Solutions Manual

Solutions Manual
This document contains solutions to all the practice questions that appear at the end of each
chapter.
Chapter 2 Solutions
1.
Year Cost Index

(000s)
2006 50.1 100.0
2007 65.3 130.3
2008 68.6 136.9
2009 72 143.7
2010 76.6 152.9
2011 78.3 156.3
2012 88.7 177.0
2013 90.5 180.6
2014 99.3 198.2
2015 112.9 225.3
2.
Year Cost () Index (Year 1 = 100)

1 650 100
2 580 89.2
3 410 63.1
3.
Calculating the index for 2015 using 2013 as the base year
Resource Index (2013 = 100)

Ink 138
Card 160
Labour 107
So card has increased the most (up 60%)
4.
Essential Quantitative Methods: For Business, Management and Finance 2016 Les Oakshott
For more resources visit http://www.palgrave.com/oakshott-eqm6
Calculating both indices for 2014 based on 2013
2013 2014
p0 q0 pn qn p0q0 pnq0 p0qn pnqn
Ink 0.55 4000 0.58 5000 2200 2320 2750 2900
Card 0.05 115000 0.05 124000 5750 5750 6200 6200
Labour 6.5 25 6.9 34 162.5 172.5 221 234.6
8112.5 8242.5 9171 9334.6
8242 9334.6
Laspeyres index = 8112.5 100 = 101.6 Paaschs index = = 101.8
9171
For 2015 based on 2013
2013 2015
0.55 4000 0.76 3000 2200 3040 1650 2280
0.05 115000 0.08 110000 5750 9200 5500 8800
6.5 25 7 30 162.5 175 195 210
8112.5 12415 7345 11290
12415 11290
Laspeyres index = 8112.5 100 = 153.0 Paaschs index = = 153.7
7345
From 2013 to 2014 there has been very little change as prices have not changed much.
However during 2014 to 2015 prices for ink and card have shown a big increase.
5.
2005 2015
Bread 28 6 78 6 168 468 168 468
Milk 20 15 150 12 300 2250 240 1800
Tea 96 1 75 2 96 75 192 150
564 2793 600 2418
Laspeyres' 495.21 Paasche's 403.00
6.
1998 2011
Company 160 200 520 500 32000 104000 8000 260000
A 0
Company 350 650 265 250 227500 172250 8750 66250
B 0
Company 105 600 140 400 63000 84000 4200 56000
C 0
Company 53 100 159 200 5300 15900 1060 31800
D 0
327800 376150 2201 414050
00
Laspeyres' 114.75 Paasche's 188.12
7. Simple index for year 2 using year 1 as the base year
Supervisor: 14/12 x 100 = 116.7
Skilled: 10/9 x 100 = 111.1
Unskilled: 7/6 x 100 = 116.7
For year 3
Supervisor: 15/12 x 100 = 125.0
Skilled: 11/9 x 100 = 122.2
Unskilled: 8/6 x 100 = 133.3
So unskilled increased wages the most.

Year 2 based on year 1
Supervisor 12 10 14 12 120 140 144 168
Skilled 9 120 10 160 1080 1200 1440 1600
Unskilled 6 200 7 200 1200 1400 1200 1400
2400 2740 2784 3168
Laspeyres index 114.2 Paasche index 113.8
Year 3 based on year 1

Supervisor 12 10 15 15 120 150 180 225
Skilled 9 120 11 160 1080 1320 1440 1760
Unskilled 6 200 8 180 1200 1600 1080 1440
2400 3070 2700 3425
Laspeyres index 127.9 Paasche index 126.9
Both indices very similar and show a steady increase in total wages over the 3 years.
8.
186.7
105,000 is worth 105,000 214.8 = 91,264
So price has dropped by 140,000 91,264 = 48,736
9.
Year Turnover RPI Real T/O

2005 15.3 192.0 15.30
2006 10.3 198.1 9.98
2007 12.1 206.8 11.23
2008 15.2 214.8 13.6
2009 24.4 213.7 21.9
2010 34.7 223.6 29.8
10.
2003 2011
p0 q0 pn qn p0q 0 pn q 0 p0 q n pn q n
7.50 120 9.00 158 900 1080 1185 1422
10.00 41 12.50 52 410 512.5 520 650
8.00 25 10.00 30 200 250 240 300
18.00 21 22.40 25 378 470.4 450 560
sum 1888 2312.9 2395 2932
Laspeyres 122.51 Paasches 122.42

11.
2006 2011
p0 q0 pn qn p0q 0 pn q 0 p0 q n pn q n
3.63 3 4.49 2 10.89 13.47 7.26 8.98
2.11 4 3.26 6 8.44 13.04 12.66 19.56
10.03 1 12.05 1 10.03 12.05 10.03 12.05
4.01 7 5.21 5 28.07 36.47 20.05 26.05
57.43 75.03 50.00 66.64
Laspeyres 130.65 Paasches 133.28
12(a)
Month Price Index
Aug 155 100.00

Sept 143 92.26
Oct 120 77.42
Nov 139 89.68
Dec 165 106.45
Jan 162 104.52
(b)(i) August to January change = 104:52 - 100 = 4:52%

165143
(ii) September to December change = 143
= 15.38%
13.
Average Average Wages deflated

Year wage RPI to 2008 values
2008 255.1 214.8 255.1
2009 271.3 213.7 272.7
2010 290.7 223.6 279.3
Real wages grew more between 2008 and 2009, due to the
decline in the RPI.
14.
Year Average RPI House price Real house price
2008 214.8 162 000 162 000
2010 223.6 157 500 153 030
Real drop in price = 11 179 at 2010 prices, i.e. 6.9%

15
(a) (b) (i)
(ii) (iii)
% change
Average RPI since since
Year RPI (2008 = 100) 2008 2009
2008 214.8 100
2009 213.7 99.5 -0.5
2010 223.6 104.1 4.1 4.9
Chapter 3 Solutions
1.
Year Population Sample
Total Female Male Total Female Male
1 320 192 128 20 12 8
2 250 150 100 16 10 6
3 230 138 92 14 8 6
Total 800 480 320 50 30 20
2.
Need to ask council tax payers. A survey could be sent out with council tax demands but that
may produce a bias sample as only those people with strong views will respond. Maybe better
to send questionnaire to random sample and follow up with letters. Could also allow residents
to use online questionnaire.
3.
The survey will probably be in the form of a simple questionnaire, possibly web based. Not
necessary to have a random sample for this type of survey so anyone who applied for tickets
could be asked to complete the questionnaire.
4 Various issues including asking age (use ranges), Vague questions (question 5), no
option for other in question 6. Unlikely that respondents would know what socio-economic
group they are in.
5 1, 2 and possibly 5 would have a sampling frame.
6 (a) The sample would be Steve, Kim, Chris, Jane, Stuart, Jill. Average age is 40.3
and 3 read the Mirror, one the Sun, one The Times and one the Express.
(b) As we have 6 males and 3 females we need a sample of 4 males and 2 females.
If we separate out the two sexes and use the same random numbers for each group then we
will have Steve, Chris, Stuart, Kim, Julie and Jill. Average age is 39.2 and two read the Sun,
two the Mirror and one each of The Times and Telegraph.
(c) Julie
7. Answers depend on sample chosen.
8.
Target population would be all council tax payers. A good sampling frame would be a
database of council tax payers in a local council region. Alternatively the electoral register
could be used although this will contain people not paying council tax (e.g. children over 18
living with parents or at university).
9.
Not all supporters belong to the supporters club so their views will be ignored.
A simple random sample may not contain the right proportion of male/female
supporters or the different ages.
A better target population may be all people who have ever watched a game (if this
data is kept) or even all people in the local town.
A stratified sample would be best as it would take into account male/female and ages
issue.
A systematic sample may also be possible as every nth person leaving the ground
could be sampled.
10.
Probably use multi-stage sampling methods. Randomly select a number of council regions
then randomly select a number of schools within each region. A random sample of school
leavers could then be selected from each chosen school.
11.
The simplest and cheapest survey would be a quota sample of shoppers in the town centre.
However, this would ignore people who dont shop in the town. An alternative would be an
advert in a local newspaper, but this would bias towards those who read the paper and have a
strong opinion. The most expensive would use the electoral register to contact a stratified
sample of voters in the town.
12 (a) Both stratified and quota sampling aim to produce a sample that represents the
target population in terms of the proportion contained within relevant sub-groups. However,
stratified sampling requires the use of a sampling frame and is therefore a probabilistic
sampling method.
Although it would be possible to obtain a list of all students it would probably be sufficient to
employ a quota sampling method as this will be much cheaper to administer. The proportions
of male/female and age ranges should be easy to obtain, which would make the quota sample
reasonably accurate.
(b) No, since all students would be on campus.
(c) Possible answers include:
Misleading or ambiguous questions. Could be overcome by a pilot study.
Asking the wrong question.
Asking personal questions.
Asking leading questions.
Making the questionnaire too long.
Difficulty in obtaining information about the student population.
Method used to conduct survey. Face-to-face will probably be most effective but this
may be too expensive.
Poor response rate. This can be improved by face-to-face interviews.
Bias sample if only certain members of the population are included in the sample (e.g.
if only interview respondents in a particular part of the university).
Conducting survey during vacations or exam periods!
13.
(a)
Randomly select 40 employees from the database of all 200 employees

Choose every 5th employee from the list
For the shop-floor department there are 80 employees which is 40% of the total so we
need a random sample of 40% of 40 = 16 employees from shop-floor. The other
departments are calculated in a similar way
Dept Number Proportion of total Sample size
Shop floor 80 .4 16
Service engineers 15 .075 3
Quality control 20 .1 4
Marketing 25 .125 5
Accounts 15 .075 3
Personnel 10 .05 2
Administration 25 .125 5
Catering 10 .05 2
Total 200 1 40
Randomly select the required number from each department
For quote sampling a survey could be conducted in the works canteen using the above
numbers as the quota for each department; i.e. once the required number of employees
has been chosen from one department no one else from that department would be
selected
14 (a)
Target population all first year students and all staff.
Sampling frame database of student and staff records.
Stratified sampling sample contains same proportion of relevant categories as
population.
Multi-stage sampling hierarchical structure. If the students were based on different
campuses it might be necessary to pick one or two of these campuses randomly to
reduce travelling time.
(b) Possible questions are:
Are you student or staff? (It will probably be necessary to be able to compare the
responses between staff and students.)
How strongly are you in favour of a 3rd semester? using a Likert scale. (This will be
better than asking for a yes or no answer.)
How much would you be prepared to pay/receive to do a 3rd semester? (Will need
to ascertain how financially viable the idea is. Students will need to pay in order for
lecturing staff to be paid.)
At what faculty are you based? (There may be differences between faculties.)
For students only What is your age (within a range)? (Older students might be
more interested in completing their degrees quicker.)
(c) A postal or email questionnaire will probably be the best method here. Alternatively,
could be face-to-face if quota sampling used.
15. There are many reasons such as:
The bins are being refilled when empty so difficult to judge how empty the bin is.
Were all the bins filled to the same level?
Do people prefer different colours?
Positioning of the bins might make a difference. Do you take the nearest regardless of
your views?
Do people associate colour with flavours?
And, of course, this was not a random sample!
Chapter 4 Solutions
1.(a) Continuous as there could be a fraction of an hour. However if the answer is obtained by
counting the number of hours of sunshine then it could be classed as discrete.
(b) Continuous
(c) Continuous
(d) Ordinal
2. (a) Time spent online
(b) Download speed
(c) Number of websites visited
(d) Rating given to a site
(e) Choice of social networking site
3.
Class interval Frequency

