Professional Documents
Culture Documents
Instructor Manual For Matching Supply With Demand 2e
Instructor Manual For Matching Supply With Demand 2e
2nd Edition
by Gerard Cachon and Christian Terwiesch
Instructor Manual
March 2008
This document describes our pedagogical strategy, summarizes the materials provided along with
the text to aid instructors (i.e., instructor materials), details our copyright policy regarding the
instructor materials, and provides a brief description of how we use the text in our own teaching
(e.g., cases, preparation questions, etc.)
Pedagogical strategy:
Our guiding philosophy with the design of this text is real operations real solutions: we
provide real company cases and real solutions to their operational challenges while presenting
the material in a manner that non-engineering student can understand. In fact, we also like the
expression rigor with relevance: we do not shy away from rigorous mathematical analysis but
our analysis is always focused on relevant operational problems.
Our primary target audience is students in MBA degree programs (daytime, weekend or
executive). Even though the models we present are quite rich (e.g., they allow for different
objective functions and different performance measure evaluations) we find that MBA students
are capable of mastering this material and they appreciate that the course has not been dumbed
down. In fact, we find that both quant jocks and poets enjoy the course. Because the text
emphasizes models and the qualitative insights from those models, we find that junior professors
are successful in the MBA classroom even with their limited experience. Finally, we suspect that
the text will work quite well with business or engineer undergraduate students.
The following is a list of our design features that make this text a useful teaching tool:
1. Multiple levels of detail. Each chapter walks students through a case analysis in great
detail. But we find that students sometimes want a quick how do I do X solution, so
we provide exhibits within the chapter that explain to students the steps need to do a
particular calculation. Furthermore, at the end of each chapter and at the end of the text
there are lists of key equations for when students remember the process to do the
calculation but cant remember the particular equation details.
2. Little mathematical notation. Students do not have the time or the desire to remember
the meaning of many different mathematical symbols or variables. So in many cases we
write out variable names so that there is no confusion, e.g., it is clear what Expected
Sales means.
3. Plenty of practice problems. Students learn by repetition. So we provide enough
practice problems to satisfy even the most eager student.
4. Presentation slides linked to the text. We find that the less quantitatively oriented
students appreciate our in class lectures that cover examples from the text, because then
they can read the text latter for additional explanations and reinforcement. Some of the
quantitatively strong students would prefer less redundancy, but when we consider the
needs of all students, the redundancy is a net winner.
5. Excel spreadsheets. We give students spreadsheets that perform the analysis of many of
the models. Students that are able to absorb this material quickly and need to minimize
their time on the course appreciate these tools.
6. Statistics Tutorial. Students do not always receive the statistics they need from their
statistics course. (For example, at our school the Poisson distribution is never mentioned
in the statistics core class.) So we provide a statistics tutorial in the appendix that gives
the students exactly the statistics they need for this text.
7. Advanced materials. Some students, the quant jocks, often want to know how
equations are derived. To satisfy their curiosity, we provide that supplemental
information in the Appendix.
Summary of Instructor Materials:
1. Lecture slides. (Power Point files). These are slides that we use in our course. Many
contain references to cases and problems in the text, but there is additional content as
well.
2. Solutions to all end-of-chapter problems in the text. (Word files).
3. Text figures. (Power Point). These files contain the figures that are used in the text (and
may not be contained in the lecture slides).
4. Case materials. We use a number of cases along with the text in our course and have
prepared materials associated with the cases, e.g., data students can use to answer
preparation questions, case analysis slides, etc.
5. Model solvers. We have developed several Excel spreadsheets that help students solve
several of the model discussed in the text.
6. Sample questions. We provide questions beyond those at the end of each chapter,
including both quantitative and qualitative questions. We have used these questions in
final exams and homework assignments. All of our homework and final exams are
multiple choice with no partial credit. We have had a very positive experience using this
testing format: students find the format to be fair, grading errors are minimized (in
particular, there is no need to train graders on how to offer partial credit), grading
ambiguities are minimized (students with short answer questions can be strategic in how
they answer even if they do not know the answer), grading time is very quick and
question development does not require much more time than typical qualitative or
quantitative questions.
