REVISED

M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 217

C H A P T E R

Waiting Line and Queuing Theory Models

14

TEACHING SUGGESTIONS
Teaching Suggestion 14.1: Topic of Queuing. Here is a chapter that all students can relate to. Ask about student experiences in lines. Stress that queues are a part of our everyday lives and how things have changed at banks, post ofces, and airports in just the past decade. (We now wait in a common line for the rst available server.) Teaching Suggestion 14.2: Cost of Waiting Time from an Organizational Perspective. Students should realize that different organizations place different values on customer waiting time. Ask students to consider different scenarios, from a drive-through restaurant to a doctors ofce to a registration line in their college or motor vehicle ofce. It becomes clear that organizations place different values on their customers time (with most colleges and DMVs unfortunately placing minimal cost on waiting time). Teaching Suggestion 14.3: Use of Poisson and Exponential Probability Distributions to Describe Arrival and Service Rates. These two distributions are very common in basic models, but students should not take their appropriateness for granted. As a project, ask students to visit a bank or drive-through restaurant and time arrivals to see if they indeed are Poisson distributed. Note that other distributions (such as exponential, normal, or Erlang) are often more valid. Teaching Suggestion 14.4: Balking and Reneging Assumptions. Note that most queuing models assume that balking and reneging are not permitted. Since we know they do occur in supermarkets, what can be done? This is one of many places to prepare students for the need for simulation, the topic of the next chapter. Teaching Suggestion 14.5: Use of Queuing Software. The Excel QM and QM for Windows queuing software modules are among the easiest models in the program to use since there are so few inputs. Yet students should be reminded of how long it would take to produce the programs in Chapter 14 by hand. Teaching Suggestion 14.6: Importance of Lq and Wq in Economic Analysis. Although many parameters are computed for a queuing study, the two most important ones are Lq and Wq when it comes to an actual cost analysis. Teaching Suggestion 14.7: Teaching the New England Foundry Case. Here is a tip for this very teachable case. About half the students who tackle the case forget that time walking to the counter must be noted and that the return time also needs to be added.

ALTERNATIVE EXAMPLES
Alternative Example 14.1: A new shopping mall is considering setting up an information desk manned by one employee. Based on information obtained from similar information desks, it is believed that people will arrive at the desk at the rate of 20 per hour. It takes an average of 2 minutes to answer a question. It is assumed that arrivals are Poisson and answer times are exponentially distributed. a. Find the probability that the employee is idle. b. Find the proportion of the time that the employee is busy. c. Find the average number of people receiving and waiting to receive information. d. Find the average number of people waiting in line to get information. e. Find the average time a person seeking information spends at the desk. f. Find the expected time a person spends just waiting in line to have a question answered. ANSWER: a. b. c.
L
2

20/hour
1 0.66 20 30 20 1 20 30

30/hour
.33 33%

P0

2 people

d. e. f.

Lq
W

( 1

(20)2 1.33 people ) 30(30 20) 1 0.1 hour 30 20

Wq

l ( l)

20 30(30 20)

0.0667 hours

Alternative Example 14.2: In Alternative Example 14.1, the information desk employee earns $5/hour. The cost of waiting time, in terms of customer unhappiness with the mall, is $12/hour of time spent waiting in line. Find the total expected costs over an 8hour day. a. The average person waits 0.0667 hour and there are 160 arrivals per day. So total waiting time (160)(0.0667) 10.67 hours @ $12/hour, implying a waiting cost of $128/day. b. The salary cost is $40/day. c. Total costs are $128 $40 $168/day.

217

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 218

218

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

Alternative Example 14.3: A new shopping mall is considering setting up an information desk manned by two employees. Based on information obtained from similar information desks, it is believed that people will arrive at the desk at the rate of 20 per hour. It takes an average of 2 minutes to answer a question. It is assumed that arrivals are Poisson and answer times are exponentially distributed. a. Find the proportion of the time that the employees are idle. b. Find the average number of people waiting in the system. c. Find the expected time a person spends waiting in the system. ANSWER: (servers). a. P 20/hour, 30/hour, M 2 open channels