0 to 2 5
3 to 5 4
6 to 8 2
9 to 11 2
12 to 14 3
15 to 19 0
20 to 30 2
A bar chart could be drawn as below
This bar chart is not very good as it doesnt take into account the different class intervals.
(See the solution to question 8 for a better chart).
4.
(a)
Type of coffee Frequency Relative frequency (%)
Instant 8 40
Filter 8 40
Ground 4 20
(b), (c) See Excel file (Chapter 4 Q4)
5.
See Excel file (Chapter 4 Q5)
Component bar chart shows that sales have increased since 2009 but this has been the result of a good
performance by the Menswear department. Furniture sales were the largest proportion of total sales in
2006 but this proportion has declined since then.
6. (a), (b), (c), (d) See Excel file (Chapter 4 Q6)
(e) Total sales declined during 2007 to 2009 but some recovery since then. South East gives
largest proportion of sales with Wales the smallest. Multiple bar chart shows that the decline in South
East sales occurred later but the recovery has not started despite improvement in other regions. Wales
has shown very little change during the 5 years.
7. (a)
Weight (gm) Frequency

20-<22 2
22-<24 8
24-<26 16
26-<28 17
28-<30 21
30-<32 16
(b)
(c)
2 20 45
2 21
5 22 148
10 23 25588
16 24 055688
26 25 2223455688
33 26 1225588
(10) 270000124468
37 28 0000122568
27 29 00022234466
16 30 000024446688
4 31 0444
20|4 20.4 gms
(d)
Could use this chart to find the median and quartiles
(e)
Data varies from 20 to 32 gm. The mean is 27.2 gm and the median is a little higher at around
27.5 gm. The distribution is left skewed
8.
As the class widths vary it would be easier to plot the frequency density. This is achieved by
dividing each frequency by the class interval (see table below). As this is discrete data the
ends of each group will be 2.5, 5.5, 8.5, 11.5, 14.5 and 19.5. A sketch of the histogram is
shown below (note it is not possible to use SPSS to draw this histogram).
Class interval Boundaries Frequency Frequency density

0 to 2 5 2.5
3 to 5 2.5 4 2
6 to 8 5.5 2 1
9 to 11 8.5 2 1
12 to 14 11.5 3 1.5
15 to 19 14.5 0 0
20 to 30 19.5 2 0.2
9.
The histogram shows that the vast majority of people work around 40 hours but the range is from
almost zero to nearly 100 hours. The histogram is fairly symmetrical.
10.
This clearly shows that over 80% of households have at least one car.
11.
This bar chart clearly shows that the majority of family type is still the traditional married
couple with children.
12.
(a)
(b)
(c)
(d)
The No qualifications is the highest category.
Male and female are approximately the same on all categories except the other
qualifications where males there are more males.
Levels 2 and levels 4/5 have about the same total percentage (just under 20%).
13.
Assuming that this is discrete data then the ends of each group will be 2, 5.5, 9.5, 13.5 and
21.5. As the class intervals vary from 2 to 7 it is easiest to divide each frequency by the class
interval to get the frequency density as in the table below.
Days off work Boundaries Number of Frequency density

