Sei sulla pagina 1di 36

Chapter 11 - Managing Capacity and Demand

CHAPTER 11
MANAGING CAPACITY AND DEMAND

TEACHING NOTE
Many of the approaches to matching capacity and demand discussed in the chapter should be familiar
to students and prompt lively discussion. The controversial practice of "overbooking" is treated in
economic terms, but the behavioral implications of this policy for both service personnel and
consumers should be discussed. This chapter presents a number of techniques to assist management
in matching supply and demand, but leaves to the imaginative student the opportunity for other
creative approaches. The chapter concludes with a discussion of yield management, a computer-based
approach for maximizing revenues. Instructions for playing the yield management game are given in
the textbook and students should be encouraged to familiarize themselves with the material before
“game day.” The tally sheets and other instructional materials are presented in this manual.
SUPPLEMENTARY MATERIALS
Case: American Airlines, Inc. Revenue Management (HBS 9-190-029)
Following deregulation, American Airlines embarked on an ambitious program of revenue
management by building on its SABRE reservation system. The case presents the reader with pricing
decisions on two routes and describes the development of yield management.
Case: University Health Services: Walk-In Clinic (HBS 9-681-061)
In response to complaints of excessive waiting, a triage system was introduced and one year later the
director reviews the results to decide whether or not the patient waiting time is acceptable.

LECTURE OUTLINE
1. Generic Strategies of Level Capacity and Chase Demand (Table 11.1 and Figure 11.1)

2. Strategies for Managing Demand

Customer-induced Variability (Table 11.2)

Segmenting Demand (Figure 11.2 and Table 11.3)

Offering Price Incentives (Tables 11.4 and 11.5)

Promoting Off-peak Demand

Developing Complementary Services

Reservation Systems and Overbooking (Tables 11.6 and 11.7)

11-1
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

3. Strategies for Managing Capacity

Defining Service Capacity

Daily Workshift Scheduling (Figures 11.3, 11.4, 11.5)

Weekly Workshift Scheduling with Days-off Constraint (Table 11.8)

Increasing Constomer Participation

Creating Adjustable Capacity

Sharing Capacity

Cross-training Employees

Using Part-time Employees

Scheduling Part-time Tellers at a Drive-in Bank (Figures 11.6 and 11.7)

4. Yield Management (Figures 11.8, 11.9, 11.10, 11.11, 11.12)

Yield Management Applications

TOPICS FOR DISCUSSION


1. What organizational problems can arise from the use of part-time employees?
The use of part-time employees can be very helpful to businesses that have peak demand periods such
as restaurants, supermarkets, and banks. Part-time employees usually are paid lower wages and they
enjoy fewer, if any, of the company benefits that are provided for full-time employees. Also, it is not
generally feasible for a company to offer career development incentives to part-time employees and it
is more difficult to fit them into the organizational structure. In addition, part-time employees
generally have lower experience levels than do full-time workers. As a result, part-time employees
might have poorer attitudes and less loyalty and commitment, which could affect reliability,
performance, and the quality of work. This situation can have a direct impact on customers and the
business. In view of these conditions, part-time employees might require greater supervision and
control than would be necessary for full-time employees. Also, there usually is a greater turnover in
part-time employees, so more time must be spent in training new employees. Finally, when a business
must hire more part-time employees than full-time people to staff positions, more administrative work
for work scheduling, personnel records, and payroll is required.
2. How can computer-based reservation systems increase service capacity utilization?
The main function of the reservation system is to pre-sell the service. A reservation system allows the
customer to reserve a service long before it actually is utilized. Allowing the customer to make
reservations has numerous advantages.
The reservation system can be used to deflect demand to other times or locations where
service capacity is available. For example, if a passenger wants a flight that is full, a reservation clerk
can suggest immediately alternative flights that are available. Thus, the demand for service

11-2
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

capacity can be rerouted to under-utilized capacity effectively. Reservation systems also


allow the service to overbook its capacity when it reasonably expects to have no-shows.
Finally, when computer-based systems are standardized among different service organizations
in the same industry, each participant can make use of a larger information base. For example, a
passenger can call a Delta Air Lines reservation agent and book a flight on American Airlines if a
Delta flight is unavailable. Such a system benefits all of the participants and the customers.
3. Illustrate how a particular service has implemented strategies for managing both demand and
capacity successfully.
Because a service is produced and consumed simultaneously, a failure to provide enough capacity to
serve results in idle servers and facilities. Public schools have been experiencing this variability in
demand recently owing to fluctuations in the numbers of school age children, a move of families from
central cities to the suburbs, and increasing enrollments in private schools. The agrarian pattern of
closing school during summer months has contributed to idle servers and facilities during that period.
As a public service, education has been somewhat slow at adapting to its fluctuating demand,
but the recent trend of more competition for scarce resources in public services and the public outcry
for educational accountability has spurred some strategies for managing the demand and supply. Some
examples follow:
Managing demand:
 Expand educational services by offering adult education, early childhood education programs, and
before- and after-school childcare programs.
 Offer social services such as community education programs, job retraining programs, programs
for senior citizens, and summer recreation programs.
 Offer incentives to attend particular schools by making them magnet schools, open-area schools,
or open-enrollment schools, which any child in the community may attend.
 Promote off-peak demand in some areas by offering summer school classes.
 Partition demand by using educational voucher systems.
Managing capacity:
 Use school buildings where classes are no longer held because of declining enrollment for
warehousing school supplies, or for administrative offices, or lease those to other organizations
until school demographics change.
 Arrange flexible schedules on a daily basis and a school-year basis for crowded schools. For
example, on a daily basis in an elementary school, schedule hours for kindergarten through third
grade students from 8 a.m. to 2 p.m. and hours for fourth through sixth grade students from 10
a.m. to 4 p.m. Another alternative is to stagger the students throughout the calendar year instead
of offering classes only during the typical nine-month school year, i.e., some students would
attend school December to September, others would go from June to March, and still others would
attend the typical September to June period.
 Merge two or more grades in under-utilized schools or, in a crowded grade school, shift sixth
graders to a junior high/middle school.

11-3
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

 Hire teachers who are trained in more than one subject in order to cope with a fluctuating demand
for courses.
 Hire part-time teachers in small schools for support areas such as art, music, and physical
education.
 Share capacity with other public service agencies to utilize space better.
4. What possible dangers are associated with developing complementary services?
One of the possible dangers associated with developing complementary services is an increase in the
firm's liability because of those added services. For instance, adding cold and hot sandwiches at
convenience grocery stores incur the possibility that a customer might suffer food poisoning and take
legal action against the store. Another example is seen with the addition of self-service gas pumps,
which involve a risk of fire, damage, and injury.
Another problem associated with offering complementary services can occur when the
complementary service attracts customers who might hurt business. Consider a shopping mall that
installs a video arcade. The arcade can become a haven for noisy and rowdy teenagers who will drive
away those customers who want to shop in peace.
Finally, a complementary service should enhance the firm's image; as noted above, a poorly
selected complementary service could compromise business seriously.
5. Will the widespread use of yield management eventually erode the concept of fixed prices
for any service?
Most services exhibit, to some extent, the capacity-constrained dilemma faced by airlines and hotels
that are unable to inventory their products (seats on a flight or rooms for a night). To avoid losing the
revenue from this perishable capacity, capacity-constrained services are motivated to pre-sell the
inventory when possible by using reservations and giving discounts to avoid lost sales. For example,
travelers find that the published room rates (rack rates) for underutilized hotels are quickly abandoned
if the guest requests a discount (e.g., government or AARP). The exceptions to businesses that
practice this strategy are the budget motels that fill all of their rooms each night. Yield management
has alerted customers to the perishable nature of capacity-constrained services and this knowledge
destroys the myth of fixed prices for many such services and leads to price negotiation for all services.
6. Go to http://en.wikipedia.org/wiki/Yield_Management and discuss the ethical issues associated
with yield management.
Yield management is a form of price discrimination, and thus can be seen as unfair. For example,
why should two customers receive the identical service (coach seat on an airline flight) and pay
different prices? However, demand-responsive pricing in which last minute purchases are more
expensive than purchases far in advance is considered by most customers as reasonable.

