You are on page 1of 4

Quiz 1

Q1: A firm has to decide on its advertising campaign for the following quarter. It has three
choices -- television, newspaper and poster advertising. For the limitation of its funds, the
firm can choose only one media. The actual return on the advertising is measured by the
number of potential customers who have the opportunity to see each media. This will
depend on the weather, poor weather favouring television viewing and newspaper
reading, whereas better weather gets people outdoors. The number of potential customer
(in thousands) is given in Table 1. (6 points)
Table 1: The number of potential customer (in thousands)
Weather
Media Poor Moderate Good
Television 22 21 13
Newspaper 18 17 15
Poster 13 16 16

Q1.1: Which media the firm will choose with five criteria for the decision making under
uncertainty (1) The maximax criterion; (2) The maximin criterion; (3) The Hurwicz
criterion (The optimism parameter is l=0.5); (4) The Laplace criterion; (5) The minimax
Regret criterion.
Q1.2: According to the Hurwicz criterion, the firm’s final choice depends on the selected
value for l. At which range of l, the firms will always choice Television?

Q2 The managers of a food company are interested in determining the effect on their sales
of a competitor’s television advertisements. An analysis of sales records for the last 120
weeks gives the following results:
Table 2: Sales records for the last 120 weeks (in no. of weeks)
Level of sales (no. of weeks)
Low Medium High Total
Competitor advertised 32 14 18 64
Competitor did not advertise 21 12 23 56
Total 53 26 41 120

Assuming that these past data are a reliable guide to the future, please determine the
following probabilities: (4 points)
(i) p(next week’s sales will be high)
(ii) p(next week’s sales will be high | the competitor advertises)
(iii) p(next week’s sales will be high | the competitor does not advertise)
(iv) Do the events ‘competitor advertises’ and ‘high sales’ appear to be independent?
And why?

Answer for Quiz 1

Q1:

Q1.1:
Transposed pay-off table
Media Maximal
Weather Television Newspaper Poster Pay-off
Poor 22 18 13 22
Moderate 21 17 16 21
Good 13 15 16 16
maxmax 22 18 16
maxmin 13 15 13
Herwitz 17.5 16.5 14.5
Laplace 56/3 50/3 45/3

Regret Value
Media
Weather Television Newspaper Poster
Poor 0 4 9
Moderate 0 4 5
Good 3 1 0
minmax 3 4 9
Q1.2: sensitivity analysis to l
ET(l)=22l+(1-l)13=9l+13
EN(l)=18l+(1-l)15=3l+15
EP(l)=16l+(1-l)13=3l+13

22

21

20

19

18
Hurwiz

17

16

15

14

13
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
namda

If l∈(1/3,1], Television will be the best choice for the firm.

Q2

Table 2: Sales records for the last 120 weeks (in no. of weeks)
Level of sales (no. of weeks)
Low Medium High Total
Competitor advertised 32 14 18 64
Competitor did not advertise 21 12 23 56
Total 53 26 41 120

Assuming that these past data are a reliable guide to the future, please determine the
following probabilities: (4 points)
(v) p(next week’s sales will be high)=41/120=0.34
(vi) p(next week’s sales will be high | the competitor advertises)=18/64=0.28
(vii) p(next week’s sales will be high | the competitor does not advertise)=23/56=0.41
(viii) Do the events ‘competitor advertises’ and ‘high sales’ appear to be independent?
And why?
p(‘competitor advertises’ and ‘high sales’)=18/120=0.15

If ‘competitor advertises’ is independent with ‘high sales’, according to multiplication


rule, we have:
p(‘competitor advertises’ and ‘high sales’)= p(‘competitor advertises’)×p(‘high sales’)
=(64/120) ×(41/120)=0.18

Since p(‘competitor advertises’ and ‘high sales’) is not equal to p(‘competitor


advertises’)×p(‘high sales’), we have ‘competitor advertises’ is dependent with ‘high
sales’.

You might also like