Professional Documents
Culture Documents
Testbank Problems
Testbank Problems
1
Table of Contents
Model Building
1. Merger
2. Skihaus
3. Evergreen
4. Kurbe
Database Analysis
1. Population
2. Tissues
3. Airline tickets
4. Rental trucks
5. IPO
Optimization
1. Portfolio of Assets
2. Flight Management
3. Staff Scheduling
4. Advertising and Production
5. Environmental Planning
6. Cash Matching
7. Carson Staplers
8. Portfolio of Stocks
9. Learning Curve
Simulation
1. Hybrid Car
2. Production Planning
3. Endowment
4. Buy Now
5. EToys Simulated
6. Doctors
7. Mutual Fund
2
Model Building Problems
The St. Mary (‘Mary’) and Mt. Sinai (‘Sinai’) hospitals are planning to merge. Together, they
must design a new employee benefits plan that will cover all employees in the merged hospital.
To save on administration costs, they have decided to adopt for the entire merged hospital either
Mary’s plan or Sinai’s current plan.
The table below shows that Sinai’s average benefits are more generous than Mary’s. For
example, an employee with a family receives, on average, $550 in annual benefits under Mary’s
plan and $830 in benefits under Sinai’s plan. In total, the employee with a family would receive
an average of $550+$90+$40+$15=$695 under Mary’s plan and significantly more under
Sinai’s. Overall, 65% of all employees at both hospitals enroll under a family plan and the
remainder enrolls under the individual plan.
Type of
Benefit category employee Mary Sinai
Medical Individual $300 $400
Family $550 $830
Dental $90 $85
Life Insurance $40 $80
Disability
Insurance $15 $15
The table displays average benefits, but either decision will significantly reduce benefits for
some employees. Mary has 1200 employees, and if Sinai’s plan is adopted then 12% of Mary’s
employees are expected to leave because of the reduction in benefits. Sinai has 750 employees,
and if Mary’s benefits are adopted then 30% are expected to leave Sinai. On average, it will cost
$45,000 to replace an employee who leaves either hospital. Replacement employees are similar
to current workers: average costs are equal to those in the table and 65% enroll in a family plan.
The hospitals would like to assess the total cost over the next 8 years of applying each plan to the
entire merged hospital. Assume that the following events will occur now: the new plan is
announced, some employees leave because of a reduction in benefits, and these employees are
immediately replaced. Then, the costs described in the table above will be booked after 1 year.
During years 2-8, total benefits costs will grow by 8% annually. Management uses a discount
rate of 5%.
a. Which benefits plan has the lower cost (expressed as an NPV)? Plan: ________
NPV of the lower-cost plan: $___________________
b. At what growth rate in benefits costs do the two plans have an equal NPV? _______%
3
2. Assortment Planning at Skihaus
Skihaus is a retail chain that specializes in the sales of Alpine skis. Skihaus sells its own brand of
skis in three styles: S, DX and LX. The three skis cost Skihaus $260, $400, and $480,
respectively, to purchase from the manufacturer. Throughout the selling season the chain will
sell them to consumers for $300, $450, and $550, respectively. At the end of the season if the
skis are not sold then they will be sold to a liquidator for $50, $230, and $290. Skis are
purchased from the manufacturer once, before the selling season.
Skihaus has identified two distinct segments of customers who purchase from the store: early
buyers who prefer top-of-the-line skis (Earlies) and bargain-hunters (Bargains). As their name
implies, Earlies arrive first in the season, and you may assume that all Bargains arrive at the end
of the season, after all Earlies.
The two segments have different preferences for skis. Earlies prefer to buy LX skis. If LX skis
are sold out, they will always buy DX skis, and if LX and DX skis are sold out, Earlies will
always buy S skis. Bargains prefer S, then DX, and will not buy LX skis. Assume each customer
buys one pair of skis.
Skihaus expects to see 5,000 buying customers during the season, of which 60% are Earlies and
40% are Bargains. Skihaus’ purchasing manager plans to begin its season with 1,300 LX skis,
2,200 DX skis and 1,900 S skis, but she is considering alternate order quantities.
a. Build a spreadsheet model that evaluates the base-case plan described above. You are not
expected to optimize this model. The model should be flexible, allowing all of the parameters
and decisions to change. You may, however, assume here that the preferences described
above for each market segment will not change (e.g., you may assume in your calculations
that Earlies will always prefer LX to DX to S).
b. In the real world, customer preferences can be uncertain. Market research may show, for
example, that Earlies really prefer DX to S and will not buy LX. Take your model from part
(a), copy it to a new worksheet, and change it so that customer preferences can be updated
easily. In other words, parameterize customer preferences.
Once you have built the model, find the objective function value with all parameters the same
as in part (a) except that Earlies now prefer DX to S and will not buy LX: $_____________
Evergreen College holds an endowment currently worth $700 million (an endowment is a
permanent fund of money that generates income and also can be used to pay for projects). The
Chief Investment Officer of the endowment has hired you to predict the performance of the
endowment over the next 10 years.
4
Given the current investment strategy, money invested in the endowment is expected to grow by
8% per year. Over the next 7 years, Evergreen’s development office is running a major capital
campaign among its alumni. During the entire 7-year campaign the college expects to collect a
total of $300 million in contributions that will be added to the endowment, although the
campaign may collect a total as low as $150 million and as high as $400 million.
A portion of the endowment is also spent each year to support ongoing activities at the college
(this is called the draw). During this past year, the draw was $65 million. During the next ten
years the draw in each year will be calculated according to a formula that is the sum of two
terms: (i) 70% of the previous year’s draw and (ii) a certain percentage (p) of the value of the
endowment at the beginning of the year. This percentage p depends on the level of the
endowment itself, and is given in the following table:
For example, if the endowment is $650 million at the beginning of the year, then the draw is
equal to (70%)*(previous year’s draw) + (2%)*($650 million).
Build a spreadsheet to calculate the value of the endowment after the next 10 years. Your
spreadsheet will be graded both for its technical correctness and for its adherence to the
principles of spreadsheet engineering. Note that documentation of each calculation is not
required, nor is extensive formatting for appearance. However, if you need to make any
assumptions beyond the problem description given above, you should include an explanation of
these assumptions at the top of the spreadsheet.
b. In addition to the draw, the college may spend part of the endowment for a new science
building. The decision to begin spending and construction for the building in any particular year
depends on two goals being met: (i) the capital campaign reaching at least $200 million by the
beginning of the year and (ii) the total endowment reaching at least $800 million by the
beginning of the year. If and when both these goals are met, then the college will spend $50
million from the endowment during the first year of construction and $60 million during the
second, for a total building cost of $110 million. If the two goals are not met during the 10 years
then the building will not be built.
Copy your model from part (a) onto a new worksheet and then extend the model to include this
new information. Again, if you need to make any assumptions beyond the problem description
given above, you should include an explanation of these assumptions at the top of the
spreadsheet.