11-4
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

INTERACTIVE CLASS EXERCISE


Watch the PowerPoint presentation concerning the overbooking experience at the Doubletree Hotel
in Houston, Texas. How could this situation been handled differently?

Clearly, the night desk clerk was not trained properly to handle such a late arrival for a reserved
room. The desk clerk should have apologized for the problem and immediately secured alternative
accommodations and transportation, if necessary. Note the guests contributed to the situation by not
notifying the hotel in advance of their late arrival. Figure 6.13, the service recovery framework, from
the Service Quality chapter can be displayed and used as a focus of the discussion.

EXERCISES
11.1 (a)

Day Capacity Walk-ins Appointments


Mon. 75 50 25
Tue. 75 30 45
Wed. 75 40 35
Th. 75 35 40
Fri. 75 40 35

(b) Appointments should not be scheduled at their maximum level because the walk-ins are only
expected values. There will be some variation both in time of arrival and in the actual demand. We
learned from queuing theory that capacity to serve must always exceed demand. Thus, some slack
time should be scheduled into the system in order to avoid excessive patient waiting.
(c) In order to avoid physician waiting, a few appointments should be made at the very beginning of
the day. The majority of appointments should be made in the late afternoon to avoid excessive waits
for walk-in patients.
11.2

Now Cu  $100 and C o remains $100

Cu 100
P(d  x)   0.50
Cu  C o 100  100
From Table 11.6 in the textbook, a strategy of overbooking three rooms guarantees that 48
percent of the no-shows will be covered. This represents a change from the previous policy of
overbooking two rooms.

11-5
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

11.3

No Shows Frequency Probability P( d  x )


0 6 .30 0
1 5 .25 .30
2 4 .20 .55
3 3 .15 .75
4 2 .10 .90
If overbook by 1, then P(d  x ) must be at least .30 and less than .55.
Cu 20
P(d  x )    0.30
Cu  Co 20  Co

Solving for Co : Co 
20  0.30(20)  $46.67 (maximum overbooking opportunity loss)
0.30

11.4

Daily
Demand Frequency Probability P( d  x )

10 1 .05 0
11 1 .05 .05
12 2 .10 .10
13 2 .10 .20
14 2 .10 .30
15 3 .15 .40
16 3 .15 .55
17 2 .10 .70
18 2 .10 .80
19 1 .05 .90
20 1 .05 .95
Cu (30  10) 20
P(d  x )     .67
Cu  Co (30  10)  10 30
Crazy Joe should lease 16 canoes.

11-6
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

11.5

Using a payoff matrix approach:

Passengers Overbooked
No
Shows P(NS) 0 1 2 3 4
0 .30 180* 130 80 30 – 20
1 .25 100 180 130 80 30
2 .20 20 100 180 130 80
3 .15 – 60 20 100 180 130
4 .10 –140 – 60 20 100 180
Expected Profit: $ 60 $101 $109.50 $ 92 $ 55
* Cell value = 6  $80  $300  $180
Thus, overbook 2 passengers (i.e., take 8 reservations) and expect a profit of $109.50 per flight.

Using the critical fractile model:

Cu 80
P(d  x )    . 61
Cu  Co 80  50
Where:

Cu = Cost of underestimating no-shows (i.e., there are more no-shows than expected and
the airline must fly empty seats)

Co = Cost of overestimating no-shows (i.e., there are fewer no-shows than expected and
the passengers who cannot be accommodated get free lift tickets)

11.6 Emergency Room Nurse Staffing

Objective Function:
Minimize x1  x2  x3  x4  x5  x6  x7
Constraints:
x1 x2 x3 x4 x5 x6 x7 RHS
Sun. 0 1 1 1 1 1 0  3
Mon. 0 0 1 1 1 1 1  6
Tue. 1 0 0 1 1 1 1  5
Wed. 1 1 0 0 1 1 1  6
Th. 1 1 1 0 0 1 1  6
Fri. 1 1 1 1 0 0 1  6
Sat. 1 1 1 1 1 0 0  5

11-7
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

For: i  1 to 7, xi  0 and integer

Where:

xi = Number of nurses assigned to a shift that has two consecutive days off beginning
with day i (i = 1 for Sunday, 2 for Monday, etc.)

Solution Summary for Emergency Room


Objective Objective
Variable Solution Coefficient Variable Solution Coefficient
x1 1 1 x5 2 1
x2 1 1 x6 0 1
x3 1 1 x7 3 1
x4 0 1
Minimized OBJ = 8 Iteration = 7 Elapsed CPU seconds = .109375

Weekly Staff Schedule for Emergency Room


Day x1 x2 x3 x4 x5 x6 x7 Req. Staff Excess
Sun. - 1 1 0 2 0 - 3 4 1
Mon. - - 1 0 2 0 3 6 6 0
Tue. 1 - - 0 2 0 3 5 6 1
Wed. 1 1 - - 2 0 3 6 7 1
Th. 1 1 1 - - 0 3 6 6 0
Fri. 1 1 1 0 - - 3 6 6 0
Sat. 1 1 1 0 2 - - 5 5 0

11.7 Sheriff Patrol Problem

Objective Function:
Minimize x1  x2  x3  x4  x5  x6  x7
Constraints:
x1 x2 x3 x4 x5 x6 x7 RHS
Sun. 0 1 1 1 1 1 0  6
Mon. 0 0 1 1 1 1 1  4
Tue. 1 0 0 1 1 1 1  4
Wed. 1 1 0 0 1 1 1  4
Th. 1 1 1 0 0 1 1  5
Fri. 1 1 1 1 0 0 1  5
Sat. 1 1 1 1 1 0 0  6

11-8
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

For: i  1 to 7, xi  0 and integer

Where:

xi = Number of deputies assigned to a shift that has two consecutive days off beginning with
day i (i = 1 for Sunday, 2 for Monday, etc.)