SOLUTIONS TO DISCUSSION QUESTIONS AND PROBLEMS

14-1. The waiting line problem concerns the question of nding the ideal level of service that an organization should provide. The three components of a queuing system are arrivals, waiting line, and service facility. 14-2. The seven underlying assumptions are: 1. Arrivals are FIFO. 2. There is no balking or reneging. 3. Arrivals are independent. 4. Arrivals are Poisson. 5. Service times are independent. 6. Service times are negative exponential. 7. Average service rate exceeds average arrival rate. 14-3. The seven operating characteristics are: 1. Average number of customers in the system (L) 2. Average time spent in the system (W) 3. Average number in the queue (Lq) 4. Average time in the queue (Wq) 5. Utilization factor ( ) 6. Percent idle time (Po) 7. Probability there are more than K customers in the system

1 1 20 0 ! 30
2 3
1 2 3
0

1 20 1 ! 30
1

1 20 2(30) 2 ! 30 2(30) 20

1 4 2 9
1 3 1 2

60 (60 20)
50%

b.

(20)(30)(20 / 30)2 1 (1)[(2)(30) 20]2 2

20 30

( 800 / 3) 1
1, 600 2
L 3/ 4 20

2 3

1 12

8 12

9 12

3 people p 4

14-4. If the service rate is not greater than the arrival rate, an innite queue will eventually build up. 14-5. First-in, rst-out (FIFO) is often not applicable. Some examples are (1) hospital emergency rooms, (2) an elevator, (3) an airplane trip, (4) a small store where the shopkeeper serves whoever can get his or her attention rst, (5) a computer system set to accept priority runs, (6) a college registration system that allows juniors and seniors to register ahead of freshmen and sophomores, (7) a restaurant that may seat a party of 2 before a party of 4 even though the latter group arrived earlier, (8) a garage that repairs cars with minor problems before it works on major overhauls. 14-6. Examples of nite queuing situations include (1) a rm that has only 3 or 4 machines that need servicing, (2) a small airport at which only 10 or 15 ights land each day, (3) a classroom that seats only 30 students for class, (4) a physician who has a limited number of patients, and (5) a hospital ward with only 20 patients who need care. 14-7. a. Barbershop: usually a single-channel, multipleservice system (if there is more than one barber). Arrivals Waiting line Service customers wanting haircuts seated customers who informally recognize who arrived rst among them haircut, style, shampoo, and so forth; if service involves barber, then shampooist, then manicurist, it becomes a multiphase system

c.

3 hr. 80

0.0375

Alternative Example 14.4: Three students arrive per minute at a coffee machine that dispenses exactly 4 cups/minute at a constant rate. Describe the operating system parameters. ANSWER: 3/minute
2

Lq

2 (

4/minute 9 2( 4)( 4 3)

Wq

1.125 people in queue on average 3 2 ( ) 2( 4)( 4 3)

0.375 minutes in the queue waiting 3 L Lq 1.125 4 1.87 people in the system 1 1 Wq .375 4 0.625 minutes in the system

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 219

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

219

b. Car wash: usually either a single-channel, single-server system, or else a system with each service bay having its own queue. Arrivals Waiting time Service dirty cars or trucks cars in one line (or more lines if there are service parallel wash systems); always FIFO either multiphase (if car rst vacuumed, then soaped, then sent through automatic cleaner, then dried by hand) or single-phase if all automatic or performed by one person

14-8. The waiting time cost should be based on time in the queue in situations where the customer does not mind how long it takes to complete service once the service starts. The classic example of this is waiting in line for an amusement park ride. Waiting time cost should be based on the time in the system when the entire time is important to the customer. When a computer or an automobile is taken into the shop to be repaired, the customer is without use of the item until the service is nished. In such a situation, the time in the system is the relevant time. 14-9. The use of Poisson to describe arrivals: a. Cafeteria: probably not. Most people arrive in groups and eat at the same time. b. Barbershop: probably acceptable, especially on a weekend, in which case people arrive at the same rate all day long. c. Hardware store: okay. d. Dentists ofce: usually not. Patients are most likely scheduled at 15- to 30-minute intervals and do not arrive randomly. e. College class: number of students come in groups at the beginning of class period; very few arrive during the class or very early before class. f. Movie theater: probably not if only one movie is shown (if there are four or more auditoriums each playing a different movie simultaneously, it may be okay). Patrons all tend to arrive in batches 5 to 20 minutes before a show.