5
Given the parameters described above, including those in part a, find the value of the endowment
after the next 10 years.
Now suppose that the capital campaign collects only $150 million. What is the final endowment
value?
Now suppose that the capital campaign collects $300 million, but that the initial endowment
value is only $250 million. What is the final endowment value?
Kurbe Marketing Research (KMR), located in Boston, specializes in marketing research in the
consumer goods industry. Each year KMR conducts hundreds of research studies for a variety of
clients utilizing focus groups, internet surveys and telephone surveys. KMR is currently studying
the resources it dedicates to telephone surveys.
Demand for KMR’s telephone research varies throughout the year. KMR conducts the vast
majority of its research in-house, but if it cannot handle all of the work at peak demand times, it
is forced to subcontract some of the phone interviewing, which results in a lower profit margin.
Also, outsourcing the phone interviews means that KMR loses some control of the process. This
can result in problems such as missed deadlines and a decrease in quality. However, dedicating
enough resources to ensure no outsourcing would not be very efficient, since it would result in
more costs and decreased productivity due to idle time during slower periods. Nonetheless, based
on the recent amount of outsourcing, KMR management feels that the company has too few
operators conducting telephone survey research. You have been assigned the task of studying
this problem and making a recommendation.
The annual fixed costs (management, facilities and other overhead charges) associated with the
telephone survey group are $320,000. Every operator needs a computer-assisted telephone
interviewing station (known as a CATI). Currently KMR has 48 CATIs and they are considering
purchasing additional ones this year. Of the 48 CATIs currently on hand, 16 per year will be
replaced each year for the next 3 years. New CATIs will not need replacement over this time
period.
The purchase cost of a CATI is $36,000 and is likely to remain stable for the next 3 years. The
cost of an operator (one per CATI) per month is $2,200. Assume that only additional CATIs
6
(beyond the current number of 48) require training (replacement CATIs do not since they already
have operators).
The time frame for the study is the next three years. Demand varies by month and forecasted
demand for the coming year is shown in the table below (demand is expressed in dollars of
revenue). A CATI can handle approximately $12,000 of work per month. Demand is expected to
grow by 10% per year (the forecasted amounts in the table have already factored this in for next
year). Assume that each month’s demand will grow by this amount in years 2006 and 2007. This
increase in revenue will come from increased volume of work rather than increases in pricing
(the forecast assumes stable pricing).
Management is looking for the best level of in-house operation (number of CATIs). KMR
receives 15% of the revenue for the work that is outsourced to an outside vendor. Finally, for
studies such as this, KMR uses an annual discount rate of 10% and their tax rate is 34%.
a. Build a spreadsheet to help KMR analyze this situation. In your base case, evaluate the NPV
of net income after taxes assuming KMR maintains 48 CATIs in all years.
You may use the following page to sketch your spreadsheet (although your sketch will not be
graded). Your spreadsheet will be graded both for its technical correctness and for its adherence
to the principles of spreadsheet engineering. Note, however, that documentation is not required,
nor is extensive formatting for appearance.
b. Construct a graph on a separate sheet in your workbook to show the sensitivity of the NPV of
net income after taxes to the number of CATIs. Sketch your results here, showing numerical
scales for the horizontal and vertical axes, and clearly identify the optimal number of CATIs.
c. In the base case (with 48 CATIs) how much lower does KMR’s tax rate have to be to increase
the NPV of net income after taxes by 10%?
7
Database Analysis
Note: To answer this question you will need to use the database PopulationData.xlsx, which
shows U.S. state populations from 1990 to 1999.
a. Which state(s) had a 1990 population less than 5 million and a 1999 population more than 5.5
million?
b. Which state had the largest absolute growth in population over the period 1990-1999?
c. Which state had the largest relative (percentage) growth in population over the period 1990-
1999?
d. By how many more people has the average of the population of the 50 states grown from
1995-1999?
Note: To answer this question you will need to use the database TissueData.xlsx, which describes
purchases of bathroom tissues at stores in a grocery chain over a period of several weeks.
a. What was the total number of cases sold for all brands over this period (calculate a single,
grand total)?
b. What was the total number of cases sold by Kleenex over this period?
c. What was the total number of cases sold by Kleenex in Size 12 over this period?
d. Which brand had the highest total number of rolls sold over this period? Also state the number
of rolls.
brand: ____________________
______________________________________________________________________________
3. The Airline Ticket Database
The file TicketData.xlsx contains a random sample of one-way airline tickets from the first
quarter of 2006. Each record represents one ticket. Note that a single ticket may allow multiple
passengers to fly (e.g., these ‘tickets’ may be bulk purchases by corporations and travel
agencies). Copy this database to a clean worksheet and use it to answer the following questions.
8
a. List the 3 highest MARKET_FARES for tickets operated by DL.
b. List the MARKET_FAREs of all tickets operated by Delta (DL) from LAX to SAV.
The file TruckData.xlsx contains repair data on 1884 rental trucks. Copy this database to a clean
worksheet and use it to answer the following questions.
b. How many distinct types of transmission manufacturers are represented in the database?
c. List the vehicle numbers for all the trucks from vehicle manufacturer B.
d. What is the odometer reading for all trucks with a gear ratio over 4.0, age less than 60, and
transmission J?
e. What is the average odometer reading for all records in the list?
g. Which transmission manufacturer has the lowest average odometer reading for gear ratios of
3.90?
i. Create (and record below) a table for the maximum age by transmission manufacturer and
engine manufacturer (exclude blanks.)
9
5. The IPO Database
The file IPOdata.xlsx contains raw data on 78 IPOs (Initial Public Offerings of stock) issued
between 1994 and 1999. Copy this database to a clean worksheet and use it to answer the
following questions.
a. What is the average revenue, by underwriter, for the IPO year (REV IPO YR-0) and for the
following year (REV IPO YR-1)?
Average Revenue
Underwriter YR-0 YR-1
b. What was the net income in the IPO year (NI IPO YR-0) for the five smallest IPOs in terms of
the number of employees?
Net Income
1
c. Among IPOs issued in 1999, which one had the largest difference between the high and low
ranges as a percent of the actual price?
10
Sensitivity Analysis and Bugs
You are the team leader of a unit of a US non-profit organization based in Banjul, Gambia. The
non-profit’s mission is to ensure that rural populations worldwide have access to health and
sanitation-related supplies. Due to the sudden departure of one of your team leaders, you are
taking over the responsibility for ordering certain medical supplies for three villages.
The spreadsheet Banjul.xlsx has been built to assist in ordering supplies for five products listed in
column A. Columns B and C give the number of items per package and the cost per package.
Columns D-G give the number of items needed per person for four population groups and the
number in each of those groups.
The results are calculated starting in column I. Columns I and J determine the number of each
item and the number of packages needed, respectively. Columns K and L give the number
currently in stock (currently set at zero for all items) and the number that need to be ordered.