7. Sample syllabi. These syllabi are from our courses that use the text and include course
policy, timetables, case preparation questions, etc.
Copyright policy:
All instructor materials that are not included in the actual text are copyrighted by the
authors, Gerard Cachon and Christian Terwiesch. Instructors that adopt the text as a
requirement for their course are free to use these materials and to modify them as they see
fit. All others are required to obtain explicit written permission from the authors.
Suggested course outline:
In our own teaching we use the text in 24 sessions (divided between two courses, each
with 12 sessions) and each session is 80 minutes of class time. The following are the
cases we use and the chapters that contain materials associated with those cases:
Chapter
2-4
2-4
5
6
7-8
7-8
9
10
11
12
12
13
14
Case
Toshiba: OME works (HBS 9696-059)
National Cranberry (HBS 9-688122)
Paul Downs
Executive Shirt (HBS 9-696-071)
Beau Ties (UVA-om-0836)
Manzana Insurance (HBS 9-692015)
Catapult exercise
Toyota (HBS 9-693-019)
Le Club Francais (Wharton case)
or, LL Bean (HBS 983-003)
Timbuk2 (Wharton case) or
National Bicycle (Wharton case)
Sport Obermeyer (HBS 9-695022)
Hewlett-Packard (Stanford)
15
16
16
Topic
Process analysis
Process analysis
Link between operations and finance
Process analysis (with batching)
Queuing
Queuing, process design
Quality
Lean manufacturing and the Toyota
production system
Newsvendor Model
Newsvendor, reactive capacity
Reactive capacity
Order upto model, delayed
differentiation
Risk pooling
Revenue management
Supply chain coordination, bullwhip
effect
Supply chain contracting
The following outlines the sequence in which we present the material. The first set of 12
sessions covers process analysis, quality and lean manufacturing. The second set covers
inventory and supply chain management.
Session #
1
2
3
4
5
6
7
8
9
10
11
12
Session
13
14
15
16
17
18
19
20
21
22
23
24
Introduction
Reading:
Session 2
The Toshiba case illustrates a classic assembly line operation. We use the case to
reinforce several of the key concepts and terms in process flow analysis.
Reading:
Text, Chapter 3
Case:
What are the key elements of Toshibas business strategy in notebook computers?
In what way do Omes operations support this strategy?
What is Toshiba doing to achieve high performance on cost, quality and
flexibility?
Assuming the assembly line prototype is implemented as shown in Exhibit 1,
calculate the following quantities:
- Process Capacity
- The maximum number of computers that can be produced in a 7.5 hour
shift.
- Direct Labor Content per notebook computer (i.e., the amount of time a
worker actually works on a computer while it is on the assembly line).
- Direct Labor Idle Time per notebook computer assembled (i.e., the amount
of time workers are idle per computer assembled).
- Inventory on the assembly line.
- Flow Time for a notebook computer.
Note that Station 9 is somewhat more complex than the others. Two facts are important:
(1) software loading does not require an operator (its like waiting for your computer to
start up) and (2) Station 9 occupies three spaces on the line. Because the conveyor belt
moves continuously, a given computer therefore spends three times as long in Station 9 as
in the other stations. The worker for Station 9 moves as needed among the three
computers within Station 9 to perform the tasks requiring an operator.
Session 3
The Operations Finance Link / Setup times
In the first part of this session we talk about the link between operations management and
finance. Two topics will be emphasized. First, we will talk about inventory turns and
other aspects of working capital management. Second, we will outline the link between
financial performance metrics and operational decisions using the case of a small
furniture company.
In the second part of the session, we discuss the impact of set-up times on capacity and
inventory levels.
Reading:
Text, Chapter 5
Session 4
Bottleneck Analysis: Continuous Process
The National Cranberry case is a classic and has become a point of reference for nearly
everyone who has attended business school. A common pitfall in analyzing the case is to
become mired in too much detail, so be careful to maintain the big picture while
addressing the questions. For the purposes of your analysis, you may make the following
assumptions:
The Flow Time of the National Cranberry process (starting after the holding bins)
is 1 hour (i.e., it takes 1 hour for a cranberry to flow through the plant).
During a high-volume period the dryer operator can start at whatever time you
choose, rather than 11 AM as shown in Figure E,
The amount of inventory other than in the bins is negligible.