employees
Less than 2 2 45 22.5
2 to 5 5.5 89 29.7
6 to 9 9.5 40 13.3
10 to 13 13.5 25 8.3
14 to 21 21.5 5 0.7
22 to 29 29.5 2 0.3
The distribution is right skewed.
14. (a) (b) See Excel file (Chapter 4 Q14) Comments should include:
General decrease in sales from 1990 to 1992.
The average sales growth has fallen fastest during 1990 to 1991.
Inner London shows least reduction.
South Eastern shows largest reduction.
South Western and Outer London are only regions to show some recovery during 1991 to 1992.
15. (a) See Excel file (Chapter 4 Q15) for histogram. The distribution is right skewed.
(b) See Excel file (Chapter 4 Q15) for ogive.
(i) About 20% of sales are in excess of 70
(ii) About 62% (84 - 22) of sales are between 52 and 72
(iii) The value of purchases exceeded by 10% of sales is about 75.50
(These answers can be seen on the ogive in the Excel file.)
Chapter 5 Solutions
1.
(a) Mean is 189.53 and median is 174.34. The mean is influenced by the high value of
274.50 so the median is probably the better average to use.
(b) Standard deviation is 48.83 (based on n1)
(c) The mean is increased by 20 to 209.53 but the standard deviation stays the same at
48.83
(d) The mean increases by 5% to 199.01 and the standard deviation also by 5% to 51.27
2.
(a) Mean = 24.5 0C, Median = 24.5 0C Mode = 19 0C
A holiday maker normally takes a holiday during the summer so an average annual
temperature is not that helpful.
25 0.5 0.25
27 2.5 6.25
29 4.5 20.25
30 5.5 30.25
33 8.5 72.25
36 11.5 132.25
Total = 511
Variance = 46.45454545
SD = 6.815757145
(b) Standard deviation is 6.82 0C
(c) The mean should be 26.5 0C; the standard deviation will not change.
(d) The mean temperature in 0F is found by using the formula on the mean and is 76.10. The
standard deviation is found by multiplying by 9 and dividing by 5 only and is 12.27 0F
3.
(a) Median because it tells the students where in the class they are placed. Mean might also
be useful in looking at the distribution of scores.
(b) Both median and mean. Median is useful as it tells you that 50% of people earn less than
this amount. The mean is generally the average quoted.
(c) Mean as it can be used in statistical analysis (quality control).
(d) Mode (the most frequent sold) is the only sensible average.
(e) Mean as it can be used in statistical analysis (quality control).
4.
Number Frequency
of rejects f
x fx x2 fx2
0 12 0 0 0
1 45 45 1 45
2 36 72 4 144
3 30 90 9 270
4 20 80 16 320
5 5 25 25 125
6 0 0 36 0
Sum 148 312 904
Mean = 2.1, standard deviation = 1.3
5.
(a) Mean and median are 1.0794 and 1.0730 respectively.
(b) Standard deviation is 0.0339
(c) This data is a time series and as the exchange rate appears to be increasing during the 10
days the mean and median represents the exchange rate half way through the period.
6.
The mean and median are around 40 hours and the modal class 40 to 50 hours. If the
distribution is assumed normal then the range is about 80 which represents 6 standard
deviations so one standard deviation will be 13.3.
7.
Average mark = 74.4 + 56.6 = 63.2%
8.
Total for 12 light bulbs is 12180.6 = 2167.2
Total for 13 bulbs = 2167.2 + 200 = 2367.2
Mean for 13 is 2367.2/13 = 182.1
9.
(a) 5.3 days
(b) 4 days
(c) 2 to 5 days
(d) 5.5 days
(e) 4.3 days
(f) 82%
(g)
2 4 7.5
0 10 20 30
The distribution is right skewed.
10. (a)
A B
mean 24.8 26.2
median 25 26.3
mode 24.26 26.28
(b) IQR = 2.5 (A); 2.7 (B)
SD = 1.96 (A); 2.04 (B)
(c) A: 7.9% B: 7.8%
(d)
Machine B
24.9 26.3 27.6
Machine A
23.5 25 26
(e)
Machine B produces item with slightly longer lengths than Machine A
The spread of lengths is about the same on both machines
The distribution of lengths is approximately symmetrical although for Machine A the
distribution has a slight left skew since the median is to the right of centre of the box
and is greater than the mean.
11.
Mean = 57.67 Median about 58.80
12.
The mean wage is 134.18
Proposal 1: Increase in pay = 10% of current mean wage = 13.42
This will cost the company 98 x 13:42 = 1315:16
The median wage is about 130
Proposal 2: Increase in pay = 12.5% of current median wage
= 0:125 130 = 16:25
This will cost the company 98 16:25 = 1592:50 = approximately 1600.
13. The purchasing manager has simply calculated the mean of the four percentage figures:
10+4+0+46
= 15%
4
However, the mean increase in costs must be calculated as a weighted mean, i.e., weighting each
raw material percentage price increase by the mean expenditure on that raw material.
Thus, the mean increase in costs is:
51000
= 10.022%
5000
The purchasing manager unwittingly gave far too much emphasis to the huge percentage
increase in Additives, which in fact accounts for a very small proportion of total
expenditure.
14.
Production target is 300 widgets per week
Total production target in all 13 weeks = 13(300) widgets.
Now let x = average to be reached in final 3 weeks in order to achieve target. Production in
first 10 weeks = 10(284.7) widgets.
Production in final 3 weeks = 3x widgets.
Total production in all 13 weeks = 10(284.7) + 3x widgets.
We want:
Total production = Total production target.
i.e.,10(284.7) + 3x =13(300)
Solving for x, we get x = 351
Thus an average of 351 widgets per week must be produced in the final 3 weeks in order to
achieve the original target of 300 widgets per week over all 13 weeks.
15.
100 100
Total time to cover all 200 miles = + = 2.9167 hours
80 60
200
Average speed over the 200 miles = 2.9167 = 68.57 mph
16.
(a) (i) Total sales = 6745
(ii) Mean = 26.66
(iii) Standard deviation = 19.60
(iv) coefficient of variation = 19.6/26.6 100 = 73.7%
(b) Mid points used
(c) (i) Less variation in second shop
(ii) Could be either
(d) (i) Median = 20
(ii) Q1 = 12, Q2 = 34, IQR = 34 - 12 = 22
(iii) Over 50 = 11%
Chapter 6 Solutions
1. 0.25
2. 0.125
3. 0.2778
4. There are 5 ways of getting 8 and 36 combinations altogether so probability is 5/36 =
0:1389
5. P(first ace) = 4/52 = 0:07692;
P(second ace) = 3/51 = 0:05882;
P(third ace) = 2/50 = 0:04;
P(all 3 aces) = :07692 x 0:05882 x 0:04 0:00181
26 25 24
6. P(3 reds) = 52 51 50 = 0.1176
7.
(a) (i) 5/10 = 0.5 (ii) 3/10 (iii) 7/10
(b) (i) (ii) 3/20
(c) (i) 2/9 (ii) 1/6
8.
(a) 4/52 (b) 26/52 (c) 4/52 + 26/52 28/52 = 28/52
9.
4/50 3/49 = .00490
10.
(a) .05.05.05 = .000125
(b) If P(sale) = S= .05 and P(no sale) = S = .95
Then either SSS or SSS or SSS will give exactly 1 sale and the probabilities for each of these
is the same so.
P(1 sale) = 3.05.95.95
= .1354
(c) P(at least 1 sale) = 1 P(no sales)
= 1 - .953
= .1426
11.
Let R = respond and R dont respond (= .8)
(a) All 8 respond = (.2)8 = 2.5610-6
(b) No one responds = (.8)8 = 0.1678
(c) There are many combinations; for example RRRRRRRR (in fact there are 28 and in
Chapter 8 you will see that this is in fact a binomial situation).
P(exactly 2 respond) = 28 (.2)2(.8).6 = .2936
12.
(a) subjective
(b) empirical
(c) a priori
(d) subjective
13. The 3 machines operate independently, so for example the probability of Machine A
breaking down in independent of the probability of Machine B breaking down. Thus we use
the multiplication rule for independent events:
P(A and B) = P(A) P(B)
(a) P(all three machines will be out of action)
= P(Machine A will be out of action)
P(Machine B will be out of action)
P(Machine C will be out of action)
= 0:1 0:05 0:20
= 0.001 or 0.1% or 1 in 100
(b) P(none of the machines will be out of action)
= P(Machine A will not be out of action)
P(Machine B will not be out of action)
P(Machine C will not be out of action)
= (1 0.1) (1 0.05) (1 0.20)
= 0.90 0.95 0.80
= 0.684 or 68.4%
14. We assume that bad weather and geological problems are independent events.
Thus we again use the multiplication rule for independent events:
P(A and B) = P(A) P(B)
P(project will be delayed by both bad weather and geological problems)
= P(project will be delayed by bad weather)
P(project will be delayed by geological problems)
= 0.30 0.20
= 0.06 or 6%
15. P(both plugs will be substandard)
= P(first plug will be substandard)
P(second plug will be substandard j first plug was substandard)
6 5
19= .0789 or 7.89%
20
16. P(shares will rise in value)
= P(shares will rise in value FT30 index rises) P(FT30 index rises)
+ P(shares will rise in value FT30 does not rise)
P(FT30 index does not rise)
= 0.8 0.6 + 0.3 (1 0.6)
= 0.48 + 0.12 = 0.60 or 60%
17.
(a) .9 .6.7 = .216
(b) P(at least 1) = 1 P(none)
= 1 (.1.4.3)
= .988
c) .012
18. (a) (i) P(bulb defective) = 0.3
(ii) P(bulb defective bulb is 60 w) = 20/60 = 0.333
(b) The probability that a box will be accepted is the sum of the probabilities on the Accept
branch ends:
P(box is accepted) = 0.4879 + 0.1494 + 0.1494
= 0.7867 or 78.67%
19.
P(research says high) = .12 + .32 = .44
P(Hresearch says high) = .12/.44 = .273
20.
.9
Has Condition
.02
.2
8 Test +ve
.04
.05
.76
P(test +ve) = .18 + .04 = .22
P(test ve) = .02 + .76 = .78
P(not have condition|Test +ve) = .04/.22 = .1818
P(have condition|Test ve) = .18/.78 = .2308
Chapter 7 Solutions
5
1. C3 = 10
2. P(3H) = 10 .53 .52 = 0.3125
3.
(a) Either get a job or doesnt so 2 states and probabilities are constant and independent.
(b) P(r>2) = 1 [P(0) + P(1) + P(2)]
P(0) = (.52)9 = .00278
P(1) = 9.48(.52)8 = .02309
P(2) = 36(.48)2(.52)7 = .08527
P(r>2) = 1-.1111
= .8889
4(a) P(male) =105/205
= .5122
(b) P(3 males) = 4C3 (.5122)3 .4878
= .2622
4
(c) P(2 males) = C2 x (.5122)2 (.4878)2
= .3746
5.
Want to find the probability of two or more defective items given a defective rate of 3%
P(at least 2 defectives) = 1 (P(0) + P(1))
P(0) = 20C0 (.03)0 (.97)20
=.544
P(1) = 20C1 (.03)1 (.97)19
= .336
P( 2) = 1 - .544 - .336
=.12
So there is a 12% change of getting two or more defectives from a batch of 20. This is not
particularly surprising.
6. Probability of egg undamaged = .8
(a) P(0) = 4C0 (.8)0 (.2)4
= .0016
(b) P( 3) = 1 [P(0)+ P(1) + P(2)]
P(1) = .0256
P(2) = .1536
P(3) = 1 (.0016 + .0256 + .1536)
= .8192
7. (a) P(< 4 calls) = P(0) + P(1) + P(2) + P(3)
P(0) = e-6:7 = .00123
Using the recursive properties of the formula:
P(1) = :00123 x 6:7 = .00825
P(2) = .00825 x 6.7/2 =.02763
P(3) = .02763 x 6.7/3 = .06170
P(< 4) = .00123 + .00825 + .02763 + :06170 .0988
(b) P(> 7) = 1 P( 7)
P(4) = .1033; P(5) = .1385; P(6) = .1546; P(7) = 1480
P(> 7) = 1 (.0988 + .1033 + .1385 + .1546 + .1480) = .3568
(c) Mean number of calls in 1 minute = 6.7/5 = 1.34
8. This is a Poisson problem.
(a) Mean number of flaws 1500/1000 = 1.5 per mm
P(2) = 1 [ P(0) + P(1)]
P(0) = e-1.5 = .2231
P(1) = .2231 x 1:5 = .3347
P(2) = 1 [.2231 + .3347) = .4422
(b) Need to find the probability of no flaws in 5 mm.
Mean number of flaws in 5 mm is 5 x 1:5 = 7:5
P(0) = e-7:5 = :0006
9.
(a) P(0) = e-3.2 = .04076
(b) P(r>4) = 1-[P(0) +P(1)+P(2)+P(3)]
P(1) = .13044
P(2) = .20870
P(3) = .22262
P(r>4) = 1 - .6025 = .3975
(c) In 8 minutes m = 6.4
Using Tables P(r<4) = .2351
10.
0.1056
11.
.0735 .0071 = .0664
12.
1 (.2514 + .1469)
= .6017
13.
65
(a) = = .667
1.5
P(Z>.667) = .2514 (from tables)
Area = .5-.2514 = .2486
45
(b) = = .667
1.5
Same area as in part (a), that is .2486
(c) area = 2.2486 = .4972
6.55
(d) For 6.5, = = 1.0
1.5
P(Z>1) = .1587
7.55
For 7,5 = = 1.67
1.5
P(Z>1.67) = .0475
Area = .1587 - .0475
= .1112
14.
2500020000
(a) = = .694
7200
P(Z>.694) = .2451
So probability that demand greater than 25000 litres is .2451 or about 24.5%
1700020000
(b) = = .417
7200
P(Z<-.417) = .3372
So P(demand>17000) = 1-.3372
= .6628 or about 66.3%
(c) P(20000<demand<25000) = .5 - .2451
= .2549 or about 25.5%
(d) For demand = 30000
3000020000
= = 1.39
7200
P(Z>1.39) = .0823
For demand = 35000
35000 20000
= = 2.08
7200
P(Z>2.08) = .0188
P(30000<demand<35000) = .0823 - .0188
= .0635 or about 6.4%
15.
For 80000
80000 45121
= = 1.438
24246
P(Z>1.438) = .0749 which is the probability that someone will earn more than 80000
For 100,000
10000 45121
= = 2.26
24246
P(Z>2.26) = .0119
So number of staff earning more than 100,000 = 500000.0119
= 5950
16.
60 48
= = 1.2
10
P(Z>1.2) = .1151
P(Call answered in less than 60 seconds) = 1 - .1151) = .8849
So there is approximately a 88.5% chance of the call being answered in less than a minute
(note in theory this probability includes time less than zero but error is negligible).
17.
Need to find the probability less than 4.5 mm and greater than 6 mm.
4.5 5.5
= = 2.5
.4
P(Z<-2.5) = .0062
6 5.5
= = 1.25
.4
P(Z>1.25) = .1056
P(between 4.5 and 6 mm) = 1-(.0062 + .1056)
= .8882
So, about 88.8% of bolts will be accepted.
18.
(a) (i) Undersize
5054
= = -1
2
P(Z<-1) = P(Z>1) =.1587
So 15.87% are undersize
(ii) Oversize
Z = .5
P(Z > :5) = .3085
So 30.85% will be oversize
(b) Out of 1000 parts, 53.28% will be usable and 308.5 on average will need to be
shortened. The 158.7 parts that were too short will have to be remade. Total cost will
therefore be:
1000 x 10 + 158:7 x 10 + 308:5 x 8 = 14 055
(Note: this is an underestimate as the 158.7 parts that need to be remade will also contain
some rejects.)
19. For 5% in the upper tail, Z = 1.645
Let x be the minimum customer spend to get a free gift
135
1.645 =
55
x = 225.48
If the upper tail is now 8% then Z = 1.405 and
225.48
1.405 = 55
= 225.48 55 1.405 = 148.21
20.
(a)
300 320
(i) Z 1.92
10.4
P(Z < 1.92) = 0.0274
325 320
(ii) Z 0.481
10.4
P(Z > 0.481) = 0.3156
(iii) P(318 <x< 325) = 1 ((P(x<318) + P(x>325))
318 320
P(x<318) = P(Z< 0.192 )
10.4
= 0.4247
P(318 <x< 325) = 1 (.4247 + .3156)
= 0.2597
(b)
500 0.0274 = 13.7 (around 14 packets would be underweight)
(c) The Z value corresponding to an area of .03 is 1.88
300 310
1.88

10
5.32
1.88
21. Let x be the demand for the product.
(a) We want to find the probability that x is greater than 2500
2500 2000
= =1
500
P(Z > 1) = 0.1587 from the normal table
(b) x > 2800
2800 2000
= = 1.6
500
P(Z > 1:6) = 0.0548
(c) x < 1600
1600 2000
= = 0.8
500
P(Z < -0.8) = P(Z > 0.8) by symmetry
= 0:2119
22.
130 100
(a) Z 2
15
P(Z>2) = .0228 or 2.28% will spend over 130
120 100
Z 1.33
(b) 15
(Z>1.33) = .0918 or 9.18% will spend over 120
70 100
(c) Z 2
15
P(Z<-2) = .0228 or 2.28% will spend less than 70
(d) (100<x<130) = 50 2.28
= 47.72%
(e) For 115
115 100
Z 1
15
P(Z>1) = .1587 or 15.87%
(115<x<130) = 15.87 2.28
= 13.59
(f) For 10% in the tail Z = 1.28
x 100
1.28
15
x = 151.28 + 100
= 119.20
(g) Z = 1.88
x 100
1.88
15
x = 128.20
23.
Let x be the IQ scores.
115 100
= =1
15
P(Z > 1) = 0.1587
(b) x > 130
130 100
= =2
15
P(Z > 2) = 0.0228
(c) 115 < x < 130
P(115 < x < 130) = P(x > 115)- P(x > 130)
= 0.1587 0.0228
= 0.1359
24. Let x = number of gallons.
Need to find the value of x that gives an area of 0.06 in the right hand tail. From tables, Z
=1.555
2000
1.555 = =1
500
x 2000 = 500 1.555
x = 2338 gallons
25.
(a) Find area less than 15.75 and greater than 16.75 g and add to give required proportion
15.75 16
Z 0.5
0.5
P(Z<-.5) = .3085
16.75 16
Z 1.5
0.5
P(Z>1.5) = .0668
So required area is .3085+.0668
= 0.3753
So 37.53% of output will be scrapped
(b)
Number scrapped = 10000.3753 = 3753
Cost = 37537 = 26,271
(c) Need to recalculate the Z values
15.75 16
Z 1.25
0.2
P(Z<-1.25) = .1056
16.75 16
Z 3.75
0.2
P(Z>3.75) 0
So number scrapped = .105610000 = 1056
Cost of scrapping = 10567 = 7392
Total cost including new machine = 7392 + 10,000 = 17,392
As the combined cost is less than 26,271 it would be worth hiring the machine.
26.
Need to find the chance of profit being less than 0
For product A
0 15
Z 1.67
9
P(Z<-1.67) = .0475
Product B
0 10
Z 2
5
P(Z<-2) = .0228
So should choose product B
27.
For the Poisson approximation the value of p should be small and the number of trials large.
Both are true. We need the mean which is 5. If we use Excel or tables we get the probability.
( x> 8) = 1 - .8666
=0.1334
Which agrees well with the binomial answer of 0.13
For the normal approximation np and n (1-p) should be greater than 5. np is the mean
and is exactly 5
The standard deviation is
= 100 .05 .95
= 2.179
Now need the probability that x is greater than 7.5 (because of the continuity correction)
7.5 5
=
2.179
= 1.147
Using tables this is a probability of .1251
So again a good approximation even though the conditions were only barely met
28.
m = 2.65
2.65 2.658
P(r=8) =
8!
= 0.00426
So, a very low probability.
Chapter 8 Solutions
1.
(a) 16,000 (b) 50 (c) 20/50 = 40% (d) 40%
4060
(e) SEP = 50
= 6.9
95% confidence interval = 40 1.966.9
= 26.5% to 53.5%
2. =9:62 g
= 0.7731
0.7731
Sem = = 0:3156
6
The value on t degrees of freedom = 2.571
95% confidence intervals = 9.62 2:571 x 0.3156
= 9:62 .81
= 8.81 g to 10.43 g
3.
Sem = 0.26
Confidence interval = 169.51.960.26
= 168:99 cm to 170.01 cm
4. P = 75%
7525
SEP = = 5:5902
60
95% confidence interval = 75 1.96 x 5.5902
= 75 10.96
= 64% to 86%
5.
Mean = 8.72 g standard deviation = 1.2523 g
Sem = .3960,
(a) t = 2.262
95% confidence interval: 8.72 2.2620.3960
= 7.82 g to 9.62 g
(b) t = 3.250
99% confidence interval: 8.72 3.2500.3960
= 7.43 g to 10.01 g
3675
6. (a) SEM = = 245:0
225
99% confidence interval = 16 450 2.575 x 245
= 16 450 631
= 15 819 to 17 081
3675
(b) 2.575 = 200

