You are on page 1of 5

BUSN 6250 | DASC 6210 | MNGT 4710 • W23 • Case #2

DUE: Friday, March 31st @ 18:00 ß “soft” deadline; hard deadline is Sunday April 2nd @ 23:59
Note: This is a week later than originally set in the Course Outline

You can do this case on your own, with another student, or with two other students — your
choice. In all cases you must sign up into a group using the at Moodle.

Please use the Case2_Starter.xlsx spreadsheet available at Moodle. It will have the
starting information you require for this case. You/your team will submit your final version of
the spreadsheet in the dropbox @ Moodle.

January Tab — Mama Mia Margarine (MMM) is manufactured by refining raw oils and blending
them together. There are two categories of oils — Vegetable Oils (VEG1, VEG2), and Non-
vegetable Oils (OIL1, OIL2, OIL3). The same information presented here is in a wee table on the
January tab in the starter spreadsheet.

The costs to purchase the raw materials in January are:


VEG1 110 ($/ton), VEG2 120, OIL1 130, OIL2 110, and OIL3 115.

The final MMM product is sold to industrial bakeries — the revenue on that is $150/ton.

Use this information to generate an Objective Function with these six decision variables.

Due to differences in refining requirements, in any given month, it is not possible to refine more
than 200 tons of Vegetable Oils (VEG1 + VEG2), and no more than 250 tons of Non-vegetable
Oils (OIL1 + OIL2 + OIL3). For the purposes of this simplistic model, there is no weight loss in the
refining process* and the cost of refining is ignored.

Use this information to create three constraints — two for the maximum refining capacities,
and one “balance” constraint that brings together the raw materials and the final product.

Finally, a key parameter of the Mama Mia Margarine product is that its “hardness” must lie
between 3.0 and 6.0 hardness units. Too hard and it won’t spread without tearing the bread;
too soft and it won’t stick to a knife.

The hardness of the raw materials is as follows:


VEG1 8.8, VEG2 6.1, OIL1 2.0, OIL2 4.2, and OIL3 5.0.

*
*this information is known however — check out the "Weights, Measures, and Conversion Factors for
Agricultural Commodities and Their Products" from the USDA
https://www.ers.usda.gov/webdocs/publications/41880/33132_ah697_002.pdf

Case #2 • W23 • Page 1 of 5


Hardness blends linearly. For example — if you have four units of OIL1 (hOIL1 = 2.0) and two
units of OIL3 (hOIL3 = 5.0), the hardness of the blend would be:
[4(2) + 2(5)] ÷ (4+2) = 18 ÷ 6 = 3.0 ß which would be the minimum hardness.

With your six decision variables, two processing constraints, one balance constraint, and two
hardness constraints (a lower limit and an upper limit) solve this first LP.

You should get this optimal solution:

è As before, I will grade this spreadsheet by setting the DVs to zero and running the Solver
dialog you have created.

Multiperiod Tab — now that your logic is working for one month of margarine production,
we’re going to introduce inventory to the mix. The future prices for the various raw materials
are in the Multiperiod tab of the starter spreadsheet:

The idea is this … buy some OIL2 in February (say) when the price is low ($90/ton), keep it in
storage, then use it later in April (say) vs buying it for $120/ton at that time.

Case #2 • W23 • Page 2 of 5


Of course, we can’t store the raw materials indefinitely and we only have so much capacity!
Each raw material inventory at the start of January is 500 tons. The total storage capacity for
each type of material is 1,000 tons. The cost of storage is 5$/ton/month.

Inventory is managed through balance constraints:


starting inventory + how much is bought = how much is used + ending inventory.

So if we take VEG1 for example, if I had 500 tons to start in January and I bought 10 tons, I have
510 tons to use in the creation of margarine. If I used 159.26 tons, that means I have 350.74
tons in storage at the end of the month ß January’s ending inventory will of course be
February’s starting inventory.

500 + 10 = 159.26 + 350.74 ç the numbers in yellow represent new decision variables for
amount bought, amount used, and amount stored. Those are values of course, let’s abstract
from the numbers and create DV’s like:

BUYRAW,MONTH USEDRAW,MONTH STOREDRAW,MONTH ß double subscripted variables 😉

So, our inventory constraints start to look like this:

January 500 + BUYVEG1,JAN = USEDVEG1,JAN + STOREDVEG1,JAN


February STOREDVEG1,JAN + BUYVEG1,FEB = USEDVEG1,FEB + STOREDVEG1,FEB
March STOREDVEG1,FEB + BUYVEG1,MAR = USEDVEG1,MAR + STOREDVEG1,MAR
Note: you will have to move all DVs to the LHS and just keep a value on the RHS at some point!

: : : : :
June STOREDVEG1,MAY + BUYVEG1,JUN = USEDVEG1,JUN + 500

Mama Mia has a business rule that requires 500 tons of material to be in storage at the end of
June to begin the next 6-month planning cycle.