Costs are calculated in column M and totaled in cell M10.
Note: each question should be answered independently of the others, starting with the base case
for each question.
a. How many additional Other Adults (over the current 9000) could be added to the population
before the total cost would exceed $7,000?
b. A check of the stock on hand shows an inventory of 124 packages of A-bandages, 16 packages
of B-bandages, 82 packages of C-bandages, 72 rolls of tape, and 4 hearing aids. What amounts
will be required for ordering supplies?
A bandages ___________________
B bandages ___________________
C bandages ___________________
c. Construct a table showing how the total cost for the order would vary if the unit cost of a
package of all three types of bandages rose by $0.25, $0.50, and so on, up to $3.00 per package.
d. This model uses 23 parameters in the range B4:G10. If each one were to vary by 15% around
its base case value, which three would have the biggest impact on total costs?
11
e. Assume that the population of Children varies around its base case value by 15%, the
population of Teens varies by 10% and the population of Seniors by 5%. List the size of the
ranges of the Total Cost of Order (the distances from the bottom of the range to the top, in $) as
each of these parameters varies.
Children __________________________
Teens __________________________
Seniors __________________________
2. MediDevice
A U.S. medical device manufacturer, MediDevice Inc., has developed a new blood analyzer for
rapid in-the-doctor’s-office testing. The much-anticipated device will receive final regulatory
approval in time for a January 2005 launch. MediDevice has the next 12 months to work out the
details for their commercial plans. The date for the analysis is January, 2004.
This device is revolutionary and has worldwide appeal; target markets are U.S., Europe and
Japan.
Demand for this product is driven by the fraction of doctor offices adopting the technology.
Forecasts indicate 10% of doctor offices in any country would potentially adopt this type of
device. The time it takes to achieve this technology adoption (starting from 0% adoption) is
expected to be 2-5 years, depending on factors including the effectiveness of the sales force in
each country.
Each office would buy only one analyzer, but once an office buys the analyzer MediDevice
receives recurring revenue from each office from royalties on analyzer supplies (sold by a third
party). These annual royalty revenues will continue until MediDevice's analyzer is displaced by
next-generation technology.
MediDevice expects to be first to market but a major Japanese competitor is working on a
superior next-generation device. When this competitive device hits the market (a number of
years away), it will most likely displace the MediDevice analyzer from a portion of
MediDevice's installed base of doctor's offices – but some of the installed base will continue in
use until third-generation competitive devices ultimately take over the entire market.
The MediDevice board has commissioned you to perform a cash flow analysis of this new
device. Their assessments of the key parameters are summarized below.
Market size
U.S. doctor’s offices 600,000
Europe doctor’s offices 450,000
Japan doctor’s offices 140,000
12
Market penetration
Launch year 2005
Peak year 2009
Initial share 0%
Peak share 10%
Note: Market share is expected to rise linearly from the Launch Year to the Peak Year
and remain constant thereafter until new products arrive in the market.
Price controls
Date introduced 2008
Price reduction 10%
Note: The regulatory authorities are expected to introduce price controls on this device
sometime in the future.
Prices
Device $10,000 per device
Royalty on Supplies $1,000 per year
Costs
Unit manufacturing $4,000 per unit
Sales force $15 million
US plant cost $25 million
Rest of world plant cost $20 million
The spreadsheet model is in MediDevice.xlsx. In the model, cash flows in each column occur at
the end of the year.
This question is in two parts. In Part A you will take the model as given and perform sensitivity
analyses on it. Do NOT alter the model for Part A. In Part B you will identify bugs in the model.
Part A:
Note: In each of the following questions you are to start with the base case model.
a. Sketch the relationship between the deprecation lifetime and the NPV. Consider values for the
depreciation lifetime between 5 and 15 years. Show the exact numerical values for lifetimes of
5 and 15 years on your sketch.
13
b. Sketch the relationship between the unit manufacturing cost and the NIAT in Year 2009.
Consider values for the unit manufacturing cost between $1 and $10. Show the exact numerical
values for costs of $1 and $10 on your sketch.
c. Vary the discount rate between 5 and 15% and the peak share for market penetration between
0 and 10%. Describe business insights about the device that you learn from the results.
d. Vary all parameters in the model except the dates by +15%. List below the three parameters
that have the strongest impact on the NPV.
Part B:
The model as currently built has a number of bugs. Your task in this part is to find these bugs.
The only types of bugs you are to look for include:
wrong input: a numerical input does not correspond to the problem description
hard-coded input: the numerical value of an input appears in a formula (this excludes the
value zero in a cell)
wrong reference: a formula uses an incorrect cell reference (i.e., points to the wrong cell)
You should assume the accounting logic of the model is correct unless it explicitly contradicts
the problem statement. Thus, for example, negative taxes are allowable in row 81.
List the cell address or addresses of the bugs you find below and give a brief explanation of why
it is an error and how you would fix it.
1.
2.
3.
Etc…
_____________________________________________________________________________
14
3. Finding bugs in the Material X model
This model was built to estimate the market potential for a consumer product that uses Raw
Material X as a key ingredient. Unfortunately, this model has bugs.
You may assume that the three supporting worksheets (Market Share Projections, Market Data,
and Price Data) are error-free. All the bugs are on the Model worksheet.
Identify all the bugs and fix them as you go. List them below, giving the cells involved, the type
of bug, and a short description of how you fixed the bug.
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
15
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
Note: The number of lines does not necessarily reflect the number of bugs.
Once you have made your changes and the spreadsheet is correct, what is the NPV in cell C10?
__________________
______________________________________________________________________________
4. Capacity model with bugs
Your task in this question is to identify all the bugs in a spreadsheet built for you. (Note: a “bug”
here refers to an error that causes the spreadsheet to return the wrong value of the output, not a
poor choice of layout or format or other poor practices. A bug can involve an error in the data
or an error in a formula.) The situation being modeled is described below. You may assume that
the accounting in the model is conceptually correct.
Lang Drug needs to determine the proper capacity level for a new drug, Niagara.
Their goal is to maximize the expected NPV earned from the drug over the next 15
years, assuming a discount rate of 10 percent per year. It costs $10 to build enough
capacity to produce one unit of drug per year. All construction costs are assumed to
be paid at time 0, while the first profits come in at the end of the first year. It costs $1
per year to maintain a unit of annual production capacity. In year 1, we know demand
will be for 160,000 units of Niagara. The annual percentage growth of demand for
Niagara is 15 percent. During year 1, each unit of Niagara sells for $8. The price of
Niagara will grow at 5 percent per year. Unit variable cost is known to be 40 percent
of sales price. The depreciation rate is 10 percent and the tax rate is 35 percent.
________________________________________________________________________
16
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
Note: The numbers of lines does not necessarily reflect the number of bugs.
b. Once you have made your changes and the spreadsheet is correct, what is the NPV?