c. Laundromat: basically a single-channel, multiserver, two-phase system. Arrivals Waiting line Service customers with dirty clothes usually rst-come, rst-served in terms of selecting an available machine rst phase consists of washing clothes in washing machines; second-phase is again queuing for the rst available drying machine

d. Small grocery store: usually a single-channel, singleserver system. Arrivals Waiting line customers buying food items customers with carts or baskets of groceries who arrive rst at the cash register; sometimes not FIFO; grocer may care for regular customers rst or give priority to person making a small, quick, purchase ringing up sale on cash register, collecting money, and bagging groceries

Service

14-10.
NUMBER OF CHECKOUT CLERKS 1 Number of customers Average waiting time per customer Total customer waiting time Cost per waiting hour Total waiting costs Checkout clerk hourly salary Total pay of clerks for 8-hour shift Total expected cost 300 1 hour 6 (10 minutes) 50 hours $10 $500 $8 $64 $564 Optimal number of checkout clerks on duty 2 300 1 hour 10 (6 minutes) 30 hours $10 $300 $8 $128 $428 3 3 300 1 hour 15 (4 minutes) 20 hours $10 $200 $8 $192 $392 4
1 hour 20

300

(3 minutes) 15 hours $10 $150 $8 $256 $406

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 220

220

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

14-11.

a. The utilization rate, , is given by

=
3 8 0.375

d. The utilization rate, , is given by 10 0.8333 12 e. The probability that no cars are in the system, P0, is given by:

P0

1 0.8333 0.1667

b. The average down time, W, is the time the machine waits to be serviced plus the time taken to perform the service. 1 W
1 8 3

14-13. 210 patrons/hour, 280 patrons/hour. a. The average number of patrons waiting in line, Lq, is given by 2 2102 44, 100 Lq ( ) 280(280 210) 280(70)
44, 100 2.25 patrons in line 19, 600 b. The average fraction of time the cashier is busy, , is given by 210 0.75 280 c. The average time a customer spends in the ticketdispensing system, W, is given by 1 1 1 W 280 210 70 0.0143 hour in the line 0.857 minute 51.4 seconds d. The average time spent by a patron waiting to get a ticket, Wq, is given by

0.2 day, or 1.6 hours c. The number of machines waiting to be served, Lq, is, on average,
2

Lq

32 8(8 3)

0.225 machine waiting d. Probability system Pn k that more than one machine is in the

2
k 1

3 9 0.141 Pn 1 8 64 Probability that more than two machines are in the system:

Wq

210 210 ( ) 280(280 210) 280(70) 210 0.011 hour 0.64 minute 19, 600

Pn

3 8

27 512

0.053

38.6 seconds e. The probability that there are more than two people in the system, Pn 2, is given by
Pn
k

Pn
Pn

3 8
3 8

81 4, 096
243 32 , 768

0.020

0.007

14-12.

10 cars/hour,
2

12 cars/hour.
102 (12)(2)

210 0.422 280 The probability that there are more than three people in the system, Pn 3, is given by Pn
2

a. The average number of cars in line, Lq, is given by

Lq ( ) 102 12(12 10)

210 0.316 280 The probability that there are more than four people in the system, Pn 4, is given by Pn
3

4.167 cars b. The average time a car waits before it is washed, Wq, is given by
Wq ( ) 10 12(12 10) 10 (12)(2)

Pn

210 280

0.237

0.4167 hours c. The average time a car spends in the service system, W, is given by 1 1 1 W 0.5 hour 12 10 2

14-14. 4 students/minute, 6 q 0 w 5 students/minute a. The probability of more than two students in the system, Pn 2, is given by

Pn Pn

4 5

0.512

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 221

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

221

The probability of more than three students in the system, Pn 3, is given by 4 4 Pn 3 0.410 5 The probability of more than four students in the system, Pn 4, is given by 5 4 0.328 Pn 4 5 b. The probability that the system is empty, P0, is given by
4 1 0.8 0.2 P0 1 1 5 c. The average waiting time, Wq, is given by 4 0.8 minute Wq ( ) 5(5 4) d. The expected number of students in the queue, Lq, is given by
2

f (part c). The average waiting time, Wq, for the twochannel system is given by 1 Wq W where

m

(m 1)!(m

)2
2

P0

Then

Wq

4 5 5
5(0.64)