Summarized Report for Sheriff Patrol


Opportunity Objective Minimum Maximum
Number Variable Solution Cost Coefficient Obj. Coeff. Obj. Coeff.
1 x1 +1.0000000 0 +1.0000000 0 +1.5000000
2 x2 +2.0000000 0 +1.0000000 + .66666663 +1.0000000
3 x3 +1.0000000 0 +1.0000000 + 1.0000000 +1.5000000
4 x4 +2.0000000 0 +1.0000000 + .66666663 +1.0000000
5 x5 0 0 +1.0000000 + 1.0000000 +1.5000000
6 x6 +1.0000000 0 +1.0000000 0 +1.0000000
7 x7 0 0 +1.0000000 + 1.0000000 + Infinity
Minimized OBJ = 8 Iteration = 7 Elapsed CPU seconds = .109375

Weekly Staff Schedule for Summer Months


Day x1 x2 x3 x4 x5 x6 x7 Req. Staff Excess
Sun. – 2 1 2 0 1 – 6 6 0
Mon. – – 1 2 0 1 0 4 4 0
Tue. 1 – – 2 0 1 0 4 4 0
Wed. 1 2 – – 0 1 0 4 4 0
Th. 1 2 1 – – 1 0 5 5 0
Fri. 1 2 1 2 – – 0 5 6 1
Sat. 1 2 1 2 0 – – 6 6 0

11-9
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

11.8

F = $79

D = $59

p = 0.8

 = 75

 = 15

( F  D) (79  59)
P(d  x )    0.316
p( F ) 0.8(79)
Full-fare seats reserved =  +z = 75 + (-.48)(15) = 67

11.9 (a)

Model Rental Discount % Discount Daily Standard Fleet


Car Rate (F) Rate (D) Seekers (p) Demand () Deviation () Size
Compact $30 $20 80 50 15 60
Midsize $45 $30 60 30 10 30

30  20
P(d compact  xcompact )   . 417  xcompact  50  (. 21)(15)  46
30(.8)

45  30
P(d midsize  xmidsize )   .555  xmidsize  30  (.14)(10)  31
45(. 6)
(b)

Because the current fleet of midsize cars is smaller than the optimal number of midsize cars that
should be reserved for full-price paying customers, an expansion of the midsize fleet should be
considered.

11-10
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

CASE: RIVER CITY NATIONAL BANK


Assignment
As Ms. Chen's top aide, you are assigned the task of providing an analysis of the situation and
recommending a solution. This is your opportunity to serve your company and community as well
as to make yourself "look good" and earn points toward your raise and promotion.
Gary Miller is faced with the problem of fluctuating demand at the recently constructed drive-in
facility. The customer complaints of excessive waiting occur for the most part on Fridays. Little
congestion is experienced during the remainder of the week. This represents the classical service
problem of managing capacity. The problem can be addressed from two perspectives: managing both
demand and supply.
Altering demand:
The bank customers could be encouraged to use the facility on off-peak days, thus avoiding delays in
service. The customers can be informed of the peak times with notices in their monthly statements or
signs outside the remote facility. A further inducement to try the facility on an off-peak day could be
accomplished by giving those customers some small gift such as a pocket calendar or pen. A more
ambitious incentive would be an offer of free or reduced monthly service charges for customers who
bank regularly on off-peak days. Cashing payroll checks represents a major transaction on Fridays.
Customers should be informed of the convenient direct deposit option where employers can deposit
paychecks directly into employee bank accounts. The bank might wish to approach large employers
in the community to promote the direct deposit idea. Finally, customers might need encouragement to
be prepared to conduct their transactions as soon as they reach the tellers' windows. A preparation
stand containing deposit slips and pens could be placed just before the entry to the drive-in station so
customers could do their paperwork while waiting for the customer currently being served.
Controlling capacity:
With the use of part-time tellers and proper shift scheduling, the bank could match service capacity
with the fluctuations in daily demand better. Because there are no problems on Saturdays, we will
consider only weekdays. Our analysis begins with an assessment of teller availability. Tellers
available for the remote drive-in include 2 full-time tellers who work the morning shift only (7 a.m. to
2 p.m.), Monday through Friday, and 4 part-time tellers who work various combinations of morning
shifts (10 a.m. to 2 p.m.) or afternoon shifts (2 p.m. to 7 p.m.). Each part-time teller works 20 hours
per week, excluding Saturdays. This yields a total of 80 part-time teller-hours per week to supplement
the 70 full-time hours for the two full-time tellers. The problem involves allocating these 150 teller
hours to meet the daily demand levels for the week.
Table 1 is developed from the data provided in Table 11.10 of the case in the textbook. This
table summarizes the daily transaction demand by calculating the mean and standard deviation for
each day by morning and afternoon shift. The last column of the table gives the 95 percent demand
level assuming normal distribution. This demand level divided by teller capacity represents the
percent of utilization. From a queuing standpoint, this percent utilization must be less than 100
percent and probably not more than 80 percent to avoid excessive customer-waiting times.

11-11
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Table 1
Statistical Analysis of Transactions by Day of Week and Shift
Mean Standard 95 Percent
Number Deviation Demand Level
Monday: Morning 168.75 14.20 192.18
Afternoon 133.75 26.42 177.34
Tuesday: Morning 131.75 20.73 165.95
Afternoon 97.00 27.13 141.76
Wednesday: Morning 147.80 43.87 220.19
Afternoon 138.80 47.84 217.74
Thursday: Morning 140.80 25.74 183.27
Afternoon 125.40 30.07 175.02
Friday: Morning 227.80 20.63 261.84
Afternoon 232.00 39.11 296.53

In order to calculate teller capacity it is assumed that each teller can service two lanes and that
all eight lanes can be used when needed. The number of transactions per lane that are possible is
based on a typical transaction, which takes approximately five minutes per customer when we account
for all the delays such as moving the car into position, filling out forms, counting money, and pulling
away. Thus, each open lane can accommodate 12 transactions per hour. For the morning shift (7 a.m.
– 2 p.m.), one teller can accommodate 168 transactions, a teller working during the midday period (l0
a.m. – 2 p.m.) can process 96 transactions, and one afternoon (2 p.m. – 7 p.m.) teller can handle 120
transactions. Because of the prevailing policy, the two full-time tellers are scheduled to cover the
morning shift each day. The remaining 80 hours of part-time teller capacity is allocated to the midday
and afternoon to create a schedule that has a reasonably consistent percent utilization across the
weekdays. Table 2 shows the proposed teller schedule. Note that 81 part-time teller hours have been
scheduled. In Table 3 the percent utilization for each period is calculated. It is interesting to note that
now customers might perceive Monday and Thursday afternoons as being busier than Friday.

11-12
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Table 2
Teller Schedule
Morning Afternoon
Monday: two full-time 7-2 two part-time 2-7
one part-time 10-2

Tuesday: two full-time 7-2 two part-time 2-7

Wednesday: two full-time 7-2 three part-time 2-7


one part-time 10-2

Thursday: two full-time 7-2 two part-time 2-7

Friday: two full-time 7-2 four part-time 2-7


two part-time 10-2

Total teller hours per week: 70 full-time and 81 part-time

11-13
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Table 3
Transaction Capabilities of Tellers as Scheduled in Table 2
Lanes Total Percent
Open 1 Capacity2 Utilization3
Monday: Morning 5.14* 432 44
Afternoon 4.00 240 74

Tuesday: Morning 4.00 336 49


Afternoon 4.00 240 59

Wednesday: Morning 5.14* 432 51


Afternoon 6.00 360 60

Thursday: Morning 4.00 336 54


Afternoon 4.00 240 73

Friday: Morning 6.29* 528 50


Afternoon 8.00 480 62
1 Lanes Open assumes each teller serves two lanes.