_____________________________________________________________________________________
5. Sensitivity analysis for the Northern Museum
The model Museum.xlsx is a prototype planning model for a small regional museum. The model
projects revenues, costs, operating surplus/deficit and the endowment out to 2014, based on 2003
and 2004 actuals and assumed growth rates. The most important outcome measure is the level of
the endowment in 2014, cell N39.
17
Copy this model to your computer and use it to answer the following questions. You may assume
this model is logically correct. Write in the answers here and also document the answer to each
of the questions on a separate worksheet.
Note: each question should be answered independently of the others, starting with the base case.
a. The model is based on 16 parameters in column D. Which four have the biggest impact on the
outcome? (To answer this question, you should assume that each parameter can vary by plus or
minus 10% of its base case value.) List them here in order of importance.
Parameter
1
b. Assume that all 10 growth rates for revenues can vary by +/- 20%, the five cost parameters by
+/- 10%, and the interest rate parameter by 15%. Which four have the biggest impact on the
outcome? List them here in order of importance.
Parameter
1
c. What interest rate (cell D38) would the Museum have to earn to ensure that the endowment in
2014 was equal to its value at the end of 2003?
d. One faction on the Museum Board is concerned about the vulnerability of the endowment to
the interest rate. Sketch a graph that will show how the endowment value in 2014 changes as the
interest rate varies from a low of 0% to a high of 15%. For grading purposes, clearly label the
numerical values at 0% and 15%.
e. Another faction of the Board is interested in ensuring that the endowment in 2014 exceeds $3
million. Create (and record below) a table which shows the combinations of the growth rate in
municipal grants (cell D11) and the interest rate (cell D38 that will bring this result about? (Each
parameter can vary between 0 and 15%).
18
______________________________________________________________________________
6. Office Building with bugs
Your task in this question is to identify all the bugs in a spreadsheet built for you. (Note: a “bug”
here refers to an error that materially changes the results of the spreadsheet, not a poor choice
of layout or format. A bug can involve an error in the data or an error in a formula.) The
situation being modeled is described below. You may assume that the accounting in the model is
conceptually correct.
The objective of this model is to predict the after-tax cash flow resulting from constructing and
operating an office building over a five year period. At a planned size of 180,000 square feet, the
expected cost is $80 per square foot. The investors plan to take out a mortgage for 85 percent of
the cost of the building (paying the remainder in cash), and they have been guaranteed a rate of
12 percent for a term of 30 years. The owners must pay for the cost of operating the building,
which includes taxes, insurance, maintenance, and certain utilities. They assume that the average
operating cost per square foot will be $1.20. They have also estimated the rental rate they can
charge is $15 per square foot, with an occupancy rate of 70%. The building has a depreciable life
of 20 years; the effective tax rate is 34%; the discount rate is 10%. Rents in the future are
expected to grow 5% per year, while operating expenses grow 10% and the occupancy rate drops
4 percentage points yearly as the building ages.
Copy the spreadsheet OfficeBugs.xlsx to a worksheet in your own exam workbook. Identify all
the bugs and fix them as you go (this will help you to find them all). List them below, giving the
cells involved and a short description of how you fixed the bug.
etc.
The model EToys.xlsx forecasts the total net income over four years (cell H30) from a new
product. Copy this model to your computer and use it to answer the following questions. You
may assume this model is logically correct. Write in the answers here and also document the
answer to each of the questions on a separate worksheet. Note: each question should be answered
independently of the others, starting with the base case.
19
a. The model is based on 16 parameters. Which four have the biggest impact on total net
income? (To answer this question, you should assume that each parameter can vary by plus or
minus 25% of its base case value.)
Parameter
1
b. How high would we have to set the initial sales price if our goal was to achieve a total net
income of $5,000,000? (Answer to the nearest penny.)
c. How much would we have to increase the price in year 4 in order to increase the total net
income to $3,000,000? (Answer to the nearest penny.)
d. Imagine that unit sales increase by the same percentage in years 2, 3, and 4. Sketch the graph
of total net income as the unit sales increase varies from 10 to 20% in steps of 1%. Label the first
and last values clearly (that is, to the nearest dollar).
20
Optimization Problems
One of the most common applications of optimization in finance involves selecting optimal
portfolios. An optimal portfolio is one that maximizes overall return while satisfying constraints
that represent the investor’s attitude toward risk.
The accompanying spreadsheet Assets.xlsx contains data on the annual percentage return on
stocks, gold, and Treasury bills over the period from 1987 to 2007. A certain investor would like
to determine the percentages of her portfolio to invest in these three asset categories. Her
objective is to maximize the expected return on the portfolio while not investing less than 20% or
more than 50% in each asset class. You may estimate the expected return on each asset with its
average return over the period 1987-2007.
The risk index of an asset is measured by the average absolute percentage change over time. You
may estimate the risk indices for stocks, gold, and Treasury bills over 1987-2007. The risk index
for a portfolio is the weighted average of the risk indices of its component assets, where the
weights are the percentages invested in each asset. For example, in the data we see that the risk
index for stocks is 16.9. Our investor would like the risk index of her portfolio to equal 15.
b. Describe how her average return and the percentages invested in each asset change as she
increases her acceptable risk index from 15 to 18.
______________________________________________________________________________
Part a
Alpha Airlines Flight 1, carrying 265 passengers, has just taken off from Chicago and is flying to
the Alpha hub in Denver. Because the flight was delayed in Chicago, if the airplane flies at
normal speed then it will arrive 18 minutes late. The airplane can speed up, but speeding up
increases fuel consumption and therefore increases costs. Specifically, to reduce travel time by x
minutes adds $7.5x2 to the cost of the flight (for example, reducing the flying time by 10 minutes
would cost $7.5*102 = $750). Alpha will not slow down the plane and will not speed up so much
that the airplane arrives early.
Because each delayed passenger can delay the departure of a connecting flight at the hub, Alpha
calculates that there is a significant cost for each delay to each passenger. Here, assume that
delays to Flight 1 cost $0.45/passenger/minute delayed.
21
What is Alpha’s total cost, given your answer above? _________
(ii) In this question only, assume that the cost of delay is $2.00/passenger/minute. Now,
Part b
Alpha is currently managing the speeds of 4 late incoming flights. Flights 1-4, which were
originally scheduled to arrive at the hub at the same time (assume that all 4 flights can land
together because the hub has multiple runways). All of the passengers on these flights will
change planes at the hub and then depart on four outgoing flights (Flights 5-8). The file
Flight.xlsx contains the number of passengers on the four incoming flights and the number of
those passengers who will be transferring to the four outgoing flights. For example, Flight 3 has
130 passengers who transfer to outgoing flight 6. The file also shows the number of minutes the
flights are currently late (18, 15, 8 and 6 minutes for Flights 1-4, respectively). There are no
other incoming flights, and all flights have a sufficient number of seats to hold all passengers.
