DXX

I NDIAN I NSTITUTE OF M ANAGEMENT

A HMEDABAD IIMA/QM-xxx

Merton Trucks Case Note

Abstract
We discuss Merton Trucks [Dhe90a] as a case to introduce linear
programming in the MBA program. This case adapted from Sherman
Motor Company case, was used to introduce Linear Programming
formulations as well as duality. Refer to the teaching note [Dhe90b].
Our approach differs from the approach suggested by
Dhebar [Dhe90b]. First, our audience consists pre-dominantly of en-
gineers with not too much work experience. As a result, handling
math and algebra is relatively easy. Explaining the algebraic formula-
tion, graphical approach and using the Excel solver do not consume
that much time. Second, because this case is used during the first
week of the MBA program, students are still unfamiliar with the case
methodology and we spend significant time in understanding case
facts. The circular logic used in allocating fixed costs based on the
product mix that in turn is used in deciding the product mix takes
some time to understand. Third, because of the participant back-
ground, they have difficulty in translating the model to the specific
business situation and interpreting the trade-offs involved in various
what-if analyses that are prompted by the case questions.
We return to the case when we teach duality. After explaining
duality, we analyze the case to show how some of the questions and
what-if analyses can be simplified using duality.
This note is based on our experiences with teaching three large
batches of students in our MBA programs.

Prepared by Ravindra S. Gajulapalli and Diptesh Ghosh.

Teaching material of the Indian Institute of Management, Ahmedabad, prepared as a
basis for class discussion.
Copyright
c 2006, Indian Institute of Management, Ahmedabad.

1
1 Without Duality: (2 × 70) minutes
We use the case to illustrate the following issues:

• Analyze case facts:

– What is the factual basis for some of the opinions expressed?

– Where did the overhead cost numbers shown in Table B come
from?
– Is it appropriate to allocate fixed costs as done in Table C?

• Decision making alternatives:

– Evaluate one product suggestions and current product-mix.

– Evaluate trade-offs and how to improve a solution manually
with its limitations.
– Algebraic formulation of the linear program.
– Graphical solution method and its limitations.
– Solution using Excel solver.

1.1 Objective: 5 minutes

We ask the students to articulate in plain English answers to the following
questions without going into the low level details:

• What is Merton deciding?

• On what should this decision be based?

• What constrains Merton?

1.2 Roles and Rationales: 5 minutes

We would like the students to understand the data that supports the opin-
ions expressed in the case. This is a prelude to analyzing validity of the
data.

2
 • Why do the company president and sales manager feel that 101’s
are making a loss and hence 101’s production should be stopped?
From Table B, it costs $40,205 to produce a 101-truck while it sells for
$39,000.

• Why do the president and the production manager feel outsourcing

engine assembly will help? No slack in engine assembly currently.
Show “Resource Usage” worksheet of merton facts.xls.

• Why does the controller feel that cutting back on 102s is an answer?
Overheads are the answer and they are dealt next.

1.3 Understanding the Exhibits: 15 minutes

The numbers in Table A are straight forward resource utilization num-
bers. The numbers in Table B show total costs including fixed overheads
allocated based on Table C. Allocation of fixed costs are done based on
the ratio of resource usage. The fallacy in allocating fixed costs based on
product mix to decide the product mix can be illustrated by using the “Ex-
hibits” worksheet of merton facts.xls. The grayed out areas represent data
that does not change with product mix. Try the following combinations as
a basis for the controller’s comments:

• (1000, 1500) for negative contributions of 101s and to explain current

exhibit values in Table B,

• (2500, 500) for both positive contributions,

• (2500, 125) for negative contributions of 102s and

• (500, 125) for both negative contributions.

1.4 Relevance of Overheads: 5 minutes

Allocated fixed costs should not be included in deciding the contributions.
The actual costs for (101s, 102s) are:

• direct materials - ($24,000, $20,000) from Table B,

• direct labor - ($4,000, $4,500) from Table B and,

3
 • variable overhead - ($8,000, $8,500) from Table C.