2.575 3675
= = 47.316
200
n = 2239
So a sample of about 2240 is required.
7.
P = 30%
Sep = 7.2457
Need to apply the finite population correction factor as population is small. This is 0.8967
So Sep becomes 7.24570.8967 = 6.497
95% confidence interval is 30 1.966.497
= 17.3% to 42.7%
320
8. P = 800 100 = 40%
4060
SEP = 800
= 1.732
= 40 1.96 x 1.732
= 36.6% to 43.4%
9.
P = 80%
Sep = 2.8284
95% confidence interval is 80 1.962.8284
= 74.5% to 85.5%
10. n = 400
= 230
= 42
4.2
SEP =
400
= 2.1
(a) 95% confidence interval is given by
230 1.96 x 2.1 = 230 4.12 or 225.1 to 234.1
(b) 99% confidence interval is given by
230 2.58 x 2.1 = 230 5.42 or 224.6 to 235.4
11. = 60 g
Margin of error required = 2
Margin of error = 1:96 x SEM

2 = 1.96 x

60
2 = 1.96 x

Square both sides
602
4 = 1.962 x
602
n = 1.962 x 4
= 3457
12.
1.96SEP = 4%
(a) P=10%
10 90 4
=
1.96
Square both sides
900
= 4.1649

n= 216
(b) If no estimate then use P=50%
As before then n = 600
13.
55524
SEM = 16
= 13881
(a) 95% confidence interval is 129500 1.9613881
= 129 500 27207
= 102 293 to 156 706
(b) Half width needs to be 13604
55524
So 1.96 = 13604

So n = 64
(note to half the confidence interval need to multiple n by 4 so 416 = 64
(c) The Finite population correction factor needs to be applied
(20016)
Which is = 0.9616
199
This is multiplied by SEM which results in a smaller half width ( 26161 compared to
27207) so the confidence interval will be narrower
14.
P = 51%
5149
SEP = 1975 = 1.1249
Confidence interval = 51 1.961.1249
= 48.8% to 53.2%
As the confidence interval straddles 50% we cannot be certain that there would be a majority
in favour of leaving the EU.
Chapter 9 Solutions
1. 1.96
2. 718 (assuming a single sample)
3. 12.592
4.
H0: = 5.8 H1: < 5.8
(This is a one-tailed test as it is required to see if the action has reduced the time taken.)
SEM = 2.3/10
= 0.7273
As is known we can use the Z test
4.95.8
Z= 0.7273
= -1.237
The critical value at 5% is -1.645 and as this is less than the test statistic then we cannot
reject H0 and therefore there is no evidence that the mean time has been reduced.
5.
H0: = 54 H1: 54
SEM = 5/12
= 1.4434
As is known we can use the Z test
50.554
Z= 1.4434
= -2.424
The critical value at 5% is -1.96 so can reject H0 and conclude that there is evidence that
the fuel consumption is not 54 mpg
6.
H0: = 1500 H1: 1500
SEM = 90/50
= 12.728
Can use the Z test as the sample size is large
14101500
Z= 12.728
= -7.07
The critical value at 5% is -1.96 so can reject H0 and conclude that there is evidence that
the mean lifetime is not 1500 hours. (Note as no one would complain if the lifetime was
greater than 1500 hours it could be argued that this is a one-tailed test. However, we
would still reject the null hypothesis.)
7.
H0: = 500 H1: < 500
(This is a one-tailed test as it is required to see if the mean weight is less than 500 g.)
From the sample the mean = 497.3 g and the standard deviation = 11.448 g
SEM = 11.448/6
= 4.674
497.3500
t= (Use the t test statistic as the sample size is small and we are estimating
4.674
the standard deviation from the sample)
= -0.578
The critical value at 5% is -1.645 and as this is less than the test statistic then we cannot
reject H0 and therefore there is no evidence that the company is selling underweight
products.
8.
H0: = 30% H1: > 30%
P = 60/150 = 40%
3070
SEP = 150
= 3.7417
4030
Z = 3.7417
= 2.67
The critical value at 5% significance level is 1.645 so as the test statistic is greater than
this can reject H0. Therefore we can conclude that the claim by the Building Society is
correct.
9.
This is a goodness of fit test.
H0: Accidents equally likely on any day H1: Accidents more likely on certain days
If H0 is true there should be 37/5 = 7.4 accidents each day on average
O E (O-E) (O-E)2 (O-E)2/E
6 7.4 -1.4 1.96 0.264865
5 7.4 -2.4 5.76 0.778378
6 7.4 -1.4 1.96 0.264865
8 7.4 0.6 0.36 0.048649
12 7.4 4.6 21.16 2.859459
Sum 4.216216
Degrees of freedom = 5-1 = 4
Critical value on 4 degrees of freedom at 5% significance level = 9.488
As this is greater than the test statistic of 4.216 we cannot reject H0 and conclude that
there is no evidence that accidents are more likely on certain days.
10. (a)
H0: There is no association between age and radio station people listened to
H1: There is an association
The expected values are
Age range
Less than 20 20 to 30 Over 30 Total
BBC 32.4 15.4 40.1 88
Local radio 12.2 5.8 15.1 33
Commercial 18.4 8.8 22.8 50
Total 63 30 78 171
O E O-E (O-E)2 (O-E)2/E
22 32.4 -10.42 108.60 3.35
6 12.2 -6.16 37.92 3.12
35 18.4 16.58 274.86 14.92
16 15.4 0.56 0.32 0.02
11 5.8 5.21 27.15 4.69
3 8.8 -5.77 33.32 3.80
50 40.1 9.86 97.21 2.42
16 15.1 0.95 0.90 0.06
12 22.8 -10.81 116.79 5.12
Sum 37.50
Degrees of freedom = (3-1)(3-1) = 4
Critical value on 4 degrees of freedom = 9.488
As the test statistic is greater than the critical value we can reject H0 and conclude that
there is an association between age and radio station people listened to.
(b) The largest contribution to the test statistic is the number of people under 20 who
listened to a commercial station. Only 18.4 would be expected from the null hypothesis
but the observed value was nearly twice this at 35
11.
H0: = 10 H1: 10
It is a two tailed test so critical values at 5% significance level are 1:96, at 1%
they are 2:57 and at 0.1% they are 3:29
n = 36, = 9.94:94, = 0:018
0.018
SEM = = 0.003
36
9.9410
Test statistic Z = = -20.0
.003
Since Z < -3.29 we reject H0 and conclude that there is very strong evidence at the 0.1%
level that the production process is not meeting specification.
12.
H0: = 36% H1: 36%
80
n = 200, P = 200 x 100 = 40%
3664
SEP = = 3.3941
200
4036
Test statistic Z = 3.3941 = 1.18
Since 1.18 is less than 1.96 we cannot reject H0 so there is therefore no evidence to suggest
that there has been a significant change in the percentage of people who bought the product.
13.
H0: = 40000 H1: 40000
n = 40; = 37 000; = 6000
6000
SEM = = 948.68
40
3700040000
Test statistic Z = = -3.16
948.68
Since -3.16 is less than -2.57 we can reject H0 at the 1% significance level. We therefore
conclude that there is strong evidence that the mean takings were significantly different from
the target figure.
14.
H0: = 60% H1: 60%
250
n = 500, P = 500 x 100 = 50%
6040
SEP = = 2.1909
500
5060
Test statistic Z = 2.1909 = 4.56
Since -4.56 is less than -3.29 we can reject H0 so there is very strong evidence to suggest that
the success rate is significantly different from 60%.
15.
Hypothesis test is inappropriate here as we are given population information not sample
information. Thus we can say with certainty that the mean size of order has increased from
last year.
16.
H0 :A = B H1: A B
t = 2:228 on 10 degrees of freedom (two tailed so 0.025 in each tail
= 5:517 SA = 1:507; -x B = 4:567 SB = 1:750

5 (1.507)2 + 6 (1.750)2
=
10
= 1:633
1 1
Standard error = 1:633 6 + 6
= 0.9428
5.5174.567
t= 0.9426
= 1.008
Since t is less than 2.228 H0 cannot be rejected so it is not possible to use these results to
decide on the best filter.
17.
H0 : Consumers equally likely to select each package
H1 : Consumers not equally likely to select each package
Package O E O-E ( )2

A 106 100 9 0.36
B 92 100 -8 0.64
C 102 100 2 0.04
D 100 100 0 0
1.04
Test statistic 2 = 1:04
5% significance level, df = 4 - 3 = 1, 2 = 7:815 (from tables) Accept
H0
Results are consistent with the hypothesis that the consumers are equally likely to select each
package.
18.
<21 2135 >35 Total

Liked 20 40 80 140
Disliked 30 20 10 60
Total 50 60 90 200
H0: No association between age and preference
H1 : There is an association between age and preference
O E O-E ( )2

20 35 -15 6.428
40 42 2 0.095
80 63 17 4.587
30 15 15 15.000
20 18 2 0.222
10 27 -17 10.703
37.037
Test statistic 2 = 37:037
5% critical value, df = (2 1)(3 1) = 2, from tables 2= 5:991
Reject H0
Evidence to suggest that age and preference are associated (would also reject at 1% and 0.1%
levels)
More under-21s and less over-35s than expected appear to dislike the design.
19.
H0:Pattern of demand hasnt changed
H1 Pattern of demand has changed
O E O-E (O-E) 2 (O-E) 2/E
Direct lines 120 180 -60 3600 20
Switchboard 80 75 5 25 0.33
Fax and others 100 45 55 3025 67.22
Sum 87.56
The critical value on 2 degrees of freedom at 5% SL= 5.991
Reject H0 and conclude that there has been a significant change in the pattern of demand.
20.
H0 : D = 0
H1 : D > 0
(one tailed since it is hoped that the campaign would not decrease sales)
Critical value of t = 1:895 at 5% (one tailed on 7 degrees of freedom)
Differences are:
A B C D E F G H
28 -25 40 100 60 -20 0 50
= 29.125
SD = 42.633
42.633
Standard error = = 15.073
8
29.125
= 15.073 = 1.932
Since 1.932 is greater than 1.895 H0 can be rejected at 5%. Therefore it seems likely that the
campaign has worked.
21.
This is a one tailed test and the null and alternative hypotheses are:
H0 :1 2 = 0 H1 : 1 2 < 0
30
P1 = 250 100
= 12%
40
P2 = 220 100
= 18:2%
250 12 + 220
=
250 + 220
= 14:9%
The standard error of the differences is
1 1
(1 2 ) = 14.9 (100 14.9)( + )
250 220
= 3:2917
The test statistic is:
(12 18.2) 0
=
3.2917
= - 1:88
As this is less than -1.645 we can reject H0 and therefore conclude that there is evidence to
suggest that students are working longer in 2000 than in 1995
22.
H0:Lunch habits hasnt changed
H1 Lunch habits have changed
O E O-E (O-E)2 (O-E)2/E