(2 1)[2(5) 4]2

(0.429)

Lq

42 5(5 4)

3.2 students

(0.429) 1(10 4)2 1.373 0.038 minute 2.3 sec onds 1(36) f (part d). The average number of students in the queue for the two-channel system, Lq, is given by

e. The average number of students in the system, L, is given as 4 L 4 students 5 4 f. Adding a second channel, we have 4 students/minute 60 5 students/minute 12 m 2 f (part b). The probability that the empty, P0, is given by 1 P0 n n m 1 1 1 n 0 n! m! two-channel system is

Lq
where

L
(m 1)!(m 4 4(5) 5
2

)2

P0

Then

Lq

(2 1)![2(5) 4]2

(0.429)

4(5)(0.64)

m
2

m m

1(10 4) 0.15 student

(0.429)

5.492 1(36)

1 1 4 0! 5 4 5
0

f (part e). The average number of students in the twochannel system, L, is given by
2

1 4 1 5 1
2

2(5) 1 4 5 2(5) 4 1(2)

(m 1)!(m

)2

P0

1 4 2(5) 2 5 10 4

Lq

0.153

4 5

0.95 student

14-15.

30 trucks/hour,

35 trucks/hour.

4 5
1 4 5

1 1 16 10 2 25 6
160 300

a. The average number of trucks in the system, L, is given by

L
30 35 30 30 5 6trucks in the system

or

P0

1 1 0.8 0.53

1 2.33

0.429

Thus the probability of an empty system when using the second channel is 0.429.

b. The average time spent by a truck in the system, W, is given by 1 W 1 35 30 1 5 0.2 hour 12 minutes

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 222

222

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

c. The utilization rate for the bin area, , is given by

30 6 0.857 35 7 d. The probability that there are more than three trucks in the system, Pn 3, is given by
Pn Pn
k

To determine total cost using the second clerk (a second channel): 1 P0 n m 1 1 n 1 m m n 0 n! m! m

1 1 12 0 ! 15
4 5
0

30 0.540 3 35 Thus the probability that there are more than three trucks in the system is 0.540.
e. Unloading cost:

1 12 1 15 1
2

2(15) 1 12 15 2(15) 12 1(2)

CM

16

hours trucks hours dollars 30 hour 0.2 truck 18 hour day

or

4 5

1 4 2(15) 2 5 30 12 1 1 16 30 2 25 18

1 1 4 5 480 900

16(30)(0.2)(18) $1,728/day or $12,096 per week. f. Enlarging the bin will cut waiting costs by 50% next year. First, we must compute annual waiting costs: annual waiting cost

P0

weeks days dollars 7 1, 728 year week day

$24,192 Enlarging the bin will cut waiting costs by 50% next year, resulting in a savings of $12,096. Since the cost of enlarging the bin is only $9,000, the cooperative should proceed to enlarge the bin. The net savings is $3,096 ($12,096 $9,000). 14-16. 12 calls/hour, 6 0r 15 calls/hour. a. The average time the catalog customer must wait, Wq, is given by Then

Wq

1 1 0.8 0.53 1 0.429 2.33 m

(m 1)!(m
12 15 15
2

)2

P0

Wq

(2 1)[2(15) 12]2

(0.429)

Wq

) 12 15(3) 12 45 0.267

12 15(15 12)

(0.429) 1(30 12)2 4.12 0.0127 hour 1(324)

Cost with two clerks: Ct/hour service cost
20 12

15(0.64)

0.763 seconds

16 minutes b. The average number of callers waiting to place an order, Lq, is given by
2

waiting cost

Lq

(
2

)
144 15(3) 144 45 3.2 customers

calls hours dollars 0.0127 50 hour hour call 20 12(0.0127)(50) 20 7.62 $27.62/hour

12 15(15 12)

There is a savings of 170.20 27.62 clerk should certainly be added! 14-17. This is an M/M/1 system with 30 per hour. a. b. c. d. e. f. g. P(n W L Wq Lq P0 0.167 hours 4 0.133 3.2 0.2 0.8 2) Pn
1