Alpha’s policy is to hold outgoing flights until all connecting passengers are aboard. Scheduled
connection times at the hub would allow on-time departures of outgoing flights if all incoming
flights were on-time, but connection times contain no extra time. Therefore, if a passenger’s
incoming flight is delayed by a certain number of minutes, that passenger’s outgoing flight must
be delayed by at least the same number of minutes. For example, if Flight 3 does not speed up
and is late by 8 minutes, then Flight 6 must be delayed by at least 8 minutes (Flight 6 may be
delayed longer than 8 minutes if other connecting flights are even later). Again, every outgoing
flight will take off as soon as all connecting passengers are aboard.
Here, Alpha measures its delay costs in terms of delays to passengers on outgoing flights because
passengers care about when they arrive at their final destination; they do not really care if their
arrival at the hub is delayed. Delay costs are $0.45/passenger/minute of outgoing flight delay. As
in part (a), speeding up by x minutes costs $7.5x2 .
By how much should Alpha speed up each incoming flight?
Flight 1 __________
Flight 2 __________
Flight 3 __________
Flight 4 __________
______________________________________________________________________________
22
3. Scheduling staff
You are the Director of the Computer Center for Gaillard College and responsible for scheduling
the staffing of the center, which is open from 8 a.m. until midnight. You have monitored the
usage of the center at various times of the day and determined that the following numbers of
computer consultants are required:
Minimum Number of
Consultants Required
Time of Day to be on Duty
8 AM – noon 4
Noon – 4 PM 8
4 PM – 8 PM 10
8 PM – midnight 6
Two types of computer consultants can be hired: full-time and part-time. The full-time
consultants work for eight consecutive hours in any of the following shifts: morning (8 AM –
4PM), afternoon (noon – 8 PM), and evening (4 PM – midnight). Full-time consultants are paid
$14 per hour.
Part-time consultants can be hired to work any of the four shifts listed in the table. Part-time
consultants are paid $12 per hour. An additional requirement is that during every time period,
there must be at least one full-time consultant on duty for every part-time consultant on duty.
The staffing model in the file Staffing.xlsx can be used to find an optimal staffing plan.
Note: all questions are independent and should be based on the original model.
Questions
a. Determine a minimum-cost daily staffing plan for the center. Record your results here:
b. How much could the Center save if it were to ignore the requirement that at least one full-time
consultant must be on duty for every part-time consultant?
23
Cost savings: ____________
c. Demand for consultants has been rising in time slot 3 (4 PM – 8 PM), but you have been given
a budget of $2,000 for your daily costs. At what level of demand for consultants in time slot 3
does the minimum cost first exceed your budget?
d. In the future, the minimum number of consultants required on each shift is expected to
increase by 1 to 5 people, as illustrated below:
Sketch a graph to show how a) the total number of full-time consultants and b) the total number
of part-time consultants changes as the requirements increase.
24
______________________________________________________________________________
4. Coordinating advertising and production
The Hawley Lighting Company manufactures four families of household lighting at its factory.
The product families are table lamps, floor lamps, ceiling fixtures, and pendant lamps. The table
below shows the average material costs for each of the products.
Each product is made in one of two production processes. Table lamps and floor lamps go
through the process in Department 1, while ceiling fixtures and pendants go through the process
in Department 2. Variable production costs and capacities (measured in units of product) are
shown in the table below. Note that there are regular and overtime possibilities for each
department.
25
Process Unit cost in units Unit cost in units
Department 1 $16 100,000 $18 25,000
Department 2 12 90,000 15 24,000
Average selling prices for the four products are known, and estimates have been made of the
market demand for each product at these prices. These figures are shown in the following table.
a. What is the optimal production plan for the company? (Give the optimal profit and how much
of each product to produce on regular time and on overtime.)
Hint: use four decision variables for production of each product in regular time, and four
decision variables for production of each product in overtime.
Amend your model to include the advertising decisions and use it to determine the optimal
production and advertising plan for the company. (Give the optimal profit, how much of each
product to produce on regular time and on overtime, and how much to advertise for each
product.)
______________________________________________________________________________
5. Environmental Planning
You are the operations manager of a chemical company which produces five products in a
common production facility. The following table gives potential sales for each product, along
with variables costs and revenues. Production in any year cannot exceed the sales potential.
26
Your production facility rotates through the product line because it can produce only one product
at a time. The production rates differ for the various products. It takes 0.3 hours to make 1 ton of
A, 0.5 hour for a ton of B, and 1 hour each for a ton of C, D, or E. The facility can be operated
up to 4000 hours per year.
The file EnviroPlanning.xlsx contains all the information given above in a format designed for
optimization. Use this template to answer the following questions.
a. Determine the maximum profit the company can achieve from its product line in the coming
year.
$____________________
$____________________
Parts (c), (d) and (e) are independent: changes described in one part do not apply to the other
parts. For example, part (c) involves changing the sales potential of product C from 1000 to 2500
tons. For parts (d) and (e), the sales potential should be set back to the original 1500 tons.
c. Describe in words how the optimal production plan and profit change as the sales potential on
Product C changes from 1000 to 2500 tons.
d. An engineering firm is offering to improve the efficiency of the process for manufacturing
products C, D, and E. This process improvement will reduce the time required for all three
products by up to 10%. How big an improvement to all three products is necessary in order to
offset the cost of $30,000? Your answer should be a percentage between 0% and 10% (answer to
the nearest integer percentage). Include a table of results to justify your answer.
TABLE:
e. The EPA is proposing to limit particulate emissions from your company over the next five
years to the amounts specified below. The sales potential for each of the five products is
expected to remain stable over this period.
27
3 60
4 40
5 20
The company plans to satisfy these requirements by modifying the product mix each year. What
would the optimal profits be in each of the next five years?
Year Profit
1 $______________
2 $______________
3 $______________
4 $______________
5 $______________
______________________________________________________________________________
6. Cash Matching
Integra Marketing is a small brand development firm that has just received the final payment
from a large contract. Integra’s cash level is now $1,500,000, and the CFO must decide what to
do with the money. In the short term Integra will have a steady stream of income and expenses,
but its major contracts end in 5 years. Therefore, the CFO would like to maximize the amount of
cash available at the end of year 5.
All transactions (costs, revenues, investments, return on investment, and interest on cash) are
recorded at the end of the year. ‘Now’ can be considered as the end of year 0.
During year 1, Integra will have $850,000 in costs and $600,000 in revenue. For years 2-5, the
annual rate of cost growth will be 3% and the annual rate of revenue growth will be 10%. There
will also be a one-time cost of $100,000 for a loan due at the end of year 4.
28
Investment A: Each dollar invested in A returns $1.54, 5 years later. For example, if $1
million is invested at the end of year 0 then $1.54 million is available as cash at the end of
year 5.
Any of the three investments may be purchased at the end of any year, including year 0. For all
three investments, the money invested cannot be cashed out before the stated due-date (5, 3 and 2
years for A, B and C, respectively). When any of the three investments is cashed out at the end of
a year, the money is available for immediate reinvestment at the end of the same year, or the
money can be left as cash.