Case:
Draw a Process Flow Diagram of the cranberry process beginning with Receiving
and ending with the Bailey Mills (i.e., ignore Sorting and Shipping at the end of
the process).
Compute the Capacity in barrels per hour of each process step.
Session 5
Consider a peak harvest day (18,000 barrels of berries unloaded with 70% of them
wet harvested). Assume that trucks arrive uniformly over a period of 12 hours.
Identify the Bottleneck of the process.
When would processing be completed on a peak day?
When would the last truck unload and how long would it have waited?
If you were Hugh Schaeffer, what changes would you make to improve
performance of the process? Estimate the magnitude of the costs and benefits of
these changes.
Process Design
Reading:
Session 6
Text, Chapter 4
In this class, we apply the ideas and concepts derived in production settings to the service
industry.
Case:
Case Preparation:
Session 7
Managing Variability: Waiting time problems
This lecture will introduce queuing formulas needed for the next several classes.
Reading:
Text, Chapter 7
Session 8
Managing Variability and Waiting Times
The Beau Ties case will allow us to apply the tools of variability management for the
analysis of a call center. Check out http://www.beautiesltd.com/ for more details about
the companys product line.
When answering the questions, please note the following: (a) use coefficients of variation
equal to 1 for both arrival and service processes (b) The numbers in Exhibit 5 reflect call
volumes for the corresponding time slots cumulated over a period of two months. E.g. the
18 calls in the Monday 7-8 time slot come from eight different Mondays in March and
April. Divide the numbers by 8 to get daily volumes!
Case:
Assuming that Kenersons has a target of an average wait of less than 1 minute,
develop a telephone staffing plan for December 4, 1995 assuming that the
distribution of phone calls throughout the day follows the hourly distribution in
Exhibit 5.
How will your December 4, 1995 plan change if the target is an average wait of
less than 30 seconds?
Compare the variable costs of the staffing plans from questions 1 and 2 with the
charges from AIDC.
How should Kenerson evaluate the decision to bring the telephone order-entry
system in-house?
For the questions above, we strongly encourage you to build a spreadsheet model in
Excel, rather than doing the analysis by hand.
Session 9
Reading:
Text, Chapter 8
How does the andon procedure work and what are its fundamental aspects? How
much does it cost to stop the line? What are the benefits of stopping the line?
What are the underlying causes of the problems facing Doug Friesen?
How, if at all, does the current routine for handling defective seats deviate from
the principles of the Toyota Production System?
As Doug Friesen, what would you do to address the seat problem? Where would
you focus your attention and solution efforts? What options exist? What would
you recommend? Why?
Text, Chapter 9
Exercise:
Introduction
This session will cover the course syllabus and introduce the newsvendor model.
Reading:
Session 14
This session explores the challenges of procurement for a fashion apparel catalog retailer.
Case:
Reading:
Session 15
What are the costs of having one bottle too few in inventory (underage cost)?
What are the costs of having one bottle too many in inventory (overage cost)? List
these costs qualitatively and attempt to attach numbers to them.
Assume the underage cost is 3 Euro and the overage cost is 1 Euro. How many
bottles would you order of a wine that is forecasted to sell 2000 bottles if your
objective is to maximize profits? How do these numbers change if you use the
cost numbers from your answer to Question 1?
Assume you would like to achieve a fill rate of 98%. How would the answer to
question 2 change?
How much of each wine listed in Exhibit 2 would you order? Be prepared to
explain and justify your decisions. (An excel file is provided that contains the data
in Exhibit 2.)
Reading:
What are some of the pros and cons of Timbuk2s Build your own channel (i.e., its
ecommerce channel)?
How should Timbuk2 go about deciding which options to offer customers in the
ecommerce channel? In other words, what general principles or analysis could be
used to deepen their understanding of the appropriate choices? You may want to
consider several of the options mentioned (an added handle, different color logos,
different size panels, etc.)
Estimate the cost of manufacturing a bag in San Francisco and the cost of producing a
bag in China.
Should Timbuk2 pursue the option of manufacturing in China? If so, what challenges
are they likely to face and what changes will they need to make?