2 Total Capacity is calculated as follows:

(lanes open)  (hours/shift)  (12 transactions/hour/lane)


3 Percent Utilization is calculated as follows:

95 percent demand level/total capacity


*Calculated using a weighted average, because the two full-time tellers work from 7 to 2 and
the part-time teller(s) work(s) from 10 to 2.
CASE: GATEWAY INTERNATIONAL AIRPORT
Questions
1. Assume that you are the assistant to the manager for operations at the FAA. Use the techniques of
workshift scheduling to develop an analysis of the total workforce requirements and days-off
schedule.
The most restrictive problem assumes the following constraints:
 Each controller must be allowed two consecutive days off.
 All controllers must work five 8-hour days.
 All controllers on a shift begin work at the same time (i.e., no starting times may overlap).
 The shifts run from 00:01 to 08:00, 08:01 to 16:00, and 16:01 to 24:00.
 The ratio of take-offs and landings to the number of controllers on duty cannot exceed 16.
Shown below is the number of operations and controller requirements for each hour of each day.

11-14
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Controller Requirements
Hourly Demand Adjusted for Daily Variation Controllers Required (Integer Value of Demand  16)
Hour S M T W T F S S M T W T F S
0-1 6.4 8.64 8.32 8. 8.32 8.96 7.36 1 1 1 1 1 1 1
1-2 4. 5.4 5.2 5. 5.2 5.6 4.6 1 1 1 1 1 1 1
2-3 3.2 4.32 4.16 4. 4.16 4.48 3.68 1 1 1 1 1 1 1
3-4 1.6 2.16 2.08 2. 2.08 2.24 1.84 1 1 1 1 1 1 1
4-5 2.4 3.24 3.12 3. 3.12 3.36 2.76 1 1 1 1 1 1 1
5-6 7.2 9.72 9.36 9. 9.36 10.08 8.28 1 1 1 1 1 1 1
6-7 9.6 12.96 12.48 12. 12.48 13.44 11.04 1 1 1 1 1 1 1
7-8 16. 21.6 20.8 20. 20.8 22.4 18.4 1 2 2 2 2 2 2
8-9 20. 27. 26. 25. 26. 28. 23. 2 2 2 2 2 2 2
9-10 26.4 35.64 34.32 33. 34.32 36.96 30.36 2 3 3 3 3 3 2
10-11 32.8 44.28 42.64 41. 42.64 45.92 37.72 3 3 3 3 3 3 3
11-12 35.2 47.52 45.76 44. 45.76 49.28 40.48 3 3 3 3 3 4 3
12-13 27.2 36.72 35.36 34. 35.36 38.08 31.28 2 3 3 3 3 3 2
13-14 30.4 41.04 39.52 38. 39.52 42.56 34.96 2 3 3 3 3 3 3
14-15 32. 43.2 41.6 40. 41.6 44.8 36.8 2 3 3 3 3 3 3
15-16 32.8 44.28 42.64 41. 42.64 45.92 37.72 3 3 3 3 3 3 3
16-17 33.6 45.36 43.68 42. 43.68 47.04 38.64 3 3 3 3 3 3 3
17-18 40.8 55.08 53.04 51. 53.04 57.12 46.92 3 4 4 4 4 4 3
18-19 44. 59.4 57.2 55. 57.2 61.6 50.6 3 4 4 4 4 4 4
19-20 50.4 68.04 65.52 63. 65.52 70.56 57.96 4 5 5 4 5 5 4
20-21 28. 37.8 36.4 35. 36.4 39.2 32.2 2 3 3 3 3 3 3
21-22 19.2 25.92 24.96 24. 24.96 26.88 22.08 2 2 2 2 2 2 2
22-23 13.6 18.36 17.68 17. 17.68 19.04 15.64 1 2 2 2 2 2 1
23-24 12. 16.2 15.6 15. 15.6 16.8 13.8 1 2 1 1 1 2 1

The minimum number of controllers needed for the basic problem is shown below:
Shift S M T W T F S
00-08 (night) 1 2 2 2 2 2 2
08-16 (day) 3 3 3 3 3 4 3
16-24 (evening) 4 5 5 4 5 5 4

The size of the workforce that is needed to meet the minimum requirements considering the
constraints of this problem is determined using the following integer linear programming (ILP) model:

11-15
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Let x j = Number of ATCs assigned to work schedule j

Minimize: x1  x2  x3  x4  x5  x6  x7
Right Hand Sides
Shift 1 Shift 2 Shift 3
8 a.m. - 4 p.m. 4 p.m. - midnight midnight - 8 a.m.
Subject to
Sun. x2  x3  x4  x5  x6 3 4 1
Mon. x3  x4  x5  x6  x7 3 5 2
Tue. x1  x4  x5  x6  x7 3 5 2
Wed. x1  x2  x5  x6  x7 3 4 2
Th. x1  x2  x3  x6  x7 3 5 2
Fri. x1  x2  x3  x4  x7 4 5 2
Sat. x1  x2  x3  x4  x5 3 4 2

Summarized Results for GIA (8am–4pm)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +1.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 +1.0000000 +1.0000000 6 x6 +1.0000000 +1.0000000
3 x3 0 +1.0000000 7 x7 0 +1.0000000
4 x4 +2.0000000 +1.0000000
Minimized OBJ = 5 Iteration = 11 Elapsed CPU seconds = 10.38281

Summarized Results for GIA (4pm–MN)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +2.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 0 +1.0000000 6 x6 +2.0000000 +1.0000000
3 x3 +1.0000000 +1.0000000 7 x7 0 +1.0000000
4 x4 +2.0000000 +1.0000000
Minimized OBJ = 7 Iteration = 9 Elapsed CPU seconds = 8.953125

Summarized Results for GIA (MN–8am)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +1.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 0 +1.0000000 6 x6 +1.0000000 +1.0000000
3 x3 0 +1.0000000 7 x7 0 +1.0000000
4 x4 +1.0000000 +1.0000000
Minimized OBJ = 3 Iteration = 3 Elapsed CPU seconds = 3.234375

11-16
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Work Schedules (W = workday, X = day off)


8 a.m.-4 p.m. Shift: S M T W T F S
Jean X X W W W W W
Laura W X X W W W W
Patty W W W X X W W
David M. W W W X X W W
Chon W W W W W X X

Staffing 4 3 4 3 3 4 4
Requirements 3 3 3 3 3 4 3
Excess 1 0 1 0 0 0 1

4 p.m.-midnight Shift:
Mario X X W W W W W
Yao (Joe) X X W W W W W
Shan Ling W W X X W W W
Kathy W W W X X W W
Jeff W W W X X W W
Scott W W W W W X X
John W W W W W X X

Staffing 5 5 6 4 5 5 5
Requirements 4 5 5 4 5 5 4
Excess 1 0 1 0 0 0 1

Midnight-8 a.m. Shift:


Lynda X X W W W W W
Claudia W W W X X W W
Ron W W W W W X X

Staffing 2 2 3 2 2 2 2
Requirements 1 2 2 2 2 2 2
Excess 1 0 1 0 0 0 0

Total Controllers Required: 15

If each shift is started one hour earlier, the workload will be smoothed enough to eliminate one
controller's position. The new shifts begin at 7 a.m. (day), 3 p.m. (evening), and 11 p.m. (night) and
yield the following ILP model:

11-17
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Let x j = Number of ATCs assigned to work schedule j

Minimize: x1  x2  x3  x4  x5  x6  x7
Right Hand Sides
Shift 1 Shift 2 Shift 3
7 a.m. - 3 p.m. 3 p.m. - 11 p.m. 11 p.m. - 7 a.m.
Subject to:
Sun. x2  x3  x4  x5  x6 3 4 1
Mon. x3  x4  x5  x6  x7 3 5 2
Tue. x1  x4  x5  x6  x7 3 5 1
Wed. x1  x2  x5  x6  x7 3 4 1
Th. x1  x2  x3  x6  x7 3 5 1
Fri. x1  x2  x3  x4  x7 4 5 2
Sat. x1  x2  x3  x4  x5 3 4 1

