University of London Study Month: MT2076 Management Mathematics:

Activities and Sample exam questions to be undertaken during weeks 3 & 4

 Discussion Topics:

(i) When trying to decipher large amounts of data from the statement of a
question, use a highlighter pen to highlight the information already
taken into account during the formulation of your solution. In which
types of questions is this suggestion likely to be most useful?
(ii) Looking at the actual UoL exam papers in MT2076 Management
Mathematics over the last three years (including Zone A and Zone B
where possible), analyse how often questions have occurred on every
individual topic (such as complex numbers, or input-output processes,
etc.) except for the “ever presents”: logic problems, index numbers,
forecasting, and differential/difference equations.
(iii) In a Markov matrix problem where steady state probabilities are
required, explain why you should not necessarily use a matrix
inversion method.
(iv) Discuss the four steps of a Box-Jenkins forecasting approach.

 Useful Reading:

One topic that hasn’t appeared very often in past exam papers is Input-Output Processes. The
theory is introduced briefly in section 6.2 of the Study Guide, and a relatively simple example
appeared in question 6 (a) of the 2012 Preliminary Exam. You should read section 6.2 of the
Study Guide, read and attempt the above mentioned question 6 (a), and then read carefully
through the solution to this question.

The L matrix, I-A, is due to Leontieff, and a fuller formulation showing how the Leontieff
matrix L is constructed in general is given in the document

http://econc10.bu.edu/economic_systems/Lecture_notes/Planning/input_output.htm .

 Sample Examination-type questions:

There follows six questions of varying lengths (and therefore marks). These are of final
University of London MT2076 standard. The total mark is 91, so this should take a well
prepared student around two hours and 45 minutes to complete. Usually in the University
exam there will be eight questions, so the average length of the six questions here is a little
long. Feedback on how the questions should have best been answered will be given in session
3 (video released in week 4).
Q1
(a) What is the difference between a “general solution” of a given second order
differential equation and its “complete solution”? (2 marks)

(b) The complete solution of a given second order differential equation in y as a

function of x is y = 0.5x + 1 + 2e-xcos x, where x ≥ 0. Sketch the graph of y, and
comment on its main features. (4 marks)

(c) Consider the second order differential equation D2y + 2Dy – 8y = -16 -27e-x,
where x ≥ 0. Derive its general solution. If y = 10 when x = 0 and Dy = -23 when
x = 0, find the complete solution. Can the value of y ever become negative?
(8 marks)

Q2

68 of the 124 female employees and 74 of the 127 male employees of a small company
responded to a ‘life style’ questionnaire. Two key questions were “do you live within
five miles of work?”, and “do you drive to and from work?”. Unfortunately some of the
information was lost due to a faulty computer hard drive, but the definitive information
that wasn’t lost is as follows.

 Of all 251 employees, 72 females and 60 males live within 5 miles of the company,
and 50 males and 41 females drive to and from work. It is known that 26 of the 127
male employees live within 5 miles of the company and yet drive to and from work.

 Of the 142 employees who responded, 20 of the females drive to work and 27 females
live within 5 miles of the company, whereas 30 of the males live within 5 miles of the
company and only 13 males drive to and from work.

Three data items are missing. (1) The number of females who responded to the
questionnaire who drive to work but do not live within 5 miles of the company; and, (2),
the number of males who did not respond but who both drive to and from work and also
live within 5 miles of the company. These two missing data items are known to be
equal, = x, say. (3) The number of females who did not respond and who live within 5
miles of the company but do not drive to and from work: this is known to be 2x.

(a) Draw a fully annotated Venn diagram of the data for the female employees, and –
separately – a fully annotated Venn diagram for the males. Each set should contain
its order, either as a positive integer or as a simple linear function of x. (7 marks)

(b) Calculate the minimum and maximum possible values of x. Is it possible that 9 of
the 251 employees did not respond to the questionnaire, do not drive to and from
work, and do not live within 5 miles of the company? (6 marks)
Q3

(a) A, B, C, D and E are five internet marketing companies competing for on-line sales of
tennis equipment, such as rackets. In April their respective gross revenues from sales
were £200k, £150k, £200k, £150k and £300k respectively. The total monthly gross
revenue from this market has been steady at £1 million for several months, and is
predicted to remain at this level for several years to come. Over the past few months
each company has been gaining and/or losing gross revenues according to the
following Markov matrix:

From ↓ To → A B C D E
A 0.7 0.0 0.0 0.1 0.2
B 0.1 0.8 0.0 0.1 0.0
C 0.1 0.2 0.5 0.1 0.1
D 0.0 0.0 0.0 1.0 0.0
E 0.3 0.0 0.1 0.1 0.5

