mine occur at the rate of 2 per shift.
There is only one hot plate available for vulcanizing; and it can vulcanize on an
average 5 belts snap per shift.
a) What is the probability that when a belt snaps, the hot plate is readily available?
b) What is the average number of belts in the system?
c) What is the waiting time of arrival?

12
Operations research waiting lines (queuing theory)

d) What is the average waiting time plus vulcanizing time?

Solution:
=2 belts per shift
= 5 belts per shift
and therefore
a) fraction of time the system is empty:

P0=

b) average number of belts in the system:

L s=

c) waiting time in the queue for an arrival:

Shift.

d) Average waiting time plus vulcanizing time.

W s= shift

4. Arrivals of machinists at a tool crib are considered to be Poisson distributed at an

average rate of 6 per hour. The service time at the tool crib is exponentially
distributed with an average of 3 minutes.
a) What is the probability that a machinist arriving at the tool crib will have to
wait?
b) What is the average number of machinists at the tool crib?
c) The company will install a second tool crib when convinced that a machinist
would have to wait at least six minutes before being served. By how much the
flow of machinists to the tool crib must increase, to justify the addition of a
second tool crib?
Solution:
=6 per hour

= machinist per minute

13
Operations research waiting lines (queuing theory)

= machinist per minute

and
a) probability that a machinist arriving at the tool crib will have to wait (i.e. fraction
of time the server is busy) is given by 1-P0=1-

b) Ls= machinists

c) The installation of second tool crib will be justified if the customer (machinist) has
to wait at least six minutes before being served.
Since Wq= 6, =0.33 and = (say) for second tool crib, therefore

This yields =0.21. Hence the arrival rate should become 0.21 machinists per minute to
justify the second booth.
5. Customers arrive at a one-window drive-in bank according to a Poisson distribution
with mean 10 per hour. Service time per customer is exponential with mean 5
minutes. The space in front of the window, including that for the serviced car can
accommodate a maximum of 3 cars. Other cars can wait outside this space.
a) What is the probability that an arriving customer can drive directly to the
space in front of the window?
b) What is the probability that an arriving customer will have to wait outside the
indicated space?
c) How long is an arriving customer expected to wait before starting service?
Solution:
= 10 per hour, = 60 /5 = 12 per hour
a) The probability that an arriving customer can drive directly to the space in front of
the window:
= P0+P1+P2

14
Operations research waiting lines (queuing theory)

b) the probability that an arriving customer has to wait out side the indicated space =
probability that there are at least 3 customers in the space in the front of the
window
=1-(P0+P1+P2) =1-0.42-P3

=0.58-

c) average waiting time of a customer in queue

Hours

6. A repair man is to be hired to repair machines which breakdown at an average rate of

6 per hour. The breakdown follows Poisson distribution. The non-productive time of
a machine is considered to cost Rs.20 per hour. Two repairmen, Mr. X and Mr. Y
have been interviewed for this purpose. Mr. X charge Rs.10 per hour and he services
breakdown machines at the rate of 8 per hour. Mr. Y demand Rs.14 per hour and he
services at an average rate of 12 per hour. Which repairman should be hired?
(Assume 8 hours shift per day).
Solution:
=6 per hour
=
and
For Mr. X: average number of machines in the system:

Ls= machines

There fore, machine hours lost in an 8-hours shift=3 machine hours.

Total cost = hire charges + cost of idle machines

15
Operations research waiting lines (queuing theory)

= 18
For Mr. Y: average number of machines in the system:

Ls= machine

Therefore, machine hours lost in 8 hours shift

=1
Total cost = hire charges+ cost of idle machines
= 14
Hence it is advisable to hire repairman Y.

7. In a railway marshalling yard, goods trains arrive at a rate of 30 trains per day.
Assuming that the inter-arrival time follows an exponential distribution and the
service-time (the time taken to hump a train) distribution is also exponential with an
average of 36 minutes. Calculate the following:
a) The mean line length
b) The probability that the queue size exceeds 10.
If the input of trains increases to an average of 33 per day, then what will be the
change in (a) and (b)?
Solution:

Trains / minute

Trains / minute and

The traffic intensity is given by:

a) mean line length is given by : Ls= Trains

16
Operations research waiting lines (queuing theory)

b) prob. (queue size

When the input increases to 33 trains per day, we have

and

Then

Consequently, we obtain

a) Ls= or 5 trains

b) Prob. (queue size approx.