Bring
lunch 42 88.53 -46.53 2165.041 24.45545
Skip lunch 95 27.24 67.76 4591.418 168.55
Local shop 27 40.86 -13.86 192.0996 4.70
Canteen 63 70.37 -7.37 54.3169 0.77
Sum 198.48
The critical value on 3 degrees of freedom is 7.815
Reject H0 and conclude that there has been a significant change in lunch habits since
moving to the new site. In particular the number of employees who skip lunch has
increased greatly and the number who bring their lunch has gone down
23.
H0:Demographic mix hasnt changed
H1 Demographic mix has changed
O E O-E (O-E)2 (O-

E)2/E
Less 18 6 6.80 -0.80 0.65 0.10
18-19 118 75.35 42.65 1819.19 24.14
20-24 102 134.57 -32.57 1060.67 7.88
>24 26 35.28 -9.28 86.12 2.44
Sum 34.56
Critical value on 3 degrees of freedom = 7.815
Reject H0 and conclude that there has been a significant change in demographic mix of
the users. This is caused by the number of people in the 18-19 age group using the site
more.
Chapter 10 Solutions
1.
45
40
35
Earnings (000s)
30
25
20
15
10
5
0
0 20 40
Age 60 80
There is a weak positive correlation between age and Earnings as shown by the scatter of
points on the chart above. This means that as someone gets older they might be paid more but
there is a lot of variation. Age range would be from 18 to 65.
2.
50
45
40
Consumption (units)
35
30
25
20
15
10
5
0
0 10 20 30 40 50 60 70
Age
Older people tend on average to consume less alcohol than younger people but this is
nowhere near a perfect correlation.
3.
x is already ranked. y is ranked as follows
y 9 6 7 9 5 2 1
Rank 6.5 4 5 6.5 3 2 1
As two values of y are equal we rank them as 6.5
x y d d2
1 6.5 -5.5 30.25
2 4 -2 4
3 5 -2 4
4 6.5 -2.5 6.25
5 3 2 4
6 2 4 16
7 1 6 36
Sum 100.5
6100.5
R = 1 7(491)
= 1 1.795
= 0.795
4.
Stats M Acc d d2
A 4 5 -1 1
B 3 1 2 4
C 1 3 -2 4
D 5 4 1 1
E 2 2 0 0
F 6 6 0 0
Sum 10
610
R = 1 6(361)
= 1 0.286
= 0.714
5.
y = 3 + 52
= 13
6.
(a) and (b) See Excel file (Chapter 10 Q6) for scatter graph.
The scatter graph shows there is strong positive association between height and weight for
males but much weaker for females.
(c) All data: r = 0.760; males: r = 0.906; females: r = 0.0102
7.
R = -0.3875
8.
See Excel file (Chapter 10 Q8) for scatter graph.
(a) The scatter graph suggests that there is a strong association between sales and
advertising expenditure and that this is a positive association.
(b) y = -0.24 + 9.292x, where y is sales and x the advertising expenditure.
The coefficient of determination is 93.3% so it appears that the regression explains the data
quite well.
(c) 46 200
9.
(a) Positive (weak for league as a whole but probably strong for a particular club)
(b) Negative, strong
(c) Positive, weak
(d) Positive, strong
(e) Negative, strong
10.
(a) (i) See Excel file (Chapter 10 Q10)
(ii) r = 0.8898
(b) a = 11.129, b = 0.04037
So time to pay = 11:13 + 0:0404 size of bill
(c) The 11.13 is the minimum time to pay (in days) and for every 1 increase in the size
of bill the time to pay will increase by 0.0404, i.e. the marginal increase.
(d) See Excel file (Chapter 10 Q10)
(e) (i) for 125 time to pay = 11.13 + 0.0404 125
= 16.18 or just over 16 days
(ii) for 1000 time to pay = 11.13 + 0.0404 x 1000
= 51.53 or 51.5 days
The prediction for 1000 is unreliable because it is outside the range of data (the largest
amount is only 480).
(f) 79.2% of the variation in time to pay is explained by the size of the bill. The
remaining 20.8% is not explained by this factor. (Other factors might help in explaining this
remaining variation.)
11.
(a) See Excel file (Chapter 10, Q11)
(b)
% Collected Rank % arrears Rank d d2
1 6 13 1 5 25
12 3 10 4 -1 1
16 2 8 5 -3 9
11 4 12 2 2 4
20 5 11 3 2 4
6 5 11 3 2 4
Sum 68
668
R = 1 6(62 1)
= 1-1.943
= -0.943
12.
(a) See Excel file (Chapter 10 Q12)
(b) Pearsons correlation coefficient, r = -0.5717, Spearmans correlation coefficient = -0.515
(c) H0: r = 0 H1 : r 0
Critical value on 8 df = 0.6319 for Pearsons so cannot reject H0. There is no evidence of a
significant correlation.
(d) (ii) 17.4 is for someone aged 0 and so has no practical meaning. The -0.2118
means that for every year older the days off work reduces by 0.2118
(iii) 17.4 0.2118 65 = 3.6 days. But 65 is outside the range of the data and the prediction
is unreliable.
(iv) r2 = 0.3268 or 32.7%. This tells us that only about 33% of the variation in days off sick
is explained by age.
(v) Length of service, sex, marital status, health, etc.
13.(a)
Advertising x Sales y xy x^2

50 1 50 2500
51 1.02 52.02 2601
60 1.3 78 3600
65 1.45 94.25 4225
45 1.2 54 2025
49 1.08 52.92 2401
55 1.25 68.75 3025
Sum 375 8.3 449.94 20377
7 449.94 375 8.3

=
7 20377 (375)2
= 0.0184
8.3 375
= 0.0184
7 7
= 0.199
The regression equation is therefore
Sales = 0.199 + 0.0184Advertising
(b)
(i) Sales =0.199 + 0.018440
= 0.935m
(ii) Sales =0.199 + 0.018480
= 1.671m
The prediction for 80,000 pounds advertising is outside the range of the date and therefore
not reliable.
14.
SPSS was used for this question.
(a)
This shows a strong positive correlation.
It looks quite a weak correlation. If the point corresponding to size of store = 12 is removed
there may be no correlation.
Very weak correlation. Could be negative.
(b)
Correlations
Sales Population size of store Years experience
Sales Pearson Correlation 1.000 .937** .699* -.365
Sig. (2-tailed) .000 .024 .300
N 10.000 10 10 10
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
Correlation for population is as expected. Other variables are affected by outliers.
(c)
For all variables
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) -4.152 22.351 -.186 .859
Population 4.759 .949 1.154 5.013 .002
size of store -5.574 4.374 -.288 -1.274 .250
Years experience -1.427 2.339 -.079 -.610 .564
a. Dependent Variable: Sales
R2 = .909
For population only

Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) -12.183 16.726 -.728 .487
Population 3.864 .509 .937 7.589 .000
a. Dependent Variable: Sales
R2 = ..878
The regression equation with the highest R2 is: Sales = -4.152 + 4.759P 5.574S 1.427Y
Where P = population and S= size of store and Y = years experience.
However neither size of store or years experience are significant, so the most reliable
equation is Sales = -12.183 + 3.864P
(d) Using the full equation
Sales = -4.152 + 4.75928 5.5746 1.42712
= 112.78 or 112,780
Using the shorter equation
Sales = -12.183 + 3.86428
= 96.01 or 96,010
15.
H0 0 = 0
H1 0 0
From question 13 we know that x = 375 and x2 = 20377 and the slope is .0184
We obtain Se from the table below
Advertising Sales y Predicted Residual Residual sq

x
50 1 1.119 -0.119 0.014161
51 1.02 1.1374 -0.1174 0.01378276
60 1.3 1.303 -0.003 9E-06
65 1.45 1.395 0.055 0.003025
45 1.2 1.027 0.173 0.029929
49 1.08 1.1006 -0.0206 0.00042436
55 1.25 1.211 0.039 0.001521
Sum 0.06285212
.062852
Se= 5
= 0.1121
.1121
Sb =
20377(375)2
7
= .00661
.01840
t= .00661
= 2.784
The critical value on 5 degrees of freedom is 2.571 so we can reject H0 and state that the
regression slope is significantly different from zero
The confidence interval for the slope is
.0184-2.571.00661 and .0184+2.571.00661
.00141 to .0354
1.
Emv = .22 + .8.5
= 0.8
2 (a) 0.4; (b) 26; (c) 14; (d) Decision 1; (e) 0.509
3.
(a) (i) Manufacture (ii) Sell (iii) Take royalties
(b) Manufacture (emvs are 30, 28, 20)
(c) Find the value of perfect information (30 - 80.2+40.5 + 20.3) = 12 (or 12,000)
4.
(a) (i) Bar (ii) Bar (iii) Bar
(b) Bar (emvs are 53, 68.5, 114.5)
5.
(a) (i) B (ii) C (iii) C
(b) C (emvs are 2.7, 3.3, 4)
(c) Calculate value of perfect information = 4 - 5.5+7.3 + 4.2 = 1.4 (or 1.4m)
As this exceeds 0.5m it would be worth employing the economist.
6.
EMV = 2000.25 + 1200.5 + 500.25 = 1225
So he would make a profit of 225
If he invested the 1000 for 3 months he would get 1000 8.75/4 = 21.875
As this is less than 225 he should buy the car.
7.
Expected cost of taking the car = 1000.02 + 10 = 30
As this is less than 50 his decision on financial grounds should be to take the car.
If let the probability of being caught = x then indifferent when
1000x + 10 = 50
x = .04 which is a 2 in 50 chance.
8.
(a) Best decision is to wait
(b) Probability should decrease to 0,5
9. 19m 10m
Rival store
0.7
2
0.3
Expand E
x No rival
40m
p store branch
a
n
d 20.8m 16m
Rival store
a
t
Move t
0.7
3
o
w 0.3
n
No rival
c 32m
store branch
e
n
t
r
e To maximize expected profit,
M
o
v
move to the new site
e
t
o
n
e
10 (a) 20
9.7 Fails
0.01
Earthquake
9.97 0.99
.1
Strengthen 10
.9
12.9
10
15
Fails
6 0.7
Tut Dont Earthquake
12.9 0.3
.1
1
.9
15
Fles 5
Dictator
11 .2
8 15
.
The best decision is to invest in Tutamolia but not to strengthen the building, giving an EMV
of 12.9m.
(b) - 6p + 15(1 p) = 11
21p = 4
p = 0:19
If probability greater than 0.19, then should choose Flesomnial. (Note: This wont affect
decision for Tut branch as the 9.97 value will be reduced.)
11. (a) (i) This would be the maximax rule and would choose Investment
(a)(ii) This would be the maximin rule. Max(-100, 0, - 10) = 0. Would
choose Investment B.
(b)
Investment Rising Stable Falling Largest