Summarized Results for GIA (7am–3pm)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +1.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 +1.0000000 +1.0000000 6 x6 +1.0000000 +1.0000000
3 x3 0 +1.0000000 7 x7 0 +1.0000000
4 x4 +2.0000000 +1.0000000
Minimized OBJ = 5 Iteration = 11 Elapsed CPU seconds = 10.32813

Summarized Results for GIA (3pm–11pm)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +2.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 0 +1.0000000 6 x6 +2.0000000 +1.0000000
3 x3 +1.0000000 +1.0000000 7 x7 0 +1.0000000
4 x4 +2.0000000 +1.0000000
Minimized OBJ = 7 Iteration = 9 Elapsed CPU seconds = 9.007813

Summarized Results for GIA (11pm–7am)


Variables Obj. Fnctn. Variables Obj. Fnctn.
No. Names Solution Coefficient No. Names Solution Coefficient
1 x1 +1.0000000 +1.0000000 5 x5 0 +1.0000000
2 x2 0 +1.0000000 6 x6 +1.0000000 +1.0000000
3 x3 0 +1.0000000 7 x7 0 +1.0000000
4 x4 +1.0000000 +1.0000000
Minimized OBJ = 2 Iteration = 1 Elapsed CPU seconds = 1.320313

11-18
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Work Schedules (W = workday, X = day off)


7 a.m.-3 p.m. Shift: S M T W T F S
Jean X X W W W W W
Laura W X X W W W W
Patty W W W X X W W
David M. W W W X X W W
Chon W W W W W X X

Staffing 4 3 4 3 3 4 4
Requirements 3 3 3 3 3 4 3
Excess 1 0 1 0 0 0 1

3 p.m.-11 p.m. Shift:


Mario X X W W W W W
Yao (Joe) X X W W W W W
Shan Ling W W X X W W W
Kathy W W W X X W W
Jeff W W W X X W W
Scott W W W W W X X
John W W W W W X X

Staffing 5 5 6 4 5 5 5
Requirements 4 5 5 4 5 5 4
Excess 1 0 1 0 0 0 1

11 p.m.-7 a.m. Shift:


Lynda W W W X X W W
Ron X W W W W W X

Staffing 1 2 2 1 1 2 1
Requirements 1 2 1 1 1 2 1
Excess 0 0 1 0 0 0 0

Total Controllers Required: 14

2. On the basis of your primary analysis, discuss the potential implications for workforce
requirements and days-off scheduling if assumptions (a) and (b) above are relaxed so that analysis
can be based on the hourly demand without the constraints of a preset number of shifts and no
overlapping of shifts. In other words, discuss the effects of analyzing hourly demand requirements on
the basis of each ATC position essentially having its own shift, which can overlap with any other ATC
shift to meet that demand.

11-19
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Relaxing the shift time constraint:


As shown above, beginning each shift one hour earlier smoothed the workload enough to eliminate
one ATC position. Flexibility in shift times allows a better match between service requirements and
employee scheduling. The most efficient shift times will minimize the number of controllers needed,
subject to the operations/controllers ratio.
Relaxing the shift demand with no overlap of shifts constraint:
Relaxing the shift profile of demand requirements and allowing shifts to overlap permit better
utilization of the workforce, because the service requirements and workforce capacity can be matched
better. The operations/controller ratio could be met with a smaller workforce. However, a tradeoff
exists between efficient workforce utilization and scheduling difficulties. In this case, it will become
easier to accommodate the general requirements of the problem, but the effort involved in scheduling
the workshifts will increase.
Relaxing other constraints:
Relaxing the days-off constraint would equalize the workload and facilitate the scheduling task.
Using part-time employees or a split-shift schedule can reduce workforce requirements and allow the
work schedule to match demand better. However, implementing these two strategies might be
inadvisable because they might have an adverse effect on employee morale.
3. Do you feel that this would result in a larger or smaller degree of difficulty in meeting the four
general constraints? Why?
Relaxing the shift time constraint seems an attractive option. This action improves the match between
service requirements and employee schedules and reduces the number of employees by one. Except
for the sensitive issue of eliminating jobs, this option does not have major drawbacks.
As noted in question 2, relaxing the shift demand and no-shift overlap constraint offers some
benefits but the implementation of this policy requires considerable expertise on the part of the
scheduler.
Using part-time employees and split-shift scheduling, on the other hand, could make
scheduling easier, but such a method might have adverse effects on employee morale. Moreover,
these alternatives might also have implications for maintaining a high level of ATC skills.
4. What additional suggestions could you make to the manager of operations to minimize the
workforce requirements level and days-off scheduling difficulty?
 Allow workers to choose if they want consecutive days or non-consecutive days off.
 Allow overlapping shift times. This will equalize workloads of all of the controllers, allow
smooth shift changes, and create a better match between workload capacity and requirements.
 Decrease the length of workshifts or allow controllers to work split-shifts. For example, if a
controller were to work from 5 to 8 p.m. on Wednesday, the airport would save more than one-
half of a day's ATC pay and still meet FAA requirements.
 Consider allowing controllers to work more than five days per week (within the bounds of safety)
or fewer than five days per week (within the bounds of maintaining skills).

11-20
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

 Attempt to alter demand. GIA could offer airlines incentives, such as lower take-off/landing fees
during off-peak times, to alter timing of their operations.

CASE: THE YIELD MANAGEMENT ANALYST


[This material is reprinted with permission of Kevin Baker and Robert Freund.]
YIELD MANAGEMENT GAME TALLY SHEET
Booking ID # Passengers # Passengers Carried Revenue
A (example) 100 Given in Phase III Phase III

SUBTOTALS
Passengers Gross Revenue
Oversale Adjustment (if over 100)

Spoilage Adjustment (if under 100)

NET REVENUE (Gross Revenue – Adjustment)

Airplane Capacity: 100 seats

11-21
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Historical Market Information:


Average no-show, misconnect, and cancellation rate: 20 percent
Average revenue per passenger: $400

Spoilage Penalty: $200 per empty seat


Oversale Penalty:
1-5 seats $200 per passenger
6-9 $500 per passenger
10-15 $800 per passenger
16+ $1,000 per passenger

Example Oversale Penalty (108 passengers): (5)($200)+(3)($500)=$2500

After reading the game instructions presented in the textbook and explaining the tally sheet, students
should be ready to begin the game. The game begins with revealing each of the demand opportunities
in Phase I in turn, one at a time, beginning with opportunity (A) and proceeding to the next
opportunity on the list only after everyone has made his and her decision. This demand revelation
process is accomplished using the PowerPoint presentation available on the McGraw-Hill/Irwin Web
site (http://www.mhhe.com/support). Students should make their decisions in ink, because once a
passenger group has been accepted, it is final. Also, students cannot return to an earlier passenger
group that has been rejected previously. Thus, as each group is revealed, the students must make a
non-reversible YES or NO decision to accept or reject that opportunity and be ready to move on
without knowledge of the future opportunities. Students record the group ID and number of
passengers in the first two columns of the tally sheet, keeping in mind a running total of seats sold in
order to hold some in reserve for Phase II, but also realizing that not all passengers will show up.
Phase I: Passenger Demand Received Outside of 13 days Prior to Departure