142.5/hour. Thus a second 24 per hour and

c. To decide whether or not to add the second clerk, we must (a) compute present total cost, (b) compute total cost with the second clerk, and (c) compare the two. Present total cost: Ct/hour service cost waiting cost
calls hours dollars 10 12 0.267 50 hour hour call

10

12(0.267)(50)

10

160.2

$170.20/hour

Pn

0.640

0.512

0.128

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 223

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

223

14-18. This is an M/M/2 system with 24 per hour and 30 per hour. Using QM for Windows we get the following: a. b. c. d. e. f. g. P(n W L Wq Lq P0 0.0397 hours 0.9524 0.0063 0.1524 0.4286 0.4 2) 0.1371 Pn
1

Pn

0.2286

0.0914

3 (0.4545) 4 (1)!(8 3)2 0.873 0.291 hour W 3 3 0.123 Lq 0.873 4 0.123 Wq 0.041 hour 3 L

3 3( 4) 4

0.873

14-19 and 14-20.

NUMBER OF FRUIT LOADERS 1 Truck arrival rate ( ) Loading rate ( ) Average number in system (L) Average time in system (W) Average number in queue (Lq) Average time in queue (Wq) Utilization rate ( ) Probability system empty (P0) Probability of more K than K trucks in 0 system 1 2 3 3/hour 4/hour 3 trucks 1 hour 2.25 trucks 3 hour 4 0.75 0.25 2 3/hour 8/hour 0.6 truck 0.2 hour 0.225 truck 0.075 hour 0.375 0.625

By looking back to Problems 14-19 and 14-20, we see that although length of the queue and average time in the queue are lowest by opening the second platform, the average number of trucks in the system and average time spent waiting in the system are smallest when two workers are employed loading at a single platform. Hence we would probably recommend not building a second gate. 14-23. The queuing systems in this problem are the M/M/2, M/M/3, and the M/M/4 systems. a. Wq 0.0643 for 2 channels; Wq Wq 0.0015 for 4 channels; 0.0079 for 3 channels;

b. The total time spent waiting is Wq(10 hours per day). This is 19.29 hours with 2 channels, 2.37 hours with 3 channels, and 0.45 hours with 4 channels. c. The total daily waiting time cost is given in the table below:
Service Service # cost per cost per Channels hour day 2 3 4 $20 $30 $40 $200 $300 $400 Total Total waiting time waiting Wq(10hr.) cost 19.29 2.37 0.45 $1929 237 45 Total cost $2129 $537 $445

0.75 0.56 0.42 0.32

0.375 0.141 0.053 0.020

These results indicate that when only one loader is employed, the average truck must wait 3 r hour before it is loaded. Furthermore, there are an average of 2.25 trucks waiting in line to be loaded. This situation may be unacceptable to management. Note the decline in the queue when a second loader is employed. 14-21. Referring to the data in Problems 14-19 and 14-20, we note that the average number of trucks in the system is 3 when there is only one loader and 0.6 when two loaders are employed.
NUMBER OF LOADERS 1 Truck driver idle time costs ( average number trucks hourly rate) (3)($10) Loading costs Total expected cost per hour 2

The minimum daily cost is $445 with 4 channels. 14-24. This is an M/M/1 system with 15 per hour. a. b. c. d. e. Wq Lq W L P0 0.1333 hours 1.333 0.2 hours 2 0.333 10 per hour and 10 per hour and

$30 36 $36

$6 12 $18

(0.6)($10) (2)($6)

14-25. This is an M/M/2 system with 15 per hour. a. b. c. d. e. Wq Lq W L P0 a. 0.0083 hours 0.083 0.075 0.75 0.5 (8 hours per day) 10(8)

The rm will save $18/hour by adding the second loader. 14-22.

P0

1 1 1 3n n 0 n ! 4 2( 4) 1 3 2 ! 4 2( 4) 3
2

14-26.