What are the contributions of 101s and 102s? ($3000, $5000).

1.5 Evaluate current plan: 15 minutes

We use the “Evaluate” worksheet of merton facts.xls to set up the model
close to the algebraic formulation discussed next.
Analyzing the president and sales manager’s decision to stop 101s is
easy. The constraints imply that the number of 102s produced should be
given by min[4000/2, 6000/2, 4500/3]. This gives $7.5 million, worse than
the current plan contributions of $10.5 million.
Current plan for productions is (101s, 102s) = (1000, 1500). Is this the
best?
Reduce one model 102 and see how many extra 101s can be produced.
Reducing one 102 frees (2, 2, -, 3) resources in (engine assembly, metals
stamping, 101-assembly, 102-assembly) that can be used to produce two
101s. Show that this increases contribution by $1000 (2 × $3000 − $5000).
How long can you do this? Net resource effect of the substitution of
each 102 is (-, +2, +4, -3) on remaining resources (-, 1000, 3000, -). This
implies that we can change up to a minimum of (-, 1000/2, 3000/4, -) and
non-negativity requirement on number of 102s produced. Explain that it
is tedious to do this kind of analysis, and more than two product-lines
implies many trade-offs.

1.6 Algebraic Model: 5 minutes

How do we describe the model algebraically? Decisions are coded using
variables called decision variables. How do we express the objective and
constraints using these variables?
Decisions variables:

t101 = number of model 101 trucks produced,

t102 = number of model 102 trucks produced.

4
 The algebraic formulation is:

max 3000t101 + 5000t102 Objective: contributions,

1t101 + 2t102 ≤ 4000 Engine Assembly,
2t101 + 2t102 ≤ 6000 Metal Stamping,
2t101 ≤ 5000 101 Assembly,
3t102 ≤ 4500 102 Assembly,
t101 ≥ 0 Non-negativity,
t102 ≥ 0 Non-negativity.

1.7 Develop graphical model: 10 minutes

Show merton trucks.ppt, slides 2 through 13. We show how the trade off
that we talked about in section 1.5 moves along the engine assembly ca-
pacity constraint line. In our first 70 minute session in 4 sections of roughly
75 students each, we reached this point.

1.8 Develop spreadsheet model: 10 minutes

Excel formulation. Set it up on-line using the “Evaluate” worksheet of
merton facts.xls.

1.9 Solutions to Problems: 60 minutes

We ask the students to ignore problem 3(b) since it requires reduced costs.
Problem 1: (10 minutes)

a. Optimal monthly mix is (2000, 1000) with a contribution of $11 mil-

lion. This part is already handled while setting up the spreadsheet
optimization model in section 1.8.

b. What is it worth to add an extra hour of engine assembly capacity?

Unutilized resources (engine assembly, metal stamping, 101 assembly,
102 assembly) = (0, 0, 1000, 1500). To use the extra engine assembly
hour, we can give up a 101 truck and make a 102 truck and the net
increase in contribution for the swap (-1, +1) is $2,000. Net effect on
resources is (+1, 0, -2, +3) for the swap. The effect on a graph can be
seen in slides 12 and 13 of merton trucks.ppt.

5
c. Resolve to verify that change of engine assembly capacity from 4000 to
4100 hours increases contribution by 2, 000 × 100 = 0.2million.

d. The number of swaps that we can do is 500.


1500
min −−, −−, 2000, , = 500.
3

Note 101 assembly does not impose a constraint on the number of

swaps directly because each swap increases the unused resources by
two units. But, the number of swaps is constrained by the number
of 101s being produced currently, i.e., 2000. We can not give up more
resources than what we are using.

