WAITING LINES

(QUEUING)

Waiting line problems arise whenever demand for a service exceeds supply,
resulting in the formation of queues. Such problems occur everywhere for
example:
 Cars waiting at the traffic lights
 Passengers waiting for buses
 Customers waiting before a bank teller
 Aircraft waiting for clearances of runways
 Production process awaiting raw materials.

Components of Queue:

Customers in the queue Departing customers

Calling Source and/or Population ----------- Service Facility --------------

The Calling Source or Population can be either infinite or finite. Wherein big
population is infinite and small is referred to as finite.

Different Queues discipline: FCFS, LCFS and SIRO

Different Queuing configurations: single-queue, single server; single-queue,

multiple server; and multiple-queue, multiple-server.

The Single-server queuing model – it applies to waiting line problem which

share the following characteristics:
1. There is an unlimited single queue waiting line,
2. The number of arrival is Poisson-distributed,
3. The Calling population is infinite
4. The queue discipline is FCFS
5. The service time is exponentially distributed

Important notation to be used in this model,

λ – mean rate of customer arrivals, per time period

μ – the mean rate of service completions, per time period

ρ – λ / μ, the traffic density (or utilization rate), assumed to be less than one

P n = probability of exactly “n” customers being in the system

Using these symbols, we have the following results:

1
1. The probability that the system is idle, i.e that there is no customer at all in
the system, Po, is

Po = 1 – λ/μ; Po = 1 – ρ

2. The probability that the system is not idle, P (n>0), is therefore:

P (n>0) = ρ

3. The general formula for the probability that there are “n” units (customers)
in the system is given by:

Pn = (1- λ/μ) x (λ/μ)^n

= (1 – ρ) x ρ^n
= Po x ρ^n

where n = 0,1,2,3…..

P = 1 - Pn

4. Denoting by Ls, the expected number of customers in the system (waiting

or being serviced), we have:

Ls = λ / (μ – λ), or
= ρ /(1 – ρ)

5. Denoting by Ws, the expected time for an arriving customer in the system,
queuing and being served, we have:

Ws = 1 / (μ – λ), or
= Ls / λ

6. Similarly if Lq denotes the expected number of customers in the queue,

and Wq the expected queuing time for an arriving customer, then

Wq = (λ/μ) x 1 / (μ – λ)
= Ws – (1/μ)

Lq = (λ/μ) x (λ / (μ – λ))
= λ x Wq

2
Example 1:
Customers arrive at the Post Office at a rate of 8 per hour. There is one postal
clerk serving the customers at a rate of 20 per hour.

a. What is the probability that there is no customer in the office?

b. What is the probability that there are more than 2 customers in the office?
c. What is the probability that there is no customer waiting to be served?
d. What is the probability that a customer is being served and nobody is
waiting?
e. Calculate Ls, Lq, Ws and Wq?

The multiple server queuing model

There are situations where there is more than one server but only one queue,
this model can be applied to all waiting lines problems that has the following
characteristics:
1. There is an unlimited, single queue waiting line
2. The number of arrivals is Poisson distributed
3. The calling population is infinite
4. The queue discipline is FCFS
5. There are several servers and customer next in line will be served
6. All servers has the same rate of service, and the service time is
exponentially distributed

Formulas and conditions:

1. The probability that the system is idle, i.e that there is no customer at all in
the system, Po, is given by:

ρ = λ/μ, as before

Po =

2. The probability that there are “n” customers in the system, Pn, is given by:

There are two different expressions for Pn, as for n<c, some of the servers
are idle and the system is not using its full service capacity.

3. Denoting by Lq, the expected number of customers in the system, then:

4. Denoting by Ls, the expected number of customers in the system, then:

Ls = Lq + ρ

3
 5. Denoting by Wq the expected time for an arriving customer in the queue,
and by Ws the expected time in the system, then:

Example 2:

To register at Concordia University, students must report to the cashier’s office to

pay their fees. On the average, 5 students arrive at the cashier’s office per
minute. There are 2 cashiers working fulltime and each of them, on the average
can serve 3 students per minute. The students form a single waiting line in the
office to pay their fees. Students at the front of the queue will report to the first
vacant cashier.

a. What is the probability that the system is idle?

b. What is the probability that at least one student is waiting to be
served?
c. What is the average number of students in the cashier’s office?
d. What is the average waiting time for a student?

Economic analysis of waiting lines -trade off a reduction in waiting cost, due
to improved service against the increased cost for obtaining this service.

Example 3:

The typist at Gan’s Car Company receives on the average 22 letters per day for
typing. The typist works 8 hours a day and it takes on the average 20 minutes to
type a letter. The company has determined that the cost of a letter waiting to be
mailed is 80 cents per hour and additional costs to the company are $40 per day
(salary of typist + equipment cost).

1. What is the typist’s utilization rate?

2. What is the average number of letters waiting to be typed?
3. What is the average time needed to have a letter typed?
4. What is the total daily cost of letters waiting to be typed?

4
Example 4:

In example 3. Forced to improve the letter typing service, the company can either
hire an additional typist, or it can lease an automated typewriter. There are three
models of the automatic typewriter, the daily costs and resulting increases in the
typewriter’s efficiency is as follows:

Model Cost per day Increase in typist’s eff

I $37 50%
II $39 75%
III $43 150%

What action should the company take to minimize total costs?