80 customers per day

1 1 3 4 1 3 8 2 4 8 3
2

b. Total time spent waiting Wq(number of customers) 0.1333(80) 10.66 hours. Total waiting time cost $25(10.66) $266.5 c. With 2 tellers, total time spent waiting 0.664 hours. 0.0083(80)

0.454

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 224

224

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

Total waiting time cost $25(0.664) $16.60 d. Total cost with 1 teller $266.5 $96 $362.5 Total cost with 2 tellers $16.60 2($96) $208.60 14-27. a. Average number in line b. Average number in system c. Average wait in line 14-28. For M L 1( But
P0

1 0.88 0.62

1 0.30 0.12

0.02

1 2.944

0.34 a. Average number waiting Lq

0.666 1.333 10 seconds 0.1666 minute

(1 P0 )

1: P0 )2
1

P0

4.706 (0.66) 5 5 0.706 b. Average number broken

Lq c. P0 0.34, as seen above.

4.4 0.6
L (1 P0) (1 0.34) 1.24

0.576
1 1

1
2

d. Average time in queue

Thus

(
2

)
( (

(
1 6

(
)( )

)( )
e. Average time in system

( N L) 0.576 (5 1.24)(0.706) 0.576 = = 0.217 hour 2.65

Wq

Lq

)( )

W
0.217

Wq
0.25

This is the same formula. 14-29.

0.1667 / day
1/day N
P0

/ day
14-31. a. Entering: Exiting:
1
1 60( 6 )3 1 120( 6 ) 4 1 120( 6 )5

0.467 hour 84/minute, 48/minute, 0.1 minute 30/minute, 30/minute, 2.8 1.6

5, n

1 1 1 5( 6 ) 20( 6 )2

The manager desires that Wq Lq 8 customers in queue. Entering: If M If M If M 3, Lq 4, Lq 5, Lq

6 seconds and that

0.36 a. Number in queue

Lq N (1 P0 ) 5 7(1 0.36)

12.27 and Wq 1.00 and Wq 0.24 and Wq

0.14 minute (too high) 0.01 minute (this is okay) 0.003 minute (this is also okay) 4 or more entrances.

5 L Lq 0.52

4.48 (1

0.52 unit So the manager must open M 0.52 N 3.84 L (1 5 0.36) 1.16 Exiting: If M If M 2, Lq 3, Lq 2.8, Wq 0.31, Wq P0)

b. Number in the system 0.64 1.16 in system

0.06 minute (this is okay) 0.006 minute (also okay)

c. Number running ok

d. Average time in queue Lq 0.52 Wq ( N L) (5 1.16)(0.1667 ) e. Average wait in system

So the manager must open M 2 or more exits. Since there are only 6 turnstiles, 4 must be used as entrances and 2 as exits.
0.817 days

Wq
1

0.817
60 85

b. The students should recognize and question all the limiting queuing assumptions that have been applied in solving the case. For example, it may be reasonable to assume that arrivals at the entrance turnstiles are independent and Poisson. But are exiting passengers independent? More realistically, they arrive in batches (as a train arrives), and unless trains unload every minute or two, this assumption may be unreasonable. Other problems arise as well. If an exiting passengers card does not have the correct fare, the card is rejected and the passenger must leave the line, go to an add fare machine to correct the deciency, and enter the queue again. This resembles the reneging customer. Note: In the real-world subway station in Washington, D.C., common queues are not formed at turnstiles and the problem becomes a series of single channel queues.

1.817 days 14-30.

0.706 /hour , 0.1765

4/hour,

5, n

P0

1 1 5(0.1765) 20(0.1765)2
60(0.1765)3 120(0.1765)4 120(0.1765)5

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 225

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

225

14-32. This is an M/M/1 system with per hour. a. Wq = 0.27 hours b. d. Lq = 3.2 W = 0.33 hours
3

= 12 per hour and

= 15

f. Wq g. 14-37. c. d. P0 e. Lq Wq W P(n P(n P(n W a.

0.5 minute 1 minute 3 per minute and 3/4 1 2.25 0.75 minute 1 minute 1) 2) 3) 0.188 0.141 0.106 0.75 3/4 0.25. The cashier is idle 25% of the time. 4 per minute

b. M/M/1

c. L = 4 e. Pn = (12/15)3+1 = 0.4096 = 12 per hour and = 15

14-33. This is an M/M/2 system with per hour. a. Wq = 0.013 hours b. c. d. Lq = 0.152 L = 0.952 W = 0.079 hours

f. g. h.