A 0 25 150 150
loss
B 60 0 50 50
C
160 40 0 160
Min(150, 60, 160) = 60 so choose Investment B
(c) EMV A = -25, B = 28 and C = 30 so choose Investment C
(d) The expected value with perfect information
= :2 x 150 ~+ :2 x 50 + :6 x 50 = 70, so EVPI = 70 - 30 = 40
(e) (i) u(150) = 1. u(-100) = 0; u(90) = .5; u(25) = :3
1
1.2
1.0
0.8
0.6
0.4
0.2
-150 100 50 50 100 150 200
(ii) u(50) = .4. u(10) = .25; u(0) = .2; u(-10) = .15
Expected utilities are therefore:
A: .2 x 1 +.2 x .3 + .6 x 0 = .26
B: .2 x .5 + .2 x .4 + .6 x .2 = .3
C: .2 x .15 + .2 x .25 + .6 x .4 = :32
So the best investment is still C.
12.
(a)
NPV of High demand = 300.909 + 400.8264 + 400.7513 20 = 884
Similarly NPV for Low demand = 473
(b)
2
H 854
710 .6
success 5
443
699
employ
H
699 551 720
.55
Internet
699 H
345
.55
884
581
H
473
.55 750
375
Best decision is to launch internet site but not to employ consultants. EMV is 699,000
(d) Maximum amount is 699 678 = 21,000
(e) Find probability that makes you indifferent between two decisions. If it takes only a small
probability to change your decision then the decision is sensitive to this probability.
13 (a)
3
Joint probabilities
.009
positive
.9
allergic .1
.001
.01
positive
.99
.297
.3
.7
.693
.1
(b) b) P(test positive) = .009 + .297 = .306
(c) P(allergicpositive) = .009/.306
= .0294 (ii)
P(allergicnegative) =.001/.694 :0014

(d)
positive
-11
-11
2.85
.306
kit allergic
9.7 8.96
-21
A1
9.7 .0014
-20
.01 9
1.50
10
So should sell the A1 product, but dont supply kit in order to maximise the EMV.
(e) If supply kit annual profit = 500,0002.85 =1,425,000

If dont supply kit profit = 50,0009.7 = 4,850,000
4
If sell older version profit = 50,0001.50 = 750,000
14.
We first use Bayes theorem as follows:
Joint probabilities
0.18
Positive
0.9
Has condition 0.1

0.002
Negative
0.02
Positive
0.98
Doesnt 0.049
0.05
0.95
Negative
0.931
P(Positive response) = 0.018 + 0.049

= 0.067
P(Negative response) = 1 0.067
= 0.933
0.018
P(Condition+ve response) =
0.067
= 0.2687
0.018
P(Condition-ve response) =
0.933
= 0.0193
+ve 350
88.1 5
.067
.067 Condition
screen 1050
69.3
-ve .0193
.9807
Doesnt
50
Dont
screen 65
So cheaper not to screen.
15.
Car Consumption Top Environmental Agg.

speed Benefits
A 75 64 50 65.4
B 0 100 0 15.8
C 100 0 100 84.2
D 58 71 60 60.9
E 92 79 30 70.1
weight 100 30 60
normalised 52.6 15.8 31.6
The car with the highest aggregate benefits is car C.
16.
Possible criteria could be floor area, parking spaces, closeness to town centre, attractiveness
of area, etc.
6
100.0
90.0 C
H
80.0
Efficient A
70.0
frontier
60.0
50.0 B
40.0
30.0
20.0
10.0
0.0
5 7 9 11 13 15 17 19 21
Cost (000s)
Should consider shops A, B, C and H
17. (a)
1.4m
0.24 1.5m H
H .85
.3 1.13
a
L c L
.7 Full
(0.3m) 1.13 .15
(0.4m)
Full
2
F/H
0.269 0.269 .3
Survey (0.1m) 1.4m
1 b
F/L H
(0.1) (0.13) .15
.7 Full
3 d
Abandon L
.85
(0.4)m
0
F/H Forecast high
L/H Forecast low
(0.1m)
7
The decision should be to conduct a survey first and then if the survey indicates a favourable
response go into full production, otherwise abandon project. The EMV would be 0.269 m.
(b)
If the probability is greater than 0.35, decision changes from survey first to full production. If
the probability is less than 0.08 decision changes from survey first to abandon project.
8
1.
Year Investment A Investment B Investment C

0 -20 -15 -5
1 12 0 3
2 10 10 3
3 5 5 3
4 3 5 3
5 0 5 3
Payback 2 3 2
Average income 6 5 3
ARR 30% 33.3% 60%
Investment A and C would be preferred in terms of payback period while investment C is a
clear winner in terms of ARR.
2.
Payback first project: 3 years 2nd project: 5 years
ARR 1st project 1/2.5 100 = 40%
2nd project: the average income is 15/5 = 3m so ARR is 3/10 100 = 30%
The first project is preferred on both the payback and ARR methods.
3.
Month Property A Property B Property C

0 -200,000 -300,000 -400,000
1 400 500 800
2 400 500 800
Payback 200,000/400 = 500 600 mths =50 years 500 mths = 41.7
mths = 41.7 years years
9
Average income pa 40012 = 4800 6000 9600
ARR 4800100/200,000 = 2% 2.4%
2.4%
Property A and C are best (and equal) on both measures.
4.
Using the simple interest formula
10,000 6% = 600
5.
20,000 5% = 1000
6.
Using the compound interest formula
P5 = 20000(1+.05)5
= 25,525.63
7.
(a) P6 = 50000(1+.055)6
= 68942.14
(b) 68942.14 50000 = 18942.14
8.
Using the constant repayments formula
P5 = 3000(1.08)5 + 1000[(1.08)5 1]/.08
= 4407.98 + 5866.60
= 10,274.58
10
9.
Using the present value formula
P0 = 50000 discount factor
Discount factor = 1/(1.06)6
= 0.70496
P0 = 500000.70496
= 35248.03
10.
The answer depends on the interest rate. The table below gives the decision for different
discount rates.
r% 20,000 25,000
2 19,607.84 24,029.22
4 19,230.77 23,113.91
6 18,867.92 22,249.91
8 18,518.52 21,433.47
20 16,666.67 17,361.11
25 16,000.00 16,000.00
As you can see the 20,000 in 1 years time is the better option unless the interest rate
exceeds 25% which is highly unlikely.
11.
The sum of the present values over the 10 years is 27,026.07
12.
Year Present value

0 -2
1 0.566
2 0.534
3 0.504
4 0.475
5 0.448
NPV 0.527
11
As the NPV is positive this investment is worthwhile.
13.
Year Investment A Investment B Investment C

PV PV PV
0 -20 -20 -15 -15 -5 -5
1 12 11.32 0 0 3 2.83
2 10 8.90 10 8.90 3 2.67
3 5 4.20 5 4.20 3 2.52
4 3 2.38 5 3.96 3 2.38
5 0 0.00 5 3.74 3 2.24
NPV 6.80 5.79 7.64
All 3 investments show a positive NPV but investment C gives the greatest value so is
preferred on this measure.
14.
Payback period = 4 years
ARR = 25%
NPV = 0.05m
IRR = 8% (Note to find IRR you can use GoalSeek in Excel to find the interest rate that gives
the NPV of zero)
15. See Excel file (Chapter 12 Q15). The Goal Seek function can be used to find the value
of r to give a NPV of zero. This was found to be 9.7% and as this is less than 14% the project
should not be accepted.
16. P0 = 1000, r = 5:4%, n = 5
5.4
P5 = 1000 (1 + 100)5
= 1000 x 1.30078
= 1300.78
12
17.
Using the sinking fund formula (assume that the customer pays the first annual instalment
now giving 4 instalments in total). Let x be the annual instalment.
First instalment grows to x(1.0575)3 = 1.1826x at the end of 3 years
2nd instalment grows to x(1.0575)2 = 1.1183x at end of 3 years
3rd instalment grows to x(1.0575) at end of 3 years
Final instalment is x
Total = x(1.1826 + 1.1183 + 1.0575 + 1) = 20000
4.3584x = 20000
x = 4588.84
So annual instalment is 4588.84
(Note if assume that first payment is not until the end of the first year then x = 6297.63).
18. Need to discount the 5 annual payments of 1000. So cost at todays prices is
8546
19.
Machine A Machine B
Profit 6000 -4000 = 2000 6000 -3900 = 2100

4 3900
Payback 2x = 1:8 years* 2x 1.95 years*
4.5 4000
2000 2000
ARR = 50% = 51.3%
4000 3900
*Note: It could be argued that the payback is 2 years for both if it is assumed that the revenue
is obtained at the end of each year.
Machine B is better on profit and ARR, but not on payback.
13
Machine A Machine B
Discount Cash flow Present Cash flow Present
factor value value
0 -4000 -4000 -3900 -3900
1 0.9259 2000 1851.9 1500 1388.9
2 0.8573 2500 2143.3 2500 2143.3
3 0.7938 1500 1190.7 2000 1587.7
NPV 1185.9 1219.9
Machine B is preferred as it has a higher NPV.
Machine A Machine B
Discount rate NPV Discount rate NPV
8 1186 8 1220
25 -32 25 -76
An IRR of about 24.5%
Using formula:
1186 25 (32) 8
1186 (32)
= 24.6%
For Machine B the IRR is 24.0%
Not much difference although Machine A has a slightly higher IRR and would therefore be
preferred using this measure.
20.
Investment 1 Investment 2
Profit 18 000 -15 000 = 3000 19 000 - 15 000 = 4000
Payback 3 years 3 years
6000 6333
ARR = 40% = 42%
15 000 15 000
Investment 2 gives a better profit and ARR.
14
Investment 1 Investment 2
Discount Cash flow Present Cash Present
factor value flow value
0 -15 000 -15 000 -15 000 -15 000
1 0.9524 6 000 5 714.3 0 0.0
2 0.9070 6 000 5 442.2 0 0.0
3 0.8638 6 000 5 183.0 19 000 16 412.0
NPV 1 339.5 1 412.9
Investment 2 gives a higher NPV Using formula for investment 1
1340 10 (79.2) 5
1340 (79.2)
= 9.7%
Using formula for investment 2
= 8:3%
So Investment 1 gives a higher IRR and is preferred on this measure.
21. (a) P3 = 800(1.04)3
= 899.89
(b) Let instalment = x
First instalment will grow to x(1.052)3 at the end of the plan
Second instalment will grow to x(1.052)2 at the end of the plan
Third instalment will grow to x(1.052) at the end of the plan
Total amount accumulated = x{(1.052)3 + (1.052)2 + 1.952}
= 3.322296x
Machine will cost 899.89 so
3.32296x = 899.89
899.89
And x = 3.32296
= 270.81
Instalments should be 270.81
15
22.
Year Cash flow Present value