Money left as cash earns interest at an annual rate of 2%. Finally, the CFO requires that at least
$50,000 be available in cash at the end of each year, including year 0.
a. Build a model to evaluate the following proposed plan. Determine not only the value of the
plan at the end of year 5, but also whether it is feasible.
Investment type
Year-End (A, B or C) Amount
0 A $600,000
0 B $250,000
0 C $200,000
Report here the value of this plan. Also, list all constraints the plan must meet and whether
each is satisfied under this plan.
Constraint Satisfied?
b. Use your model to evaluate the following alternate plan. Determine not only the value of the
plan at the end of year 5, but also whether it is feasible.
Investment type
Year-End (A, B or C) Amount
0 A $600,000
1 B $250,000
1 C $200,000
29
Report here the value of this plan. Also, list all constraints the plan must meet and whether
each is satisfied under this plan.
Constraint Satisfied?
c. Devise an investment plan for Integra that maximizes the amount of cash that will be available
at the end of five years. How much cash will be available at the end of five years?
Describe the optimal investment plan (you may not need all rows):
Investment type
Year-End (A, B or C) Amount
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
d) Suppose that if Investment A is used, at least $800,000 of Investment A must be bought. All
other information remains the same, as described above.
Create a new model (or new models) to help the CFO maximize the amount of cash that will
be available at the end of five years. Describe in words the changes you made to your
original model:
________________________________________________________________________
30
Use your new model(s) to devise a new optimal investment plan for Integra. How much cash
will be available at the end of five years?
Describe the optimal investment plan (You may not need all rows):
Investment type
Year-End (A, B or C) Amount
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
__________ ____________ __________
7. Carson Staplers
31
After considering the firm’s variable labor, material and overhead costs, the accounting
department has determined the unit manufacturing cost for each component. These data, along
with the purchase price quotations by the external supplier, are as follows:
Manufacturing Purchase
Component Cost Cost
a. Build a model with which you can evaluate the following proposed plan. You must be able to
determine not only the total cost of the plan but also whether it is feasible.
Report here the total costs of this plan. Also, list all constraints the plan must meet and whether
each is satisfied under this plan.
Constraint Satisfied?
b. Determine the production and purchasing plan that minimizes total cost. Report here the plan
itself and the minimum cost.
c. What is the impact of changes in the manufacturing costs of the Base component on the total
cost? Consider variations in the cost from $0.50 to $1.00 in steps of $0.05. Report your results in
the table below.
32
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
______________________________________________________________________________
8. Optimizing a Portfolio of Stocks
Four stocks are available for creating a portfolio: Microsoft, ATT, GM, and GE. We have data
on the mean, standard deviation, and the covariance of returns. If we chose to invest certain
percentages of our assets in each of the four stocks we can determine the mean and standard
deviation of the resulting portfolio, as in the accompanying spreadsheet.
One way to formulate the problem of determining the optimal portfolio for an individual is to
combine the mean and standard deviation of the return into a single measure of the overall utility
of the portfolio. This can be done using a penalty factor (p) for the standard deviation, so that the
utility is given by the mean return less a penalty for the standard deviation:
The portfolio optimization problem then becomes one of finding the appropriate percentages, or
weights, for each of the four stocks to maximize this utility measure. Note: all four weights must
be non-negative.
a. Assume you have a client whose penalty factor equals 2.0. Determine the optimal portfolio.
Report here the four weights, as well as the mean and standard deviation of the optimal portfolio.
33
Portfolio mean _____________
b. Let the penalty factor vary from 0 to 3 in increments of 0.5. Determine the optimal portfolio
for each value of p. Report here the optimal utility, portfolio mean return, and portfolio standard
deviation.
c. Describe in words how and why the pattern of weights varies as the penalty factor p varies
from 0 to 3.
______________________________________________________________________________
(p x1+q)/(1+q).
34
We now consider a specific production setting where there is learning. Demands for our single
product over the next five periods are 100, 150, 300, 200, and 400. Holding cost per unit per period
is $.30.
a. In the worksheet Learning.xlsx, use Solver to find the best fitting learning curve (that is, find the
values of p and q) so as to minimize the sum of squared error where the predicted value of unit i is
p*(i)q. Report your results below.
Value of p: __________________
Value of q: __________________
b. Based on your results from part a), use the approximation (px1+q)/(1+q) as the cost (in dollars) of
producing a batch of size x. Assume that the learning from one time period to another is lost (so
there is no transfer of learning between time periods) and the amount we produce in a given period
is one batch. Solve the production planning problem of minimizing the sum of production and
inventory costs, while satisfying demand. Assume we must have an ending inventory in period 5 of
at least 50. Report your results below.
c. Solve the same production planning problem ignoring the learning curve, that is, assume that
every unit costs p dollars (where the parameter p was estimated in part a) above). Report your
results below.
d. Compare the optimal production plans you obtained in b) and c). Explain any differences.
35
Simulation Problems
Note: Refer to the model Hybrid.xlsx for this problem. This is a deterministic version of the
model that you must convert into a simulation model. You should assume the model logic is
correct.
ADC Auto Corporation is developing a new type of hybrid car. This car will generate sales over
the next 10 years. Analysts have gathered the following information about the financial side of
this decision from the marketing and engineering departments.
Fixed cost: Fixed costs are assumed to be normally distributed with a mean of $2.3 billion and a
standard deviation of $0.5 billion. Assume fixed costs will be incurred at the beginning of year 1,
before any sales are recorded.
Variable production cost: This cost, which includes all variable production costs required to
build a single car, is assumed to be normally distributed in the first year of production with a
mean of $7,800 and a standard deviation of $600. After the first year, variable costs will inflate
in each year by an amount given by a normal distribution with a mean of 1.05 (a 5% increase)
and a standard deviation of 0.015. Each year’s inflation factor is independent of the other years.
All production costs in a year are incurred at the end of that year.
Selling price: The price in the first year is set at $11,800. After the first year the price will
inflate by the same inflation factor that drives variable production costs. Revenues are realized at
the end of the year.
Demand: Demand is normally distributed in the first year with a mean of 100,000 units and a
standard deviation of 10,000. After the first year, demand is expected to increase each year by an
amount given by a triangular distribution, with a minimum value of 0%, a maximum value of
10%, and a most likely value of 3%. The growth rate from year to year is independent of all
other years.
Production: DC bases its production decision each year on demand in the previous year. (The
first year’s production will be 100,000.) DC’s policy is to produce an amount equal to demand in
the previous year plus a multiple k of the standard deviation of demand (10,000). For example, if
k = 0 then DC will produce just the previous year’s demand. The higher the k-factor, the higher
the likelihood that production will cover demand. If demand is higher than production in any
year, the excess demand is lost. However, if production in any year is greater than demand, DC
will sell the excess cars at an end-of-year discount of 50% off the normal price.