8. An air line has one reservation clerk on duty at a time. He handles information about
flight schedules and makes reservations. All calls to the airline are answered by an
operator. If a caller requests information or reservation, the, operator transfers the all
the reservation clerk. If the clerk is busy, then operator asks the caller to wait. When
the clerk becomes free the operator transfers to him the call of the person who has
been waiting for the longest duration. Assume that arrivals and services follow
Poisson and exponential distributions respectively. Calls arrive at the rate of 10 per
hour and the reservation clerk can a call in four minutes on the average.
a) What is the average number of calls waiting to be connected to the reservation
clerk?
b) What is the average time a caller must wait before reaching the reservation
clerk?
c) What is the average time for a caller to complete a call (i.e. waiting time plus
service time)?
Solution:

Per minute

Per minute

17
Operations research waiting lines (queuing theory)

and

a) average number of calls waiting:Lq=

b) average time a caller must wait: Wq= minutes

c) average time for a caller to complete a call: Ws= minutes

9. On an average 96 patients per 24 hour day require the service of an emergency clinic.
Also on average, a patient requires 10 minutes of active attention. Assume that the
facility can handle only one emergency at a time. Suppose that it costs the clinic
Rs.100 per patient treated to obtain an average servicing time of 10 minutes, and that
each minute of decrease in this average time would cost Rs.10 per patient treated.
How much would have to be budgeted by the clinic to decrease the average size of
the queue from 11/3 patient to ½ patient.

Solution:

Patient / minute; patient / minute

Expected No. of patients in the waiting line: Lq= (1/3)

But Lq= 1 (1/3) to reduced to =1/2

There fore substituting =1/2, in the formula

We get

18
Operations research waiting lines (queuing theory)

Which gives patient /

minute.
Hence the average rate of treatment required is minutes.
Consequently, the decrease in the average rate of treatment = 10-15/2=5/2 minutes
And the budget per patient 100+ so in order to set the required size of the
queue, the budget should be increased from Rs.100 to Rs.125 per patient.

10. In a large computer industry, the average rate of a system break down is 10
systems per hour. The idle time cost of a system is estimated to be Rs.20 per hour.
The working hours per day are 8. The manager of industry considers two mechanics
for repairing. The first mechanic A takes about 5 minutes on an average to repair a
system and demands wages Rs.10 an hour. The second mechanic B takes 4 minutes
in repairing and charges at rate of Rs.15 an hour. Assuming rate of system
breakdown is Poisson distributed and repair rate exponentially distributed, which of
the two mechanics should be appointed.
Solution:
For mechanic A:

Total cost = total wages + cost of non-productive time

= (hourly rate no. of hours) + (average no. of systems cost of idle time / hour number
of hours

TC= 10 8+

For mechanic B:

TC= 15 8+

19
Operations research waiting lines (queuing theory)

Mechanic B has the lowest total cost and also repair three systems more than Mechanic A
i.e for 8 hours, 24 systems (8 3=24). Thus mechanic B can save (24 20) =Rs.480 per
day. So that total cost of Mechanic A becomes =Rs.880+Rs.480=Rs1360.
So, mechanic B has to be appointed.

11. In the production shop of a company the breakdown of the machines is found to be
Poisson with an average of Rs.3 machines/hour. Breakdown time at one machine
costs Rs.40 per hour to the company. There are two choices before the company for
hiring the repairman. One of the repairman is slow but cheap, the other fast but
expensive. The slow-cheap repairman demands Rs.20 per hour and will repair the
broken down machines exponentially at the rate of 4 per hour. The fast expensive
repairman demands Rs.30 per hour and will repair machines exponentially at an
average rate of 6 / hour, which repair man should be hired?

Model II:
This is similar to model I with a difference only in queue discipline. Since the derivation
of Pn is independent of nay specific queue discipline, there fore in this model also we
have,
Pn=
Consequently other results will also remain unchanged in this queuing system as long as
Pn remains unchanged.
Model III:
This model is different from model I, in respect to the capacity of the system. Here the
capacity of the system is limited to N customers only and therefore different equations
derived in model I will be same for this model as long as n<N.
The system of steady state different equations for this model is
; n=0
; 1
and ; n=N
Using the usual procedure, the first two different equations become

20
Operations research waiting lines (queuing theory)

; n

and PN= ; n

Now in order to find the value of P0, use the fact that

1=

= =P0

Thus P0 =

And consequently Pn =

Here the steady-state solution exists even for this is due to the limited capacity of
the system regardless of the traffic intensity.