0 -3750 -3750
1 1310 1178
2 1310 1059
3 1310 953
4 1310 857
5 1310 770
(a) NPV = 1067
(b) By trial and error (using Excel) a discount rate of 25% gives a NPV of -227.
Using formula
106725(227)11.2
IRR = 1067(227)
= 22:6%
23. The annual amount can be found from:
[(1.07)11 1]
= 1000000
. 07
15.78x = 1 000 000
x = 63 371
An annual investment of 63 371 would give a sinking fund of 1m in 10 years.
24.
(a)
[(1.075)20 1]
0 = 650001.075)20 + .075
0 = -276110 + 43.305x
x = 6376
So annual repayment would have to be 6376
(b) This time we know the annual repayment (7000) but need to find n, the number of
years to pay off the loan.
16
7000[(1.075) 1
0 = 65000 1.075 +
. 075
28 333 x (1:075)n = 93333
(1.075)n = 3.294
n log(1.075) = log(3.294)
n = 16.48 years
How much money is saved: Paying 7000 for 16.48 years gives a total of
115 374 and paying 6376 for 20 years gives a total of 127 520. So the amount saved is
12 146.
25.
Using the Constant repayments formula
0 = -150000(1.07)25 + x[(1.07)25 1]/.07
0 = -814114.896 + 63.249x
x = 12871.59
So annual repayments are 12,871.59
26.
Using the depreciation formula
5000 = 50000(1 + r/100)5
(1 + r/100) = (0.1)1/5
r = -36.904
So rate of depreciation is 36.9%
27.
(a) 5000 = 12 000 x (1 + r/100)3
1 + r/100 = (.4167)1/3
= .7469
17
r = -25:3
So the rate of depreciation is 25.3%
(b) At end of first year
P1 = 12 000 (1 - .253)
= 8964
So car will be worth 8964 after one year.
28. Take 1000 invested for one year.
If compounded annually at 7%
P1 = 1000 1.07
= 1070
If compounded continuously at 5%
P1 = 1000 x e5/1000
= 1051
So better to use the first option
18
1 (a) and (b) See Excel file (Chapter 13 Q1)
(c) Average seasonal differences (adjusted) are -10.05, -6.17, -5.38, -4.72,
8.66, 17.66
(d) Forecasts using regression line through CMA
Trend forecast Forecast

2015 1 43.5 33.4
2 43.6 37.4
3 43.7 38.3
4 43.8 39.1
5 43.9 52.5
2 (a) See Excel file (Chapter 13 Q2)
6 44.0 61.6
The additive model is best as it gives smallest MSE. (b) and (c) See Excel file (Chapter 13
Q2)
d) Forecasts
Period Trend Sales

2 26.7 22.9
3 27.5 18.6
3. (a) and (b) see Excel file (Chapter13 Q3)
4 28.3 32.5
(c) The additive is the better model as it has the lower MSE )0.04 compared to 0.55 for
multiplicative model)
19
(d) Forecasts are
Quarter Trend Sales

2 123.5 112.1
3 124.5 136.0
4 125.5 138.1
4. See Excel file (Chapter 13 Q4)
Forecast on day 12 is 112.4.
5.
See Excel file (Chapter 13 Q5) for detailed calculations. Multiplicative was used as it had the
lowest MSE. Forecasts are:
Period Trend Sales
1 196.517 70.2
2 199.220 151.1
3 201.923 405.9
4 204.626 178.9
If demand is expected to rise by 5% from 2011 then demand in quarter 3 and 4 of 2015
should be no more than 331 and 148 respectively. As the forecasts exceed this it looks as if
they wont have enough titanium to meet demand.
20
6.
(a)
There is a downward trend and if a straight line was fitted the sales would at some point
become negative! It is more likely that the trend would flatten off.
(b) Multiplicative as the MSE is lower. The chart shows that the seasonal swings are getting
less and not constant.
(c)
Year Quarter Turnove Forecas error in New
r t forecast forecast
2013 1 138
2 134.6 138 -3.4 137.32
3 129.5 137.32 -7.82 135.76
4 124.6 135.76 -11.16 133.52
2014 1 127.6 133.52 -5.92 132.34
2 125.8 132.34 -6.54 131.03
3 125.3 131.03 -5.73 129.89
4 117.7 129.89 -12.19 127.45
The forecast for quarter 1 of 2015 is 127.45
(d)
The squared errors using the figures above are
11.56
61.15
21
124.46
35.10
42.77
32.85
148.49
And the average of these is 65.2. This is the MSE
7.
(a) Simple exponential smoothing with a low alpha tends to dampen the forecasts and puts
less weight on more recent observations. This means that for the oil consumption graph the
forecast is very slowly adapting to the increase in oil consumption from from1984 to 1996
Simple exponential smoothing is not very good at forecasting a series with trends which is
what is apparent here.
(b) If the smoothing constant was increased the forecasts would try and follow the actual
series. However, there would always be a lag between the actual and forecast. A small value
of the smoothing constant is useful if it is required to predict the long term value of
consumption. However a larger value would give more weight to more recent observations
which might be what is required here.
(c) New forecast = last forecast + error in last forecast
= 1777 + 0.5(1696-1777)
= 1873
8.
(a) See Excel file (Chapter 13 Q8).
22
(b) See Excel file (Chapter 13 Q8). The trend line suggests that the sales are rising slowly.
The Additive model might be best as the fluctuations are fairly constant from year to year.
(c) See Excel file (Chapter 13 Q8) The MSE for the additive and multiplicative models are
.52 and .63 which shows that both models would be equally good although the additive
model is slightly better which is consistent with (b)
(d) See Excel file (Chapter 13 Q8). The deseasonlised figures for quarters 1 and 2 of 2015
are 124.8 and 123.4 respectively. The first figure is consistent with the trend but the figure
for quarter 2 is less showing that the trend may be changing.
(e) See Excel file (Chapter 13 Q8). The forecasts are shown below. (Note the trend forecast
were used by fitting a regression line to the trend values. The regression line was trend =
0.7491 + 113.82)
Period Trend Sales
1 126.554 116.6
2 127.303 136.2
3 128.052 143.2
4 128.802 114.6
23
1.
50 A
45
40
Optimal solution is at B which
35
is x= 20 and y = 25 giving a Z
30 of 137,500
B
25
20
15
10
5
C
0
0 10 20 30 40 50 60 70
2.
Let P = number of units of product P and
Q = number of units of product Q
Max 600P + 400Q
Subject to
2P + Q 10
P+Q7
P, Q 0
24
Solution is P = 3 and Q = 4 giving a profit of 3400.
3.
Let D = number of diesel engines and
P = number of petrol engines to be produced each day
Max 60D + 45P
Subject to
4D + 2P 16
2D + 2P 12
P3
D0
Solution is D = 3 and P = 2 giving a profit of 270
4.
(a)
Let L = number of low wattage bulbs to produce each day
Let H = number of high wattage bulbs to produce each day
20L + 30H 36 000 (10 x 60 x 60)
10L + 30H 36 000
L +H 1500
L, H 0
25
(b) See Excel file (Chapter 14 Q4) for Solver solution.
Solution is L = 900 and H = 600 giving a profit of 8700p (87)
(c) There are 9000 seconds (2.5 hours) left in testing but both fitting and supply of
shells are scarce resources. The shadow price of the fitting resource is .2p per second (7.20
per hour) and 1p per extra shell. Increasing either the fitting time or the number of shells
would increase total profits.
5.
19
Deluxe
18
17
16
15
14
13
12
11
10
A
9
8 A possible integer
7 solution
6
4
Feasible region B
3
2
Iso profit line
1
0
Basic
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
As the isoprofit line is parallel to AB (part of the resin constraint) there will be multiple
optimal solutions as an point on AB is optimal. A sensible one would be to make 12 Basic
and 5 Deluxe dinghies a day.
(b)
(i) Shadow price of the resin resource is 5.
26
(ii) Can store an extra 34kg which will give an extra profit of 5 34 = 170 per day.
6.
(a) Max 60P +85S
ST
4P +8S 400
2P + 3S 250
10S 150
5P + 7S 480
P 2S 0
(b) P = 70, S =15 giving a profit of 5475 per day
(c) Components A and c are binding (tight) constraints. Shadow prices of A and C are 15
and 3.50 respectively. (Note increasing C would reduce profits). It would be worthwhile
increasing A since shadow price greater than cost.
(d) No upper limit but lower limit of 42.50
7.
(a) 50 radio and 30 TV ads giving a total rating of 28000
(b) Radio and budget constraints binding. Shadow price of radio ads constraint is 50 so If
radio ads increased by 50 then extra rating would be 50 x 50 = 2500
(c) Cant reduce TV ads rating.
27
8. (a) Let x1 = Amount invested in government bonds
Let x2 = Amount invested in corporate bonds
Let x3 = Amount invested in FTSE 100 stocks
Let x4 = Amount invested in Aim stocks
Max .04x1 + .05x2 + .06x3 + .08x4 + .03(550000 - x1 - x2 - x3 - x4)
i.e. Max 16500 + .01x1 + .02x2 + .03x3 + .05x4
s.t
x1 + x2 + x3 + x4 550 000
x2 50 000
x2 + x3 + x4 300 000
x1 + x2 - x3 - x4 0
x1 + x2 + x3 - 3x4 0
x1 , x2 , x3 , x4 0
(b) (i) 200 000 in government bonds, 50,000 in Corporate bonds, and 125,000 in
each of FTSE 100 and Aim stocks. This means that 50,000 is left in the bank. Total return
is
200 .04 + 50 .05 + 125 .06 + 125 .08 + 50 .03
= 29 500
28
As a percentage on capital invested = 5.36%
(ii) The shadow price for constraint 2 is -.04 which means that reducing the RHS
increases the return. Can reduce RHS by 25 000 so increase in return would be 25000 .04 =
1000.
The shadow price for constraint 3 is 0.05 so increase the RHS by the maximum which is
25000. Increase in return would be 5000 .05 = 1250. So it is better to increase the RHS of
constraint 3.
(iii) The allowable decreases for x1 , x2 , x3 , x4 are .015, infinite, .08 and .02 respectively.
The implication of this is that x1 would have to fall to 2.5% which is below the bank rate. x2
and x3 would mean a negative rate so is not realistic. x4 could fall to 6% and still remain in
solution. The allowable decrease for corporate bonds is infinite because constraint 2 says that
must have at least 50,000.
9.
(a)
L = No. of LXT models to make each week
H = No. of HXT models to make each week
Max 580L +440H
ST
5L + 10H 40 (1)
9L + 2H 40 (2)
7L + 5H 40 (3)
29
L .6(L + H)
Or
0.4L - .6H 0 (4)
L, H 0
(b)
20
19
18
17
16
15
14
9 L + 2 H = 40
13
12
Optimal solution is L = 4 units per week and
H = 2 units per week
11
The value of the optimal solution is 580 x 4 +
440 x 2 = 3200
H
10
8
7 L + 5 H = 40
0.4L -.6H = 0
4
3
A Optimum
2
B 5 L + 10 H = 40
1
C
0
0 1 2 3 4 5 6 7 8 9
30
(c)
Substituting optimal values into the constraints
5 4 + 10 2 = 40 so binding
9 4 + 2 2 = 40 so binding
7 4 + 5 2 = 38 Slack of 2 hours
.4 4 - .6 2 = 0.4 Surplus variable = 0.4
(d)
If profit becomes less than 220 then optimal moves to point A (3.4, 2.3). If greater than
1980 then moves to C (4.4,0)
(e)
Need to consider the loss in profits of making one unit of ZXT
For each hour lost in the Body shop the company loses 35 so if the new model requires 4
hours in this shop the loss of profits will be 4 35 = 140
Simly for the Assembly shop the loss of profits is 2 45 = 90
As there are 2 spare hours in the Painting shop there will be no loss of profits
Total loss of profits is therefore 140 + 90 = 230
So profit on the ZXT will need to be at least 230 to make it worthwhile to produce.
(f)
The new constraint is
31
5L + 4H 30
Substituting L = 4 and H = 2 gives
5 4 + 4 2 = 28
As this constraint is not binding it does not affect the optimal solution found in (a)
10. (a)
Max Z = 1.5S + L
ST
0.5S + 0.4L 6
0.25S +0.7L 14
S + L 30
L5
S0
(b)
32
30
25
20
No. of Lentil Puff (L)
15 Z = 15
10
Feasible Optimal solution, Z = 17
region
0
0 5 10 15 20 25 30
No. of Shepherds Ecstasy (S)
So make 8 Shepherds Ecstasy and 5 Lentil Puff giving a profit of 17
(c) The isoprofit line is now parallel to the ingredient A constraint so there are multiple
optimal solutions all along this line. The profit would now become 15
(d) (i) Yes, it could increase to 7 pies today as this is within the allowable limits, but profit
would fall by 0.4 as the shadow price is 0.2
(ii) Shadow price of A is 3 so net profit would be 2 per kg. We could increase A by 8.5 kg
so profits would increase by 2 8.5 = 17
(iii) The shadow price of A is 3 per kg so using 0.6 kg on Corn Delight the loss in profits
would be 1.8 per pie. As B is not scarce there is no cost in using this ingredient. However
as the profit is less that 1.8 she shouldnt make this new pie.
11.
33
Please note an error in the formulation on page 382. The demand constraints should be
greater than or equal to otherwise the optimal solution is to deliver nothing!
Xap Xbp Xcp Xaq Xbq Xcq Xar Xbr Xcr Xas Xbs Xcs Total Avail/Req'd
5 13 7 6 4 8 5 22 20 3 9 4
Supply A 1 1 1 1 175 <= 175
Supply B 1 1 1 1 35 <= 80
Supply C 1 1 1 1 35 <= 150
Demand P 1 1 1 60 >= 60
Demand Q 1 1 1 35 >= 35
Demand R 1 1 1 50 >= 50
Demand S 1 1 1 100 >= 100
Solution 60 0 0 0 35 0 50 0 0 65 0 35 1025
12.
(a) Using the transport algorithm the optimal solution is:
Watford to London 330 boxes
Watford to NE 170 boxes
Birmingham to London 200 boxes
Birmingham to Wales 250 boxes
Birmingham to NW 450 boxes
Bristol to Wales 100 boxes
Bristol to SW 400 boxes
This allocation means that the NE were short by 330 boxes
The total cost of this allocation is 6316
(b) There is an alternative solution (as the shadow price of the cell Bristol to NW is zero. The
alternative solution is:
Watford to London 430 boxes
34
Watford to NE 70 boxes
Birmingham to London 100 boxes
Birmingham to Wales 350 boxes
Birmingham to NW 450 boxes
Bristol to NE 100 boxes
Bristol to SW 400 boxes
(c) The shadow price of using the Bristol to London route is 0.70 per box and this will be the
additional cost of using this route.
(d) The shadow cost of this route is 0.2 therefore if the cost could be reduced by this much
(that is 9.80 or less) it would be become economic to use this route.
13.
(a) Cost = 106,000. It is not optimal because there of the negative shadow price (-1) in
route Colchester to Warehouse IV
(b) The optimal solution is
Aberdeen to I (60) III (50) IV (65)
Bristol to II (35)
Colchester to IV (35)
Cost is 102,500
35
14.
(a) Solution is to make 3.75 convection heaters and 1.5 thermal heaters giving a profit of
35.25
(b) Solution is same as part (a). Can meet all goals except Goal 4 (short by 3.25
convection heaters)
15.
Goal 1: L + H + d1 - d1+ = 1500 Minimise P1d1-
Goal 2: 10L + 30H + d2- - d2+ = 28800 Minimise P2d2+
Goal 3: 5L + 7H + d3- - d3+ = 10000 Minimise P3d3-
Using solver (see Excel file Chapter 14 Q15) solving each goal sequentially we get
Low High d- d+ Total Limit
Goal 1 1 1 0 0 1500 = 1500
Goal 2 10 30 0 0 28800 = 28800
Goal 3 5 7 1120 0 10000 = 10000
#Bulbs 810 690
So we produce 810 Low wattage bulbs and 690 High wattage bulbs. Both goals 1 and 2 are
met but we fall short of goal 3 by 1120p or 11.20
36
1.
(a) 16 weeks (b) 0 (c) 2 weeks (d) Yes by 1 week
(e) 5
37
2.
31 weeks (see below)
8 10
D
2
10 12
24 27
12 15 15 17 17 20 20 24 27 31
0 8 8 12
E H I J K L
3 2 3 4 3
A C 4
8 4
8 12 12 15 15 17 17 20 20 24 24 27 27 31
0 8 4 6
0 4 F
B
2
4 End
31
9 13
0 1
G
1
38
16 17
3. (a)
(b)
(i) 6 workers required during weeks 1 to 3