Problem 2: (10 minutes) This problem should not take much time but
students get confused with the unfortunate wording used in the case. Pro-
duction manager is talking about reimbursing the outside supplier for ma-
terials, labor and overhead. Hence the students question why are we not
considering the fixed overheads in column two of Table C when comput-
ing the results? Fixed overheads are based on Merton’s infrastructure and
do not hold for the outside supplier. Only the variable overheads are rele-
vant here.
The second doubt that arises in some students, due to a mis-reading
of the case text, is about what is being outsourced. They assume that a
specific model’s capacity is increased by the outsourcing. The text is clear
that we do not outsource any specific model’s engine capacity. Either 101s
or 102s are expected to be made by the outside supplier.
Since we are not using duality, the answer is same as in problem 1 (b)
and 1 (d). Sourcing out engine assembly is acceptable from 1 (b), and the
largest rent Merton can pay is $2000 and no more than 500 hours of engine
assembly can be purchased.
Problem 3: (10 minutes) Model 103:

a. Resources required for 103 = (.8, 1.5, 1, -) and net contribution of $2000.
Re-solve the problem and it is not worth producing. The optimal mix
does not include model 103s and hence no change in contributions.

6
 Decisions variables:

t101 = number of model 101 trucks produced,

t102 = number of model 102 trucks produced,
t103 = number of model 103 trucks produced.

The algebraic formulation is:

max 3000t101 + 5000t102 + 2000t103,

1t101 + 2t102 + 0.8t103 ≤ 4000,
2t101 + 2t102 + 1.5t103 ≤ 6000,
2t101 + 1t103 ≤ 5000,
3t102 ≤ 4500,
t101, t102, t103 ≥ 0.

b. Ignore.

Problem 4: (20 minutes) Overtime production - re-solve. Students take

time understanding that even though the trucks are identical after pro-
duction, the trucks produced during over time and regular time have to
be disambiguated in the model by using extra variables due to differing
engine assembly costs and the capacity constraints of engine assembly.
Decisions variables:

t101 = number of model 101 trucks produced in regular time,

t102 = number of model 102 trucks produced regular time,
o101 = number of model 101 trucks produced in over time,
o102 = number of model 102 trucks produced over time.

The contribution of the trucks produced during overtime reduces to

(2400, 3800) from the regular time contributions of (3000, 5000) because
of 50% extra costs for the engine assembly labor given in Table B. Some
students multiply total direct labor by 1.5 instead of multiplying only the
engine assembly labor component. The labor costs increase to (1800, 3600).
In addition to the variables, a new constraint for overtime engine assem-
bly capacity has to be added and the metal stamping, 101-assembly, 102-
assembly constraints have to incorporate the overtime variables. The full

7
algebraic formulation is:

max 3000t101 + 5000t102 + 2400o101 + 3800o102,

1t101 + 2t102 ≤ 4000,
1o101 + 2o102 ≤ 2000,
2t101 + 2t102 + 2o101 + 2o102 ≤ 6000,
2t101 + 2o101 ≤ 5000,
3t102 + 3o102 ≤ 4500,
t101, t102, o101, o102 ≥ 0.

The optimal solution yields a contribution of $11.7 million with 1500

model 101 trucks and 1250 model 102s produced in regular time and 250
model 102s produced in over time. Hence paying a monthly fixed over-
head of $.75 million is not worth the increase of $.7 million.
Problem 5: (10 minutes) Marketing constraint - re-solve. To the base
model add the constraint t101 − 3t102 ≥ 0. This leads a contribution of
$10.5 million with a plan of (2250, 750).
Instead of requiring the number of 101s produced to be at least three
times the number of 102s, if this constraint was altered to say that it should
be at least two times the number of 102s, then the optimal solution does
not change since the original optimal solution still remains feasible for the
new constraint. Based on this, we generalize about the effect of extra con-
straints on the feasible region.

2 With Duality: 30 minutes

References
[Dhe90a] Anirudh Dhebar. Merton truck company. Case 9-189-163, Har-
vard Business School, HBS Publishing, Boston, MA 02163, apr
1990.

[Dhe90b] Anirudh Dhebar. Merton truck company. Teaching Note 5-

189-171, Harvard Business School, HBS Publishing, Boston, MA
02163, may 1990.