(b) Sketch the network diagram and forecast the gross revenues of each company during
May, assuming that the above Markov matrix continues to be relevant from April to
May. In particular verify that the expected gross revenue in May of company A will
increase by over 30 percent. However, without making any further calculations,
would you regard the longer term future for company A to be promising? (8 marks)

(c) Both company A and company C have decided to carry out marketing campaigns, as a
result of which the Markov matrix is expected to change. The modified Markov matrix
is as follows, and is expected to hold for many months to come:

From ↓ To → A B C D E
A 0.8 0.0 0.0 0.1 0.1
B 0.1 0.8 0.0 0.1 0.0
C 0.1 0.2 0.6 0.0 0.1
D 0.2 0.0 0.0 0.8 0.0
E 0.4 0.0 0.1 0.1 0.4

Calculate the exact steady state monthly proportions of the total gross revenue earned
by each company, and the corresponding monthly gross revenues. If any company
will withdraw from the market if its steady state monthly gross revenue looks likely to
fall below £100k, what then should be the steady state monthly gross revenue of
company A? (10 marks)
Q4

(a) In forecasting studies, distinguish between (i) extrapolative models; and (ii) causal
models. (5 marks)

(b) A snack and ice cream parlour close to a sandy beach has been open for 16 weeks. In
weeks 1 and 2, equipment and supplies were being delivered, so sales were quite low,
and from week 3 onwards the parlour has been run by two teams. Team A has run the
parlour during the odd numbered weeks (3, 5, . . . , 15) and Team B has run it during
the even numbered weeks. The table below shows the revenue taken in each week, in
$’000:

Week 1 2 3 4 5 6 7 8
Revenue 1.2 1.7 6.8 5.5 7.4 5.8 7.6 6.3
Week 9 10 11 12 13 14 15 16
Revenue 8.1 6.6 8.0 6.4 8.8 7.1 9.3 7.5

(c) A forecast of the sales in week 17 is required, and Joe Swindle, a statistical
consultant, has suggested using an additive MA2 moving averages model in either or
both of the following cases: (i) using the data of all 16 weeks; and (ii) using the data
of weeks 3 to 16 inclusive. Comment on these two suggestions but do NOT carry out
any calculations. (3 marks)

A second consultant, Jack Better, has suggested instead that an exponential smoothing
approach should be used. He proposes two models: (iii) using the data values of
weeks 3 to 16 inclusive with smoothing constant equal to 0.5 because the successive
values are oscillating; and (iv) using only the odd weeks 3, 5, 7, 9, 11, 13 and 15, with
a smaller smoothing constant, 0.3, because these data values are relatively smooth.

Explain why you feel that model (iv) should be better than model (iii), and use model
(iv) to generate the forecast revenue in week 17. (6 marks)

Q5

(a) Briefly explain each of the three following terms:

A stochastic process; a Markov process; and a (M/M/2) queue. (6 marks)

(b) Define briefly each of the following terms: Euclidean distance; single linkage; and
complete linkage. (5 marks)

(c) Define the form of a pth order autoregressive model of a variable Yt. (2 marks)
Q6

(a) Explain briefly the difficulties involved in constructing a Paasche price index for raw
vegetables in supermarkets in a large country. (5 marks)

(b) The following table shows the quantities in thousands of tonnes of three vegetables, A
(potatoes), B (onions), and C (cabbages), in a large supermarket in the years 2010 to
2013, together with the revenues in hundreds of thousands of dollars.

2010 2011 2012 2013

Quantity Revenue Quantity Revenue Quantity Revenue Quantity Revenue
A 6.0 9.0 7.0 11.2 7.0 12.6 9.0 13.5
B 2.0 3.8 2.2 4.4 2.0 5.0 3.0 6.6
C 3.0 3.0 3.0 3.9 4.0 4.8 5.0 5.5

Construct the Paasche price index to indicate the changing price of food for the years
2010 to 2013 inclusive. (4 marks)

(c) Construct a fixed based total revenue index for the years 2010 to 2013 inclusive.
(3 marks)

(d) The following table shows the RPI (retail price index), with the index = 100 in 2005.

Year 2005 2010 2011 2012 2013

RPI 100.0 120.0 124.0 130.0 133.0

Rebase this RPI to 100 in 2010 and adjust for inflation the total revenue index in part
(c) above. Briefly comment on the resulting calculations. (4 marks)

(e) Consider again part (b) above. Given that a price index was sought, explain why you
do or do not think that a Paasche price index was the best choice. (3 marks)