Interest rate: DC plans to use a discount rate of 10% for future cash flows. Note that the fixed
cost is not discounted, since it occurs at the beginning of the first year, but costs and revenues for
the first year are discounted one year because we assume they occur at the end of the year.
36
ADC would like to determine the financial implications of these assumptions. In particular, it
would like your recommendation as to an appropriate value for the parameter k.
a. Assume k = 0 for a base case. What is the mean Total NPV that ADC can expect from this
new car?
NPV: _____________________
b. Assume k = 0. What is the probability the Total NPV will be less than one standard deviation
above the mean?
Probability: _____________________
c. Assume k = 0. What is the probability that the Total NPV will exceed $1,800 million and
Sales revenue in year 10 will exceed 3,000?
Probability: _____________________
d. What value of the safety factor k between 0.0 and 1.5 (to the nearest 0.1) should ADC choose
in order to maximize the mean Total NPV?
e. How does the optimal value for the safety factor k change with the standard deviation of
Demand in year 1 (set at 10 in the base case)? Vary the standard deviation of “Demand in
year 1” from 5 to 50 in increments of 5. Use the graph below to display your results. Show
the exact numerical values of the optimal k for standard deviations of 5, 25 and 50 on your
sketch.
37
______________________________________________________________________________
2. Production planning
Your company has agreed to produce 8000 ounces of shampoo for a retailer. This product is
manufactured in batches that take 5 to 11 days according to the following distribution (this table
is also in the file Shampoo.xlsx).
Only one batch can be made at a time. Each batch yields an uncertain amount of product, ranging
between a minimum of 600 ounces and a maximum of 1100 ounces, with a most likely value of
38
1000 ounces. Finally, 20% of the time a batch fails inspection and the entire batch must be
discarded.
Your task is to build a model with which to forecast the number of batches and the number of
days that will be required to meet the customer’s order of 8000 ounces.
c. What is the probability that the number of days required will exceed 125? ________
d. The retailer may want to order more than 8000 ounces, but only 150 days are available in the
manufacturing plant. Your company’s goal for any order is that there is less than a 2% chance
that the number of days required for the order exceeds 150 days. What is the largest order your
company can accept and still achieve the goal? Give your answer to the nearest 100 ounces.
Answer: __________
______________________________________________________________________________
3. Simulating the Evergreen College Endowment
Evergreen College holds an endowment currently worth $700 million. The Chief Investment
Officer of the endowment has hired you to predict the performance of the endowment over the
next 10 years.
You are given a deterministic spreadsheet for this purpose, Evergreen.xls. The spreadsheet
incorporates the following elements (you can assume that the spreadsheet is correct):
Annual growth in the endowment due to financial investments;
Contributions from a capital campaign;
An annual draw on the endowment for ongoing activities, where the size of the draw
incorporates percentages of the previous year’s draw and the current endowment;
The cost, over 2 years, of a new science building. Spending and construction of the new
building is conditional on meeting targets in the endowment and the capital campaign.
To improve the model you want to incorporate some measure of uncertainty. You have selected
the following probability distributions:
Annual rate of return: Each year’s rate of return is lognormal with a mean of 8% and a
standard deviation of 2%. Each year’s rate is independent of all other years.
39
Total collected in capital campaign: triangular with a minimum of $100 million, most likely
value of $300 million, and maximum of $450 million.
Length of the capital campaign: The campaign could last 6, 7 or 8 years, with these
probabilities:
All of the following questions are independent. In other words, at the start of each of (a)-(f),
return to the base case given in the original spreadsheet.
b.
(i) What is the probability that the endowment at the end of year 10 (‘final
endowment’) is greater than $1 billion?
probability that final endowment > $1 billion: ____________
(ii) What is the 90th percentile of the distribution of the final endowment?
c. The final endowment can be increased by reducing the draw. Specifically, the final
endowment can be increased by lowering the “% of previous year's spending” (cell C26 in
Evergreen.xls). What is the largest value of “% of previous year's spending” that will ensure that
the mean final endowment is greater than $1 billion? Answer to the nearest 1%.
d. Of course, the final endowment can also be enlarged by increasing the expected returns from
investments. What is the minimum average annual rate of return needed to ensure that the
probability that the final endowment is at least $1 billion is greater than 0.9? (You may assume
that the standard deviation of returns remains at 2%.) Answer to the nearest 0.1%
40
Average rate of return: _________ %
e. What is the probability that construction of the new science building begins in year 5?
Probability: _________
f. The year in which construction of the new science building begins is random, for it depends
upon the uncertain values of the endowment and the capital campaign. What is the mean and
standard deviation of the first year of construction? Note: for this problem, if construction does
not begin by the end of year 10 then you may assume that construction begins in year 11.
Mean: _________
______________________________________________________________________________
4. Buy now or pay to buy later
You are considering investing in a firm that is developing a new, promising technology. The firm
will spend $5 million during each of the next 2 years while it files a crucial patent application
and waits for the result. During year 2 the application will be approved (with probability 70%) or
will be rejected. If the application is rejected then the firm will go out of business and will have
no assets to sell off.
If the application is approved in year 2, the firm will immediately begin selling its patented
product during year 3. The initial market size for the product (the number of units sold in year 3)
will be normally distributed with a mean of 1 million and a standard deviation of 0.2 million.
The growth rate in subsequent years will be normally distributed with a mean of 10% and a
standard deviation of 5% (whatever the growth rate turns out to be, it will be the same in all
future years). The unit margin is normally distributed with a mean of $50 and a standard
deviation of $15 (whatever the unit margin turns out to be in year 3, it will be the same in all
future years). The discount rate is 15% and the analysis should cover 15 years, in total. Assume
all assets are worthless at the end of 15 years.
When modeling this problem, you should assume that all cash flows occur at the end of the year.
$ __________
b. Suppose that you are offered the opportunity to pay $100 million (paid now) in exchange for
a 50% share of the firm. This share entitles you to 50% of all cash flows in years 1-15,
including negative cash flows (e.g. at the end of years 1 and 2 you would have to pay 50% of
the firm’s expenses). For this deal,
41
(i) What is the mean NPV?
$ __________
(ii) What is the probability that you lose more than $50 million?
probability: ___________
$ __________
(ii) What is the probability that you lose more than $50 million?
probability: ___________
a. The retailer may want to order more than 8000 ounces, but only 150 days are available in the
manufacturing plant. Your company’s goal for any order is that there is less than a 2%
chance that the number of days required for the order exceeds 150 days. What is the largest
order your company can accept and still achieve the goal? Give your answer to the nearest
100 ounces.
Answer: __________
5. EToys Simulated
EToys is trying to determine the total net income they can expect to receive over the next four
years from a new toy. During the course of their analysis they have made some modifications to
the model we discussed in class on October 7. In particular, they have incorporated a demand
relationship between initial price and unit sales (see cell E8). The key parameter in this
relationship is the price elasticity (cell E5).