Characteristic of model III:

1. mean number of customers in the system:

L s=

2. Mean length: Lq = Ls-

3. Mean waiting time in the system: Ws=

4. Mean waiting time in the queue: Wq= Ws- or

21
Operations research waiting lines (queuing theory)

Problems on Model III

1. The railway marshalling yard is sufficient only for 9 trains (there being 10
lines, one of which is earmarked for the shunting engine to reverse itself from the
crest of the hump to the rear of the train). Trains arrive at the rate of 30 trains per day,
inter-arrival time follows an exponential distribution and service time distribution is
also exponential with an average of 36 minutes. Calculate the following:
a) the probability that the yard is empty.
b) The average line (number of trains in the system) length.

Solution:

Trains / minute. And trains / minute

Therefore

a) the probability that the yard is empty is given: P0 =

b) average number of trains in the system is given by: Ls=

=0.26 or 2 trains
2. If for a period of 2 hours in a day, trains arrive at the yard every 20 minutes but the
service time continues to remain 36 minutes, and then calculate for this period.

22
Operations research waiting lines (queuing theory)

a) The probability that the yard is empty

b) Average line length with the assumption that the line capacity of the yard
is limited to 4 trains only.
Solution:

Per minute and per minute

Which gives , and N=4.

a) The probability that the yard is empty: P 0 = , since (

b) Average queue length: P0

= 0.04 or 3 trains.
Model IV:
The most common and basic multi channel system contains parallel stations serving a
single queue on the first-come-first –served basis. All service stations provide the same
service. In this model there are S service stations in parallel, all having exponential
service times with mean 1/ customers enter in a queue and are served by the first
available server. This assumption eliminates the possibility that there are customers in the
queue and idle servers.
If the number of customers are less than to the number of servers (service
stations), i.e. n<s in the system, then there will be no queue and mean service rate will be
n . Alternatively, if n then all servers will be busy and mean service rate will be s .
Clearly s must be greater than if the queue is not to become bounded.
Following the procedure similar to those used in model I, we get different equations.
For n<s, we have the following equations:
; n=0.
;0
By solving these equations just as we did for the model I. we find tht

Pn= ;

23
Operations research waiting lines (queuing theory)

For n the set of equations to be solved become

;n
Solving recursively and using previous results, we have

The value of P0 may be obtained by using the fact that

Thus P0=

Characteristics of model IV
1. Mean number of customers in the queue (mean queue length):

Lq=

To evaluate the summation, let j=n-S

Then Lq=P0
2. Mean number of customers in the system:

Ls= Lq+

3. Mean waiting time of a customers in the queue

Wq=

24
Operations research waiting lines (queuing theory)

4. Mean waiting time that a customer spends in the system

Ws= Wq+

5. Proportion of time that the system will be partially idle is given by

1-

6. Probability that a customer has to wait is given by

P (n

Problems:
1. A super market has two girls at the sales counters. If the service for each customer is
exponential with mean 4 minutes, and if people arrive in a Poisson fashion at the
counter at the rate of 10 per hour, then calculate
a) the probability of having to wait for service
b) the expected percentage of idle time for each girl
c) If a customer has to wait, find the expected length of his waiting time.

Solution:
Per hour or per minute

Per minute and S=2

Traffic intensity

a) Probability of having to wait for service: P (n

Where P0=

= =

prob. (n

25
Operations research waiting lines (queuing theory)

b) the fraction of the time servers remain busy, i.e. traffic intensity
Thus expected percentage of idle time for each girl is given by
(1-
c) Expected waiting time in the system :( W: w>0)
= minutes
2. A company currently has two tool cribs, each having s single clerk in its
manufacturing area. One tool crib handles only the tools for the heavy machinery,
while the second one handles all other tools. It is observed that for each tool crib the
arrivals follow Poisson distribution with a mean of 20 per hour and the service time
distribution is negative exponential with a mean of 2 minutes.
The tool manager feels that, if tool cribs are combined in such a way that either
clerk can handle any kind of tools as demand arises, would be more efficient and the
waiting time could be reduced to some extent. It is believed that the mean arrival rate
at the two tool crib will be 40 per hour; while the service time will remain unchanged.
Compare the status of queue and the proposal with respect to the total expected
number of machines at the tool crib(s), the expected waiting time including service
time for each mechanic and the probability that he has to wait for more than five
minutes.
Solution:
a) when tool crib works independently:

Per hour or per minute

Per minute

Traffic intensity
i. average number of arrivals waiting in each tool crib:

Ls=

iii) Average waiting time of an arrival

Ws= Ls/ =2 minutes (each)

26
Operations research waiting lines (queuing theory)

b) when both the tool cribs work jointly

Per hour or per minute

Per minute and S=2

Traffic intensity

i) P0=

= =

ii) Ls=Lq+

=2.4
iii) Ws=3.6 minutes

3. A telephone exchange has two long distance operators. It is observed that, during the
peak load, long distance calls arrive in a Poisson fashion at an average rate of 15 per
hour. The length of service on these calls is approximately exponentially distributed
with mean length 5 minutes.
a) What is the probability that a subscriber will have to wait for his long
distance call during the peak hours of the day?
b) If the subscribers will wait and are serviced in turn, what is the expected
waiting time?

Solution:

a) p(n =

Where P0=

27
Operations research waiting lines (queuing theory)

p(n =

=0.48

b) Wq=

Lq

Lq=0.8

Wq = =0.8/ (1/4) =3.2 minutes

4. A bank has two tellers working on savings accounts. The first teller handles
withdrawals only. The second teller handles deposits only. It has been found that the
service time distributions for both deposits and withdrawals are exponential with
mean service time 3 minutes per customer. Depositors are found to arrive in a Poisson
fashion throughout the day with mean arrival rate of 16 per hour. Withdrawers also
arrive in a Poisson fashion with mean arrival rate of 14 per hour. What would be the
effect on the average waiting time for depositors and withdrawers if each teller
handles both withdrawals and deposits?

Solution:

per minute for both the tellers or per hour

Average arrival rate of depositors is given by

per hour and the arrival rate of withdrawers is given by
per hour
Average waiting time for depositors is given by

28
Operations research waiting lines (queuing theory)

hours or 12 minutes

Similarly, average waiting time for withdrawers is given by

Hours or 7 minutes.

If both tellers do service for withdrawers and depositors, then per hour and

S=2,

Thus P0=

And Wq=

= hours

5. A tax consulting firm has four service stations (counters) in its office to receive
people who have problems and complaints abut their income, wealth and sales taxes.
Arrivals average 80 persons in an 8 hour service day. Each tax adviser spends an
irregular amount of time servicing the arrivals which have been found to have an
exponential distribution. The average service time is 20 minutes. Calculate the
average number of customers in the system, average number of customers waiting to
be serviced, average time a customer spends in the system, and average waiting time
for a customer. Calculate how many hours each week does a tax adviser spend

29
Operations research waiting lines (queuing theory)

performing his job. What is the probability that a customer has to wait before he gets
service? What is the expected number of idle tax advisers at any specified time?
Solution:

per hour, per hour, S=4 and

1. probability of no customers in the system:

P0=

ii) Average number of customers in the system:

L s=

Average queue length Lq= Ls- customers

iv) average time a customer spends in the system

Ws=Wq+ hours

v) Average time a customer waits for service in the queue=Wq= =

hours

vi) Total time spent each week by a tax counselor:

Utilization factor, =5/6=0.833

The expected time spent in servicing customers during an 8-hours day is 8

hours.
Therefore, on an average, the tax adviser is busy for 33.30 hours based on a 40
hour week.

30
Operations research waiting lines (queuing theory)

vii) The probability that a customer has to wait: p(n =

viii) the expected number of idle advisers at any specified time can be obtained as
follows:
as the probability of no customers in the system is P 0, i.e., all the 4 counselors are
idle, therefore it is required to determine P 1, P2 and P3, i.e., the probability that 3
counselors are idle, 2 counselors are idle and 1 counselor is idle.

Now Pn=

Then P1=

P2=

P3=

There fore expected number of idle advisers is =4P0+3P1+2P2+1P3

= 4(0.021) +3(0.070) +2(0.118) +0.131=0.661
Hence less than one (0.661) adviser is idle on an average at any time.

The minimum cost of service rate

Let Cw = expected waiting cost / unit / unit time.
LS= expected average number of units in the system / unit time.
Cf = cost of servicing one unit.

The expected waiting cost per unit time (period) = Cw LS =Cw

Expected service cost per unit time (period) = Cf

Total cost = Cw + Cf

31
Operations research waiting lines (queuing theory)

This will be minimum if

32