We now have 90 decision variables — 30 BUYRAW,MONTH, 30 USEDRAW,MONTH, and 30


STOREDRAW,MONTH — since we have 5 RAW materials X 6 months.

Where we had one DV for the final MMM product in the January tab, we now have 6 DVs —
one for each of the amount of MMM produced in each of the 6 months (e.g., MMMJAN,
MMMFEB, …, MMMJUN).

96 decision variables total. Your version of Solver™ is good for 200 😉

Case #2 • W23 • Page 3 of 5


DVs are done so on to Objective Function.

Where we had the revenue for MMM less the cost of the raw materials before, now we have:
• The revenue for the 6 months of MMM’s,
• Less the cost of the raw materials bought (e.g., 130 X BUYOIL1,MAR, etc.) — use all 30 of
these costs,
• less the cost of storing the product @ 5$/ton/month. For example, if STOREDVEG2,APR =
40.74 tons, we’ll incur a cost of $203.70.

Last bit — constraints:

The processing and hardness constraints remain as before, but they too are month-wise. So
where we had “it is not possible to refine more than 200 tons of Vegetable Oils (VEG1 + VEG2)”
— we now have “it is not possible to refine more than 200 tons of Vegetable Oils (VEG1 + VEG2)
in January. It is not possible to refine more than 200 tons of Vegetable Oils (VEG1 + VEG2) in
February”, etc.

We have 30 new constraints requiring that all of the STOREDRAW,MONTH values don’t exceed the
1,000 tons storage capacity. And five constraints to set the final STOREDRAW,JUN values to 500
tons.

If you can get your multiperiod model working, you’ll end up with a total profit of $107,842.59.

PSYC 1110 Tab — one of the most popular undergrad courses at any university is the
“Introduction to Psychology” course — and TRU is no different.

This term there are 6 sections (A-F) of around 60 students. For final exam scheduling purposes,
you can’t have 60 students in a 60 seat classroom — you need to have a separation between
them — so the 6 sections are split into two A1 & A2. A1 includes students with surnames
starting A-L, and A2 is surnames M-Z. Ditto for the other five sections.

è See the data in the starter spreadsheet. Section C (for example) has 57 students total; 29 fall
into the first half of the alphabet (so C1 is 29) and the remaining 28 are in the last half of the
alphabet (so C2 is 28).

Fifteen classrooms have been tentatively allocated to be used for this exam — the capacities
given are “1/2 capacities” — where OM 1007 might normally seat 60 students, it has capacity of
30 for the final exam.

You now have 180 decision variables — just slightly below the capacity of the free version of
Solver™

Case #2 • W23 • Page 4 of 5


The person in charge of final exam scheduling wants to get as close a “match” between the
number of seats in the exam room as there were to students. For example, if the 30 students of
in the first half of section A (A1) were assigned to the 30 seats in OM 1007, it would be a
perfect match. If those same 30 students were assigned to the 62 seats in OM 1220, the “cost”
would be 32 — and would remove that room for consideration by use by another course.
Conversely, trying to squeeze those 30 students into the 25 seats in OM 2015 would be an
impossible task.

1) Formulate a linear programming model for this assignment problem.


Hint: use a compact notation with terms like Xij to simplify your work. Be explicit about
what those terms mean!

2) Translate your LP model into Excel and solve using Solver. What is the optimal
assignment? How many left-over empty seats are there with this arrangement?

3) Suppose the Scheduler wanted to assign the half-sections to classrooms so that the
number of classrooms (and hence number of proctors) was minimized — instead of the
number of left-over seats. How does the problem change? Try and come up with a
solution to this on a new tab.

Centreville tab — Centreville Amusement Park is a small amusement park on the Toronto
Islands catering (mostly) to families with young children. In a typical summer season, the
number of adult tickets sold has a uniform distribution between 10,000 and 20,000 tickets. The
number of children’s tickets sold has a normal distribution with a mean of 20,000 and a
standard deviation of 2,000. Adult ticket prices are $20 and the children’s price is $10. Revenue
from food and beverage concessions is estimated to be between $50,000 and $100,000, with a
most likely value of $60,000. Likewise, souvenir revenue has a minimum of $20,000, most likely
value of $25,000, and a maximum value of $30,000. Variable costs per person are $5, and fixed
costs amount to $150,000. Determine the distribution of profit for this business using 250 trials.
What is the probability that the park will make a profit of $250,000 or less?

The first few columns are filled with 250 rows of =RAND() functions to generate uniformly
distributed random numbers between 0.000 and 0.999. Transform these random numbers into
random variates with the requisite characteristics per the description above, and use those as
inputs to a profit model to answer the probability question and summarize your results.

I will grade your work two ways — (i) I will open your spreadsheet, press the F9 key to
recalculate it and see what changes, what doesn’t and (ii) then I will paste a predetermined set
of random numbers into the columns and I will check your solution values against a known set
of data. DO NOT MESS with these columns of data.

As always — Questions? Post ‘em to the Moodle Discussion Board…

Case #2 • W23 • Page 5 of 5

You might also like