42
They also have decided that three parameters are uncertain: the price elasticity itself (cell E5),
the unit sales increase (cell F9), and the sales price increase (cell F16). The following probability
distributions have been selected as representing their uncertainty:
Price elasticity: triangular with minimum of –2.0, most likely of –1.5, and maximum
of –1.0.
Unit sales increase: lognormal with mean of 0.12 and standard deviation of 0.04.
Sales price increase: triangular with minimum of 1.5, most likely of 2.0, and
maximum of 3.0.
a. Make a simulation run of 5000 trials with the initial price (cell E15) set at $20. Determine the
following:
(ii) (1 point) probability total net income will exceed $2 million: ____________
b. What is the minimum number of trials needed to reduce the Mean Standard Error of total net
income to below $21,000? (Answer to the nearest 1000.)
____________ trials
c. Find the initial price (cell E15) that maximizes the expected total net income. Answer to the
nearest $10, between $20 and $100. Provide a table of numbers that justifies your answer.
$______________
TABLE:
d. How sensitive is the probability that net income will exceed $2 million to the initial sale price
(for values from $20 to $100 in steps of $10). Provide a table of numbers that justifies your
answer.
ANSWER:
TABLE:
43
e. The minimum price elasticity in the base case is -2.0. How sensitive is the optimal initial price
to variations in the minimum price elasticity between -10 and -2? (The optimal initial price is the
initial price that maximizes the expected total net income). Answer to the nearest $10, between
$20 and $100. Provide a table of numbers that justifies your answer. Hint: set the number of
trials to 1000.
ANSWER:
TABLE:
______________________________________________________________________________
Part I
In the U.S. health care system at present, patients typically wait weeks, even months, to see a
doctor. However, a new trend among private doctors is to eliminate most waiting through the
“Open Access” system. In an open access system, the patient is guaranteed an appointment on
the day after they call (calls on Friday and over the weekend are scheduled for Monday). If the
time needed by the doctor to see all scheduled patients exceeds the time available on any
particular day, then the doctor works overtime on that day.
Of course, if the doctor has too many patients wanting appointments then there will be too much
overtime. The number of patients needing an appointment for each day depends upon the number
of patients the doctor agrees to serve. This number is the doctor’s panel size. Typical panel sizes
range from 1000 patients to 3000 patients.
Studies show that, on average, each patient has a 0.008 chance of needing an appointment for
any particular day. Thus a panel of 1000 will generate an average of 0.008*1000 = 8
patients/day. If the panel size is 3000, there will be an average of 0.008*3000=24 patients/day.
But of course the number of calls each day is variable; the random number of patients who need
an appointment for each day can be described with a Poisson distribution with a mean equal to
(0.008)*(panel size).
44
The doctor’s work day is divided into 18 slots, each approximately 20 minutes in length. Seventy
percent of patients require 1 slot, 20% require 2 slots, and 10% require 3 slots. The doctor can
use overtime to create more slots in the day, if that is required to serve all the patients.
A doctor makes a profit of $30 for every slot in which he or she sees a patient. This figure
represents revenue minus the variable costs of treating patients. The fixed costs of running the
business can be ignored here.
However, if the doctor uses overtime then there is an additional cost of $80 per slot. Thus a
doctor looses $50 for each overtime slot used. Despite this, under the open access system a
doctor cannot turn away a patient.
The American Medical Association would like you to build a simulation model to find the best
panel size for an individual doctor. The model should determine the doctor’s profit for one day
as well as how often the doctor will need to work overtime.
a. Assume that the doctor uses a panel size of 2000 and use 5000 trials for your simulation run.
What is the mean one-day profit?
b. From the simulation run in part (a), report a 95-percent confidence interval for your estimate
of the mean profit.
c. What is the probability that the doctor will need to work overtime? Again assume that the
doctor uses a panel size of 2000 and use 5000 trials for your simulation run.
Probability: _____________________
d. What is the average number of non-overtime slots that are idle (not used to serve a patient)?
Again assume that the doctor uses a panel size of 2000 and use 5000 trials for your simulation
run.
e. What panel size maximizes profit? Support your conclusion with a table of profits over an
appropriate range of panel sizes (answer to the nearest 100). Circle the optimal value.
45
_____________________ _____________________
Table:
Part II
Patients who request an appointment with the doctor can be divided into ‘urgent’ (70%) and
‘nonurgent’ (30%). Both urgent and nonurgent patients have the same distribution of slots, as
described above: 70% need 1 slot, 20% need 2, 10% need 3. Again, the doctor has 18 slots
available each day and the random number of patients who need an appointment for each day can
be described with a Poisson distribution with a mean equal to (0.008)*(panel size).
A new “open access” system allows doctors to create a short waiting list for nonurgent patients.
All urgent patients are seen the next day, and overtime is used, if necessary, to see all urgent
patients. However if all of the slots are filled on a day, a nonurgent patient may be delayed to the
following day. Overtime is used for nonurgent patients on Friday to clear out the waiting list so
that the doctor sees all patients before the weekend.
Build a simulation model to find the best panel size. The model should determine the doctor’s
profit for one 5-day week.
If a nonurgent patient needs more than one slot, the doctor may take care of some of
those slots at the end of one day and take care of the remaining slots on the following
day. (This is slightly unrealistic, but it introduces only a small error into the model.)
a). Assume that the doctor uses a panel size of 2000 and use 5000 trials for your simulation run.
What is the mean one-week profit?
b). What panel size maximizes profit? What mean profit value does this panel size give you?
(Answer to the nearest 100.)
_____________________ _____________________
______________________________________________________________________________
46
7. Mutual Fund Performance
The Magellan mutual fund had outperformed the Standard and Poor’s 500 index during 11 of the
13 years prior to 1990. It was widely thought that this performance indicated that it was possible
to “beat the market” on a consistent basis. But just how unusual is this performance? A simple
experiment will shed some light on this question.
Consider 50 mutual funds, each of which has a 50% chance each year of performing better than
the Standard and Poor’s 500 (S&P500). Further, assume that each year’s performance is
independent of the all other years’ performance, and that the performance of each fund is
independent of the performance of all other funds.
a. Build a spreadsheet for one mutual fund to calculate the probability that a single mutual fund
will outperform the S&P500 in 11 or more of 13 years. What is the probability that a single
mutual fund will outperform the S&P500 in 11 or more of 13 years? Use 5000 trials.
Probability: ___________________
b. Build a spreadsheet to calculate the probability that the best performer out of 50 mutual funds
will outperform the S&P500 in 11 or more of 13 years. What is the probability that the best
performer of 50 mutual funds will outperform the S&P500 in 11 or more years out of 13 years?
Use 5000 trials.
Probability: ___________________
c. What is the probability that two or more of 50 mutual funds will outperform the S&P500 in 11
or more years out of 13 years?
Probability: ___________________
______________________________________________________________________________
47