Sie sind auf Seite 1von 4

Chapter 17: Waiting Lines (Queues)

A queue is a waiting line. A queuing system involves customers arriving for service who
sometimes have to wait.
Queuing analysis provides:
Summary measures for assessing a queuing system in terms of customers and time.
A way to balance the costs of providing service and costs of congestion.
Importance of Queuing Analysis:
Servicing customers can be costly.
Retail environments are plagued with customer congestion. Managing that has
benefits.

Single-Server Queuing Model (M/M/1).


Assumes exponential distribution for times:
between successive arrivals—mean rate λ
to complete service—mean rate μ
Results:
Probability for n customers in the system λ λ
n

P0 =1 − Pn = 
µ µ
 P0
 
Mean number of customers in the system λ
L=
µ −λ
Mean customer time spent in the system W =
L
=
1
λ µ −λ
Mean number of customers waiting
λ
2

Lq =
µ (µ − λ )
Mean customer waiting time Lq λ
Wq = =
λ µ ( µ − λ)
Service utilization factor λ
ρ=
µ

Problem: Customers arrive at a supply room with mean rate λ = 25/hr. Service take
a mean of 2 minutes, so that µ = 30/hr. Do a queuing analysis. Solution:
25 1
L= = 5 cust. W= = .20 hr.
30 − 25 30 − 25

Lq =
( 25 ) 2 = 4.167 Wq =
25
= .167
30 ( 30 − 25 ) 30 ( 30 − 25 )

ρ = 25/30 = .833 P0 = 1 − .833 = .167


1 2
P1 = (25/30) (.167) = .139 P2 = (25/30) (.167) = .116
Problem (continued): Hourly cost of providing service is $10 (clerk’s wages). Lost
productivity of employee customers while getting supplies is $7/hr, the congestion
penalty. Evaluate total cost.
Solution (continued): The hourly system cost may be used to compare alternatives:
TC = hourly queuing cost + hourly service cost
= $7×W ×λ + $10 = $7(.20)(25) + 10 = $45
The queuing cost accounts for all λ customers in the system during the hour,
and collective time under a congestion penalty, W × λ , applies above.

When the congestion penalty involves just the waiting time, Wq × λ is used
for queuing cost. Retail customers resent the wait, not the time receiving
service.

Multiple servers involve greater complexity.


Chapter 17 Problems

2. Patrons arrive at the Dogpatch post office at the rate of 30 per hour. There is one clerk
on duty, who takes an average of 1 minute to serve each customer. Service times are
approximately exponential.
a) Calculate the mean customer times (1) spent waiting in line and (2) spent receiving or
waiting for service. Also, find the mean number of customers (3) in line and (4)
receiving or waiting for service.
b) Construct the probability distribution for the number of customers inside the post
office, for values of n from 0 to 6.

λ = 30/hour µ = 60/hour
(a) (1)
30 1
Wq = = hour or 1 minut e
60 (60 −30 ) 60
(2)
1 1
W = = hour or 2 minut e
(60 − 30 ) 30
(3)
 1 
L q = 30   = .5 person
 60 
(4)
 1 
L = 30   = 1 person
 30 
(b) Po = 1 − λ /µ = 1 − 30/60 = .50

n Pn = (λ /µ )nP0

0 .5000000
1 .2500000
3 .0625000
4 .0312500
5 .0156250
6 .0078125
8. Ace Airlines has one reservations clerk on duty at a time to handle information about
flight times and make reservations. All calls to Ace Airlines are answered by an
operator. If a caller requests information or reservations, the operator transfers the call to
the reservations clerk. If the clerk is busy, the operator asks the caller to wait. When the
clerk becomes free, the operator transfers the call of the person who has been waiting the
longest. Assume that arrivals and services can be approximated by a Poisson process.
Calls arrive at a rate of 10 per hour, and the reservations clerk can service a call in an
average of 4 minutes.
a) What is the average number of calls waiting to be connected to the reservations clerk?
b) What is the average time that a caller must wait before reaching the reservations
clerk?
c) What is the average time it takes for a caller to complete a call?

λ = 10/hourµ = 15/hour
(a) Lq = 4/3 calls
(b) Wq = 2/15 hour or 8 minutes
(c) W = 1/5 hour or 12 minutes

9. Suppose that the management of Ace Airlines in Problem 8 is considering installing


some visual display equipment and a new reservations system. One of the benefits of this
system is that it will reduce the average time required to service a call from 4 to 3
minutes.
a) Under the new system, what would be the average number of calls waiting to be
connected?
b) Under the new system, what would be the average waiting time for a caller?

(a) µ = 20/hour Lq = .5 calls


(b) Wq = .05 hour or 3 minutes

10. Mildred’s Tool and Die Shop has a central tool cage manned by a single clerk, who
takes an average of 5 minutes to check and carry parts to each machinist who requests
them. The machinists arrive once every 8 minutes on the average. Times between
arrivals and for service are assumed to both have exponential distributions. A
machinist’s time is valued at $15 per hour, a clerk’s time is valued at $9 per hour. What
are the average hourly queuing system costs associated with the tool cage operation?

λ = .125 customers/minute or 7.5 per hour


µ = .20 customers/minute or 12 per hour
Each machinist spends an average of W = 1(12 − 7.5) = .2222 hours for every trip to
the tool cage. Altogether, 7.5 machinists arrive on the average each hour, so that the
average hourly cost for machinists' lost time is
7.5(.2222) × $15 = $25
Adding the clerk's cost, the hourly queuing system cost is
$25 + $9 = $34

Das könnte Ihnen auch gefallen