(A) 100 Sales Representatives - $275 per passenger


(B) 60 Professors - $200 per passenger
(C) 25 MBAs - $150 per passenger
(D) 5 Person Family - $300 per passenger
(E) 45 Conventioneers - $180 per passenger
(F) 35 Conventioneers - $300 per passenger
(G) 70 Sportswriters - $350 per passenger
(H) 20 Person High School Basketball Team - $300 per passenger
(I) 40 High School Band Members - $275 per passenger
(J) 80 Government Officials - $390 per passenger
(K) 25 Marketing Managers - $450 per passenger
(L) 15 Golfers - $325 per passenger
(M) 10 Nurses for Convention - $275 per passenger
(N) 40 Republican Congressmen - $450 per passenger
(O) 45 Club Med Vacationers - $375 per passenger
(P) 20 Shouldice Hernia Patients - $300 per passenger

11-22
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Phase II: Passenger Demand Between 13 days


Prior to Departure and Day of Departure
(1) 5 Business People - $875 per passenger
(2) 1 CEO - $1000 per passenger
(3) 4 IBM Executives - $850 per passenger
(4) 2 Last-minute Persons - $1500 per passenger
(5) 10 Passengers from another airline‟s oversold flight - $1000 per passenger
(6) 8 Travel Agents - $875 per passenger
(7) 11 Sales Mangers - $900 per passenger
(8) 8 Passengers from another airline‟s oversold flight - $1300 per passenger
(9) 2 CEO‟s running up to the plane at departure - $1500 per passenger

The final stage of the game requires each student to complete the last two columns of the tally
sheet, given the actual passengers who show up for each of the demand groups they accepted (i.e.,
booking ID in column one). In Phase III the number of passengers who show up and the revenue
contribution are reported, again using the PowerPoint presentation.

11-23
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Phase III: Actual Passengers Who Show Up for Flight and Resulting Revenue
(A) 70 Sales Representatives Show Up – $19,250
(B) 50 Professors Show Up – $10,000
(C) 25 Reliable MBAs Show Up – $3,750
(D) 5 Person Family Show Up – $1,500
(E) 35 Conventioneers Show Up – $6,300
(F) 15 Conventioneers Show Up – $4,500
(G) 70 Sportswriters Show Up – $24,500
(H) 20 Person High School Basketball Team Shows Up – $6,000
(I) 25 High School Band Members Show Up – $6,875
(J) 68 Government Officials Show Up – $26,520
(K) 25 Marketing Managers Show Up – $11,250
(L) 10 Golfers Show Up – $3,250
(M 8 Nurses Show Up – $2,200
(N) 30 Republican Congressmen Show Up – $13,500
(O) 30 Club Med Vacationers Show Up – $11,250
(P) 20 Shouldice Hernia Patients Show Up – $6,000
(1) 2 Business People Show Up – $1,750
(2) 1 CEO Shows Up – $1,000
(3) 3 IBM Executives Show Up – $2,550
(4) 2 Last-minute People Show Up – $3,000
(5) 5 Oversold Passengers Show Up – $5,000
(6) 8 Travel Agents Show Up – $7,000
(7) 11 Sales Managers Show Up – $9,900
(8) 4 Oversold Passengers Show Up – $5,200
(9) 2 CEO‟s Show Up – $3,000

After Phase III is revealed, the total number of passengers and gross revenue is known. Net
revenue is determined by adjusting for the costs of oversale or spoilage. Upon completion of the
game, the instructor records the total revenue for the flight for each student on the black board. We
find a range of $20,000 to $60,000 is not uncommon.

CASE: SEQUOIA AIRLINES


1. For the forecast period (i.e., July - December), determine the number of new trainees who must be
hired at the beginning of each month so that total personnel costs for the flight attendant staff and
training program are minimized. Formulate the problem as a linear programming model and solve.
Sequoia Airlines faces the problem of planning the optimal use of its resources during the next six
months. It must hire new employees and develop them through three stages of training, first as
trainees, then as junior flight attendants, and finally as full flight attendants. Moreover, Sequoia must
meet staff requirements for all flights during the training period. This problem is analyzed using total
personnel cost as the selection criterion.

11-24
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Other resources in addition to the available personnel create constraints that must be considered when
hiring and training new employees. We will solve the problem by using linear programming and then
follow up with an analysis of the objective function and the constraints and assumptions on which
they are based. We also will offer an interpretation of the results.
Because cost is the only criterion we will use to solve this problem, the objective function is
straightforward. One assumption is noteworthy: we assume that once an experienced attendant
becomes an instructor, he or she continues to receive the instructor's salary even when there is a
surplus of instructors who act as attendants. The objective function, therefore, becomes the sum of
the number of employees in each class during each period multiplied by the appropriate wage.
Minimize:

Z  750T1  T2  T3  T4  T5  T6 

 1050 J 1  J 2  J 3  J 4  J 5  J 6 

 1400F1  F2  F3  F4  F5  F6 

 1500I 1  I 2  I 3  I 4  I 5  I 6 
We also must deal with several resource constraints. First, we must meet the requirement for hours of
in-flight attendant service. We assume that a surplus instructor works the same number of hours as an
experienced attendant does.

140 J1  125F1  125S1  14000

140 J2  125F2  125S2  16000

140 J3  125F3  125S3  13000

140 J4  125F4  125S4  12000

140 J5  125F5  125S5  18000

140 J6  125F6  125S6  20000

11-25
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Next, we must satisfy the constraint that requires the total number of hours for junior attendants to be
less than or equal to 25 percent of the total attendant hours in each month.

140 J i  0.25140 J i  125Fi  125S i  fori  1,2,3,4,5,6

Also, there must be one instructor for every five trainees.

I i  0.2Ti fori  1,2,3,4,5,6

The remaining constraints deal with the relationships between the status of employees during the
current and prior periods. In this case we make assumptions that pertain to the initial staffing levels
for the positions of junior attendant, full attendant, and instructor.

Ji  0.8Ti 1

Fi  0.95 Ji 1  0.92 Fi 1

Ii  0.95Ii 1  0.005Fi 1

Si  Ii  0.2Ti

For i  1,2,3,4,5,6
Except J1  8, F1  119, I1  6
We see a total of 42 constraints in the following computer model.