Session 16
This session studies how one fashion apparel supplier used early sales information to
increase its supply flexibility.
Case:
Reading:
Session 17
How much of each style described in Exhibit 10 should be produced if you only
had one production run? What is Sport Obermeyers expected profit?
What do you think about Sport Obermeyers forecasting process? Contrast it with
LL Beans forecasting process.
Suppose you now have the opportunity to make two production runs. The first
must be decided before the Las Vegas show, and the second is decided after the
show. Lets say the production minimums are large (i.e., you essentially get to
produce a style either in the first production run or the second, but not both).
Furthermore, suppose the first production run must total at least 15,000 units
across all styles, but then there is no capacity restriction on the second production
run. In addition, the production cost per unit of a style does not depend on which
production run it is ordered. How many units of each style in Exhibit 10 should
Sport Obermeyer order in the first production run? How much is the mismatch tax
reduced by the second production run opportunity?
How can Sport Obermeyer improve upon their system to better match supply to
demand?
The Order Up-To Model
This session studies service levels and lead times in a supply chain. Unlike with the
newsvendor model, we now consider a supply chain that has demand over a long time
horizon, so multiple replenishments are possible.
Reading:
Session 18
Text, Chapter 13
Postponement
Postponement is a strategy to redesign a product and its supply chain to increase supply
flexibility.
Case:
What are the pros and cons of the following proposals mentioned in the A case: a
European factory, better forecasting, more inventory.
Session 19
Assess quantitatively the air freight option relative to current operations. Just
consider the products for the European market. Do not forget to consider pipeline
inventory (since HP owns the pipeline inventory from Vancouver to Europe). Use
the following assumptions:
- HP wants to minimize inventory while still achieving at least a 98% fill
rate.
- The lead time from Vancouver to Europe is 5 weeks by the current method
(ocean) but 1 week by air.
- HP orders and received inventory on a weekly basis.
- There are 4.33 weeks per month and demand is independent across time.
- The product sells for $450 and marginal production cost is $300.
- Inventory carrying costs are 24% per year.
- Shipping via sea (the current operation) costs $10 per printer, whereas
airfreight costs $25 per printer.
Evaluate quantitatively the proposal in the B case for the European market.
Will the B case proposal be effective in the other major markets, North America
and Asia?
Would you support the B case proposal? If so, why? If not, why?
Managing Risk in Operations
This session explores several operations strategies for reducing and hedging uncertainty.
Reading:
Session 20
Text, Chapter 14
Internet Distribution
Internet retailing requires less inventory and retail space than brick-and-mortar retailing.
But Internet retailing introduces additional costs. We shall compare these two models
from an operations perspective.
Case:
Session 21
Revenue Management
This session studies several revenue management tools to increase revenue in the
presence of fixed capacity and variable demand.
Reading:
Text, Chapter 15
Reading:
Is there any evidence that Barilla faces the bullwhip effect? If so, what causes of
the bullwhip effect are present?
Who resisted JITD and why? How would you respond to their concerns (i.e., how
would you modify the JITD proposal to make it more acceptable)?
Would you adopt JITD?
Reading:
Video Vault needs to decide how many copies of new movie to purchase. Video
Vault has constructed the following forecast of rentals for this movie.
Number of tapes
purchased
1
2
3
4
5
6
Expected total
number of rentals
60
100
120
132
137
139
For example, if they purchase 3 tapes, then they expect those three tapes to rent a
total of 120 times. Clearly, the more tapes purchased, the greater the total number
of rentals, but each additional tape generates fewer incremental rentals over the
previous one. Suppose Video Vault can purchase each tape for $65 and Video
Vault charges $3 per rental. How many tapes should Video Vault purchase? What
is the Video Vaults suppliers profit if the suppliers production, handling and
distribution cost per tape is $8?
Say you are Video Vaults supplier and you are considering offering Video Vault a
revenue sharing contract, i.e., you will sell each tape to video vault for some
wholesale price but then you will also collect a certain fraction of Video Vaults
revenue. What terms would you offer Video Vault, i.e., what wholesale price
would you charge and what share of revenue would you let Video Vault keep.
Who are the winners and losers with revenue sharing?
What are the potential concerns with the implementation of revenue sharing?