SOLUTIONS TO INTERNET HOMEWORK PROBLEMS

14-34. a. b. c. d. e. f. g. 12/hour; 4/hour/barber; M 4 channels P0 0.0377 3.8% (from formula) L 4.528 W 0.377 hour 22.6 minutes Wq 0.127 hour 7.6 minutes Lq 1.5282 (from formula) 0.75 75% with m 5 barbers drops to 60% 6 patients/hour; 5 patients/

14-38. This is an M/M/2 system with 3 per minute and 4 per minute. Solving with QM for Windows we obtain the following: a. b. c. Lq Wq W 0.1227 0.0409 minute 0.2909 minutes 1) 0.3409, P(n 2) 0.1278, P(n 3) 0.0479.

d. P(n

SOLUTION TO NEW ENGLAND FOUNDRY CASE

1. To determine how much time the new layout would save, the present system must be compared to the new system. The amount of time that an employee spends traveling to the maintenance department added to the time that he or she spends in the system being serviced and waiting for service presently, compared to this value under the proposed system, will give the savings in time. Under the present system, there are two service channels with a single line (M 2). The number of arrivals per hour is 7 ( 7). The number of employees that can be serviced in an hour by each channel is 5 ( 5). The average time that a person spends in the system is

14-35. a. 9 A.M.3 P.M.; hour/doctor 6 1.20 5

Want Wq to be 5 minutes 0.0833 hour. Wq 0.0833 implies that Lq Wq 0.0833 or Lq 0.0833 or Lq 0.50 Thus m 3 channels or doctors are needed (with m 0.6748; with m 3, Lq 0.0904). b. 3 P.M.8 P.M.; patients/hour/doctor 4 0.80 5 Wq Lq 4 patients/hour; 5 2, Lq

W
where
P0

( / )M ( M 1)!( M
M 1 1 n 0 n!
n

P0

1 1 M!
M

M M

0.0833 hour implies that 0.03333. This means m c. 8 P.M.midnight; hour/doctor 12 2.4 5

Lq

0.0833 or Lq

0.0833 or

In this case

2 doctors. 12 patients/hour; 5 patients/

P0

1 1 (1) 1 1 7 1 5
1

1 7 2(5) 2 5 5(2) 7

0.18

Therefore,

W
Lq 0.0833 or Lq
0.0833 or Lq

5(7 / 5)2 1(10 7 )2

(0.18) 1 / 5

Want Wq 1.00. m

0.0833 hour or

0.396 hour, or 23 minutes and 45 seconds Added to the travel times involved (6 minutes total for maintenance personnel and 2 minutes total for molding personnel), the total trip takes: For maintenance29 minutes and 45 seconds For molding25 minutes and 45 seconds

4 doctors are needed. 1 per minute and 0.5 1 0.5 1w 0.5. The cashier is idle 50% of the time. 2 per minute

14-36. a. b. M/M/1 c. d. e. P0 Lq 1w

Under the new system, waiting lines are converted to singlechannel, single-line operations. Bob will serve the maintenance

REVISED
M14_REND6289_10_IM_C14.QXD 5/12/08 1:01 PM Page 226

226

CHAPTER 14

WAITING LINE

AND

QUEUING THEORY MODELS

personnel and Pete will serve the molding personnel. Bob can now service 6 people per hour ( 6). Four people arrive from the maintenance department every hour ( 4). The time spent in Bobs department is 1 1 1 hour, or 30 minutes W 6 4 2 The reduced travel time is equal to 2 minutes, making the total trip time equal to 32 minutes. This is an increase in time of 2 minutes and 15 seconds for the maintenance personnel. Pete can now service 7 people per hour ( 7). Three people arrive from the molding department every hour ( 3). The time in Petes department is 1 1 W hour, or 15 minutes 7 3 4 The travel time is equal to 2 minutes, making the total trip time equal to 17 minutes. This is a decrease in time of 8 minutes and 45 seconds per trip for the molding personnel. 2. To evaluate systemwide savings, the times must be monetized. For the maintenance personnel who are paid $9.50 per hour, the 2Z\v minutes lost per trip costs the company 36 cents per trip [2Z\v 60 0.0375 of an hour; 0.0375(9.50) $0.36]. For the molding personnel who are paid $11.75 per hour, the 8 minutes and 45 seconds per trip saved saves in monetary terms $1.71 per trip. The net savings is $1.71 0.36 $1.35 per trip. (Students may also nd the cost savings on an hourly or daily basis.) Because the net savings for the new layout is small, other factors should be considered before a nal decision is made. For example, the cost of changing from the old layout to the new layout could completely eliminate the advantages of operating the new layout. In addition, there may be other factors, some noneconomic, that were not discussed in the case that could cause you to want to stay with the old layout. In general, when the cost savings of a new approach (a new layout in this case) is small, careful analysis should be made of other factors.