11-26
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Min 1050J1+ 1050J2+ 1050J3+ 1050J4+ 1050J5+ 1050J6+ 1400F1+ 1400F2+ …


Subject to
(1) 140J1+ 125F1+ 125S1 >= 14000
(2) 140J2+ 125F2+ 125S2 >= 16000
(3) 140J3+ 125F3+ 125S3 >= 13000
(4) 140J4+ 125F4+ 125S4 >= 12000
(5) 140J5+ 125F5+ 125S5 >= 18000
(6) 140J6+ 125F6+ 125S6 >= 20000
(7) 105Jl-31.25Fl-31.25S1 <= 0
(8) 105J2-31.25F2-31.25S2 <= 0
(9) 105J3-31.25F3-31.25S3 <= 0
(10) 105J4-31.25F4-31.25S4 <= 0
(11) 105J5-31.25F5-31.25S5 <= 0
(12) 105J6-31.25F6-31.25S6 <= 0
(13) 1I1-.2T1 >= 0
(14) 112-.2T2 >= 0
(15) 113-.2T3 >= 0
(16) 114-.2T4 >= 0
(17) 115-.2T5 >= 0
(18) 116-.2T6 >= 0
(19) 1J1 = 8
(20) 1F1 = 119
(21) 111 = 6
(22) -111+ 1Sl+ .2T1 = 0
(23) -112+ 1S2+ .2T2 = 0
(24) -113+ 1S3+ .2T3 = 0
(25) -114+ 1S4+ .2T4 = 0
(26) -115+ 1S5+ .2T5 = 0
(27) -116+ 1S6+ .iT6 = 0
(28) 1J2-.8T1 = 0
(29) 1J3-.8T2 = 0
(30) 1J4-.8T3 = 0
(31) 1J5-.8T4 = 0
(32) 1J6-.8T5 = 0
(33) -.95J1-.92F1+ 1F2 = 0
(34) -.95J2-.92F2+ 1F3 = 0
(35) -.95J3-.92F3+ 1F4 = 0
(36) -.95J4-.92F4+ 1F5 = 0
(37) -.95J5-.92F5+ 1F6 = 0
(38) -.005F1-.95I1+ 112 = 0
(39) -.005F2-.9512+ 113 = 0
(40) -.005F3-.9513+ 114 = 0
(41) -.005F4-.9514+ 115 = 0
(42) -.005F5-.9515+ 116 = 0

2. How would you deal with the noninteger results?


The real number value solution for this problem is shown in Table 1. However, it should be noted that
Sequoia could only hire whole people, not fractions of people. Ordinarily we would attempt to apply
integer programming to this problem, but, unfortunately, an integer solution is not possible owing to
the large number of constraints that have decimal parameters. Nevertheless, we can still obtain an
integer solution by assigning ranges to some of the decimal parameters. An example integer solution
is presented in Table 2.

11-27
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

TABLE 1 Sequoia Airlines Real Number Solution

Final Solution for SEQUOIA AIRLINES Page : 1


Variables Opportunity Variables Opportunity
No. Names Solution Cost No. Names Solution Cost
1 J1 +8.0000000 0 16 I4 +6.7955503 0
2 J2 +4.1294594 0 17 I5 +6.9693007 0
3 J3 0 0 18 I6 +7.1805620 0
4 J4 +18.374796 0 19 S1 +4.9676352 0
5 J5 +27.182201 0 20 S2 +6.2950001 0
6 J6 +21.434549 0 21 S3 +1.9719510 0
7 F1 +119.00001 0 22 S4 0 0
8 F2 +117.08000 0 23 S5 +1.6106634 0
9 F3 +111.63660 0 24 S6 +7.1805620 0
10 F4 +102.70567 0 25 T1 +5.1618242 0
11 F5 +111.94527 0 26 T2 0 +1686.3875
12 F6 +128.81274 0 27 T3 +22.968494 0
13 I1 +6.0000000 0 28 T4 +33.977749 0
14 I2 +6.2950001 0 29 T5 +26.793188 0
15 I3 +6.5656500 0 30 T6 0 +1221.3268
Minimized OBJ = 1177115 Iteration = 35 Elapsed CPU seconds = 16.85938

Final Solution for SEQUOIA AIRLINES Page : 2


Variables Opportunity Variables Opportunity
No. Names Solution Cost No. Names Solution Cost
31 S1 +2615.9539 0 46 S10 +1280.1985 0
32 A1 0 0 47 S11 +694.49170 0
33 S2 0 +20.099571 48 S12 +1999.1627 0
34 A2 0 –20.099571 49 S13 +4.9676352 0
35 S3 +1201.0670 0 50 S14 +6.2950001 0
36 A3 0 0 51 S15 +1.9719510 0
37 S4 +3410.6787 0 52 S16 0 +7367.7900
38 A4 0 0 53 S17 +1.6106634 0
39 S5 0 +20.861754 54 S18 +7.1805620 0
40 A5 0 –20.861754 55 T19 +5.1618242 –725.91003
41 S6 0 +18.853069 56 T2O 0 –1053.4916
42 A6 0 –18.853069 57 T21 0 +4704.0610
43 S7 +3033.9885 0 58 T22 0 0
44 S8 +3421.8755 0 59 T23 0 +2512.4463
45 S9 +3550.2668 0 60 T24 0 0
Minimized OBJ = 1177115 Iteration = 35 Elapsed CPU seconds = 16.85938

11-28
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

Final Solution for SEQUOIA AIRLINES Page : 3


Variables Opportunity Variables Opportunity
No. Names Solution Cost No. Names Solution Cost
61 A25 0 0 70 A34 0 –869.93689
62 A26 0 +2607.7190 71 A35 0 +534.34448
63 A27 0 +2356.6335 72 A36 0 +2092.1052
64 A28 0 +937.50000 73 A37 0 +956.63367
65 A29 0 –542.37274 74 A38 0 +6530.5908
66 A30 0 +937.50000 75 A39 0 +5808.5732
67 A31 0 +2779.4475 76 A40 0 +7693.2349
68 A32 0 +1589.4298 77 A41 0 +1921.5211
69 A33 0 +341.14731 78 A42 0 +856.63373
Minimized OBJ = 1177115 Iteration = 35 Elapsed CPU seconds = 16.85938

TABLE 2 Sequoia Airlines Integer Solution


J1 = 8 F1 = 119 I1 = 6 T1 = 0 S1 = 6
J2 = 0 F2 = 122 I2 = 6 T2 = 0 S2 = 6
J3 = 0 F3 = 104 I3 = 6 T3 = 0 S3 = 6
J4 = 0 F4 = 102 I4 = 8 T4 = 36 S4 = 0
J5 = 32 F5 = 102 I5 = 10 T5 = 15 S5 = 7
J6 = 13 F6 = 134 I6 = 12 T6 = 0 S6 = 12

Note that the integer solution is radically different from the continuous solution. This
difference is explained by the large ranges given to the constraint parameters. Which solution should
be used? Perhaps rounding-up the continuous solution and then adjusting for constraint requirements
would provide the best solution. In fact, this solution is preferable because the integer solution
implicitly assumes that management can control the number of employees who quit each month,
which clearly is not the case. In order to use the integer solution, we might first use forecasting
methods to estimate the number of employees who will quit and then use integer programming to plan
for hiring new employees.
3. Discuss how you would use the LP model to make your hiring decision for the next six months.
Following the hiring decisions for the current month (July in this case), the program can be run again
at a later date by dropping the July requirements and adding the January requirements. This "leap
frog" approach can be used to fine tune proposed hiring levels by extending the planning horizon for
another six months. The decision on future hiring (July through December in this case) is postponed
until a commitment must be made. At that time the decision can be based on a six-month projection
of requirements.