average time in the system gives W 7.6 minutes. This plan is clearly faster. Use of an ATM with the same service rate as the clerks (20 per hour) by 20 percent of the arrivals (18 per hour) gives the same average time for these guests as the current systems30 minutes. The remaining 72 per hour form an M/M/4 or M/M/5 queuing system. With four servers, the average time in the system is 8.9 minutes, resulting in an overall average of: 0.2 30 0.8 8.9 13.1 minutes With ve servers, the average time is 3.9 minutes resulting in an overall average of: 0.2 30 0.8 3.9 9.1 minutes

INTERNET CASE STUDY

Pantry Shopper
Beth wants to get a general idea of the system behavior. She rst will need to decide whether she is interested in time waiting or time in system. Some students may use system time, but since most shoppers are relieved when it is their turn, we use waiting time as our measure. For all of our analyses, we use current service times, even though a UPC reader is going to be installed. This means that our waiting times are an upper bound for the new, better system (the M/M/s model). We begin with a rough analysis (one that is going to have a very interesting feature, by the way). We assume that there are no express lanes. Then, we want to nd the average service time and rate. The time is given by t .2(2 min.) .4 3.2 3.6 min. .8(4 min.)

SOLUTION TO WINTER PARK HOTEL CASE

1. Which of the two plans appears to be better? The current system has ve clerks each with his or her own waiting line. This can be treated as ve independent queues each with an arrival time of 90/5 18 per hour. The service rate is one every 3 minutes, or 20 per hour. Assuming Poisson arrivals and exponential service times, the average amount of time that a guest spends waiting and checking in is given by 1 Ws

0.5 hour, or 30 minutes 20 18 If 30% of the arrivals [that is, 0.3(90) 27 per hour] are diverted to a quick-serve clerk who can register them in an average of 2 minutes ( 30 per hour) their average time in the system will be 20 minutes. The remaining 63 arrivals per hour would distribute themselves equally among the four remaining clerks ( 63/4 15.75 per hour), each of whose mean service time is 3.4 minutes (or 0.5667 hour), so that 1/0.5667 17.65 per hour. The average time in the system for these guests will be 0.53 hour or 31.8 minutes. The average time for all arrivals would be 0.3(20) 0.7(31.8) 28.3 minutes. A single waiting line for the ve clerks yields an M/M/5 queue with 90 per hour, 20 per hour. The calculation of

This means that the average service rate is 60/3.6 16.67 customers per hour. Notice that this is not the same as taking 20 percent of the rate of 30 and 80 percent of the rate of 15, which would equal 18 and would be wrong. Using an arrival rate of 100 and a service rate of 16.67, the minimum number of servers is 6. (This is due to round off.) In reality, the minimum number is 7, and the average waiting time is 2.2 minutes. Trying one more server leads to a waiting time of .64 minutes. Now we separate the express and regular. Assume that all express customers go into the express (even though they can go into any lane) and assume that all non-express customers go into the proper lanes (even though we all have seen people with twenty packages get into a ten-items-or-less line). For the express lane, with an arrival rate of 20 and a service rate of 30, one server yields an average wait of 4 minutes, while two servers yield an average wait of .25 minutes. For the regular lane, with an arrival rate of 80 and a service rate of 15, 6 servers yield an average wait of 4.28 minutes and 7 servers yield an average wait of .98 minutes. If Beth uses 7 servers, they will be split this way: 6 in regular lanes and 1 in an express lane. If Beth uses 8 severs, a 62 split between regular lanes and express lanes yields an average wait of (.2)(.25) (.2)(4) (.8)(4.28) (.8)(.98) .8 .05 3.424 3.47 min. A 71 split yields an average of .784 1.584 min., which is better. However, the express lane would be slower than the regular lanes!