11-29
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

CHAPTER QUIZ QUESTIONS


True/False
1. The use of a ski-resort hotel for business conventions during the summer is an example of using
the complementary service strategy. (F)
2. Overbooking is a strategy that can be used to smooth demand. (F)
3. The strategy of segmenting demand to reduce variation makes use of the fact that demand for a
service seldom is derived from a homogeneous source. (T)
4. Differential pricing is an attempt to make peak usage periods unattractive by imposing a penalty
on the consumer for using the service during peak periods. (F)
5. Yield management is the process of allocating a fixed perishable resource to several market
segments in the most profitable manner. (T)
6. The work shift-scheduling problem is important in service organizations that face a cyclic demand
for their services. (T)
7. Work shift scheduling attempts to deal with the service utilization problem by controlling the
demand for the service and partitioning it so that utilization is uniform. (F)
8. A drawback to increased consumer participation is the fact that service quality is no longer
completely under the control of the provider of the service. (T)
9. When peaks of activity are persistent and predictable such as meal times for restaurants, off-duty
personnel can be placed on standby to supplement regular employees. (F)
10. Cross training of employees as a strategy to increase flexibility is feasible only when there exist
tasks, the demands for which peak at different times. (T)
11. Time perishability of service capacity is a challenge for service managers because customers
demand immediate service. (F)
12. Yield management is a pricing and capacity allocation system that was developed by American
Airlines. (T)
13. Yield management is a strategy that manages both demand and capacity. (T)
14. An example of segmenting demand is seen when movie theaters offer matinee prices before 6:00
p.m. (F)
15. Expected loss for an overbooking reservation strategy would be calculated by multiplying the loss
for each no-show possibility and its probability of occurrence, and then adding the products. (T)
16. The critical fractile is a cumulative probability of demand. (T)
17. Using part-time personnel at fast-food restaurants allows capacity to vary with demand. (T)
18. A disadvantage of yield management is that it cannot be implemented in real time. (F)

11-30
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

19. Some restaurants use tables and chairs instead of booths to create more flexible capacity. (T)
20. SABRE is the name for American Airlines‟ yield management system. (F)
21. Service capacity is defined in terms of an achievable level of output per unit time. (T)
22. Level demand and chase capacity are the two generic strategies for capacity management. (F)
Multiple Choice
1. The strategy of segmenting demand is feasible only when:
a. demand is not from a homogeneous source.*
b. demand is cyclic and predictable.
c. arrivals for service are random.
d. making appointments is impossible.
2. The purpose of differential pricing is to:
a. make peak period usage unattractive.
b. make off-peak usage attractive.*
c. charge customers according to their ability to pay.
d. adjust capacity to demand.
3. A good overbooking strategy should:
a. minimize the expected opportunity cost of idle service capacity.
b. balance the expected opportunity cost of idle service capacity and expected cost of turning
away customers who have reservations.*
c. minimize the expected cost of turning away reservations.
d. none of the above; services should try to avoid overbooking.
4. Bars that offer happy hours in the afternoon are using the strategy of:
a. creating adjustable capacity.
b. developing complementary services.
c. increasing customer participation.
d. promoting off-peak demand.*

11-31
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

5. Which of the following is not a strategy for managing capacity?


a. Developing complementary services*
b. Using part-time employees
c. Forecasting demand
d. Scheduling shifts
6. Which of the following strategies is inappropriate for managing capacity and demand?
a. Smooth customer demand by offering price incentives.
b. Scheduling staff to meet variations in forecasted customer demand.
c. Decrease customer participation in the service process.*
d. Promoting off-peak use of facilities.
7. A cruise ship has available a certain number of rooms in each of 10 categories of appointments
ranging from a two-bunk inside cabin to a suite with an outside patio and indoor sitting area. If a
passenger requests a sailing date in which the desired cabin is sold out, which of the following
actions would be considered the least viable alternative for the cruise line?
a. Offer an upgraded cabin at a reduced price.
b. Attempt to steer the passenger to an available date.
c. Overbook the cabin.*
d. Offer a downgraded cabin with special privileges such as a $100 certificate for use in the
casino.
8. What one of the following is not a characteristic of yield management?
a. Capacity is relatively fixed.
b. There is one homogeneous customer class.*
c. The service is considered a perishable inventory.
d. Demand fluctuates yet is somewhat predictable.
9. Which one of the following four steps in "daily workshift scheduling" is out of order?
a. Forecast demand.
b. Convert to operator requirements.
c. Assign operators to shifts.*
d. Schedule shifts.

11-32
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

10. Faced with variable demand and a perishable capacity, a service manager can smooth demand by:
a. using part-time help during peak hours.
b. scheduling workshifts to vary workforce needs according to demand.
c. increasing the customer self-service content of the service.
d. using reservations and appointments.*

11. Which one of the following is not a strategy to manage demand?


a. cross-training employees*
b. offering price incentives
c. developing reservation systems
d. partitioning demand
12. A health club offering a reduced rate membership for students to workout before 4:00 p.m. on
weekdays is:
a. promoting off-peak demand.*
b. partitioning demand.
c. using yield management.
d. offering price incentives.
13. Several approaches to demand management exist, but only _______ seeks to maximize revenue.
a. promoting off-peak demand
b. reservation systems
c. offering price incentives
d. yield management*
14. Which one of the following is not a characteristic of firms using yield management?
a. ability to segment their market
b. perishable inventory
c. variable capacity*
d. product sold in advance

11-33
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

15. Which one of the following is not an example of the differential pricing policy?
a. weekend and night rates for long-distance telephone calls
b. difference in hospital fees for walk-in and scheduled services*
c. peak-load pricing by utility companies
d. none of the above
16. In using the critical fractile criterion P(d<x) = Cu/(Cu + Co) for overbooking, the „d‟ refers to:
a. cost of overestimating demand.
b. cost of underestimating demand.
c. number of rooms overbooked.
d. number of no-shows based on past experience.*
17. For a chase demand strategy which of the following does not have a high trade-off?
a. Employee utilization
b. Labor-skill level*
c. Labor turnover
d. Supervision required
18. A restaurant that features special lunchtime combo meals is providing all but one of the following
benefits?
a. Promotes off-peak demand*
b. Increases customer satisfaction
c. Decreases service times
d. Segment demand

11-34
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

19. All but one of the following is an example of a yield management application:
a. RyerFirst
b. SABRE*
c. HIRO
d. Restaurant Catering Software
20. _________ variability is not one of the five sources of customer-induced variability.
a. Arrival
b. Capability
c. Effort
d. Demand*

21. ________ is not a strategy for reduction of customer-induced variability.

a. Adapt to customer skill levels*


b. Require reservations
c. Limit service breadth
d. Target customers based on capability

22. ________ is not a strategy for accommodating of customer-induced variability.

a. Provide generous staffing


b. Cross-train employees
c. Reward increased effort*
d. Do work for customers

23. A medical clinic has two doctors and each can treat 25 patients a day. The doctors see walk-in
patients whose arrival times cannot be controlled, and also patients who have made appointments.
Knowing the expected number of walk-ins per day, appointments are scheduled to utilize the
doctors fully. The following table gives the expected number of walk-ins for a particular week:
Day: Mon. Tue. Wed. Th. Fri.
Expected Walk-ins: 45 35 40 45 40
What is the total number of appointments that can be scheduled during this week?

a. 30
11-35
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.
Chapter 11 - Managing Capacity and Demand

b. 35
c. 40
d. 45*
24. ________ is not a characteristic appropriate for a yield management strategy.
a. Relatively fixed capacity
b. Steady demand*
c. Ability to segment markets
d. Products sold in advance
25. When yield management is implemented which one of the following does not result:
a. consumer surplus increases*
b. multiple prices are offered
c. capacity is more fully utilized
d. market for the service is segmented

11-36
© 2014 by McGraw-Hill Education. This is proprietary material solely for authorized instructor use. Not authorized for sale or distribution in
any manner. This document may not be copied, scanned, duplicated, forwarded, distributed, or posted on a website, in whole or part.

Potrebbero piacerti anche