(ii) One solution is to start C at day 4 so ends at 11. Activity D will then start at 11 and finish at 15
4.
(a)
A 0 20
F 20 50
0 20
20 30
20 50
C 2 12 D 12 13
Start B 02 E 13 19
3 13 13 14
0 13 10 1 14 20
2 6
G 12 18 H 18 22
J 50 53
40 46 46 50
6 4
50 53
I 0 16 3 Finish
53
16 34 50
K 02
51 53
2
Critical activities are A, F and J. Time = 53 weeks. Slacks are B, C, D, E = 1 and G, H = 28, I = 34 and K = 51 weeks. Would miss trade fair by
5 weeks
(b)
(i) I is not on critical path so not worthwhile
(ii) Need to consider both AFJ and BCDEFJ as both exceed 48 weeks
Find crash cost per week

Activity Weeks saved Extra cost Unit cost
A 2 6 3
E 1 2 2
F 2 12 6
I 8 8 1
J 1 10 10
Crash A by 2 weeks at a total cost of 6. Then F by 2. Then J by 1. Cant crash anymore
Path A2 F2 J1
AFJ 53 51 49 48
BCDEFJ 49 49 47 46
Crash cost - 6 12 10
Cum. Crash 6 18 28
cost
Lost profits 40 24 8 0
Total 30 26 28
Crash A by 2 and F by 2 weeks to give duration of 49 weeks. Total cost including lost profits would be 26,000
5. (a)
0 .5 .5 2.5
B C
.5 1 1 3
9 11 11 11.5
3 5 5 9
0 3
A E F H I
0 3 3 5 5 9 9 11 11 11.5
11.5 12
3 8
0 1
D G J
5.5 6.5 6.5 11.5 11.5 12

(b) Job will take 12 days and critical activities are A, E, F, H, I and J (c)
(c) Job delayed by 0.5 days
(d) 5 people required between days 6 and 8. Not possible to have F and G at the same
time. If F starts as soon as G has finished project will be delayed for 3 days.
A0 E2 F3 H4 I1 J4
1
1
C1
11
11
G2
0 2 4 6 8 10 12
Days
6.
(a)
(a)
E G
A
6 4
4
Start
C D
H End
4 2
B 5
F
2
3
Activity EST EFT LST LFT

A 0 4 0 4
B 0 2 2 4
C 4 8 4 8
D 8 10 8 10
E 10 16 10 16
F 10 13 17 20
G 16 20 16 20
H 20 25 20 25
(b) Project takes 25 days. Critical activities are ACDEGH

(c) Average duration of B = 20/6 = 3.3. As this is less than the EFT of A the project isnt
affected
(d) Cost (in 000s) is 25 + 5x6.5 = 57.5
(e) The critical path is ACDEGH but path BCDEGH takes 23 days so need to consider this
path too. Crash A by 2 days as cheapest and common to both paths and costgs 10,000 Now
both paths critical. Next crash E by 1 day which costs 7000. Now crash A and B by 1 day
each which costs 8000. Finally crash C by 1 day costing 15000
Path Duration A-2 E-1 A-1, B-1 C-1

ACDEGH 25 23 22 21 20
BCDEGH 23 23 22 21 20
Crash cost 10 7 8 15
Cum. Crash 10 17 25 40
Overheads 25 23 22 21 20
Penalty 32.5 19.5 13 6.5 0
Total 57.5 52.5 52 52.5 60
Cheapest option is crash by 3 days. Crash A by 2 days, and E by 1 day.

7.
( a)
A 05 C 58
16 69
5 3
Start B 06 D 69 F 9 13 Finish
13
0 06 69 9 13
6 3 4
E 07
29
7
Critical activities are B, D and F, project will take 13 weeks and slacks are A:1, C: 1 and E:2
(b)
Project delayed by 1 week. Activity C will now become critical instead of B and D
(c)
There are 3 paths through the network, BDF (13), ACF(12) and EF(11). Need to reduce BDF
first so crash activity B by 1 week costs 12k
Now need to crash both BDF and ACF. Choice here is between activity F (20k) and
activity B and A (17k) so chose B and A
This reduces time to 11 weeks at a total cost of 29k
F (20k) is common to all 3 paths and is cheaper than crashing B,C and E. Crash F
by 2 weeks
Now have to crash B, C and E by one week at a cost of 26k
Minimum time is 8 weeks at a cost of 12 + 17 + 40 + 26 = 95k
To summarize:
Path Time B (1) B(1), A(1) F(2) B(1), C(1),

E(1)
BDF 13 12 11 9 8
ACF 12 12 11 9 8
EF 11 11 11 9 8
Crashing 12 17 40 26
cost
Cumulative 12 29 69 95
cost
8.
(a)
A B D F
3 3 4 5
6
Activity EST EFT LST LFT
A 0 3 0 3
B 3 6 3 6
C 6 10 8 12
D 6 10 8 12
E 6 12 6 12
F 12 17 12 17
(b) Critical activities are A, B, E and F
Time = 17 hours
Slacks are C = 2 hours and D = 2 hours
(c) Normal cost 1650, penalty cost 500 and overheads 1700
Profit 10000 - 1650 - 500 - 1700 6150
(d) D is not on critical path
Path B1 E2 C 1, D 1, E 1
ABEF 17 16 14 13
ABDF 15 14 14 13
ABCF 15 14 14 13
Crash cost 100 300 500
Penalty cost 500 400 200 100
Overheads 1700 1600 1400 1300
Total 2200 2100 1900 1900
So crash by 4 hours (same cost as only crashing by 3 hours but better to finish earlier)
Crash B by 1 hour C by 1 hour D by 1 hour E by 3 hours
Total cost = 1650 + 1900 = 3550
Profit = 10 000 - 3550 = 6450

9.
5 10
F
5
5 10
2 5 5 6
D
B 1
3
2 5 6 7
6 9 10 16 16 21 21 24 24
0 2 2 4
E I J K End
5 3
C 3 6
A 2
2 7 10 10 16 16 21 21 24 24
5 7
3 3 6
2 Critical path is A, B, F, I, J and K
H
G 3 Total time = 24 weeks
1
7 10
6 7
(b)
A (3) B (2) F (0) I (2) J (2) K (2)
C(2)
G (1)
H (5)
D (1)
E (5)
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
3 5 9 7 6 5 0 2
(c)
Overload during the 4th week .
Could delay H by 1 week and E by 1 week.
Then max no. of persons required is only 7 during 5th week.
10 (a)
B
3
C
3.5
A E F H I J
Start End
2 1 10 6 2 4.8
D
3.3
G3
G
3
Activity to tp tm jJ (J2
C 2 7 3 3.5 0.6944
D 2 6 3 3.5 0.4444
J 3 6 5 4.8 0.25
(b) Details of the EST and LST and floats are given below.
Activity EST EFT LFT LST Float Critical?
A 0 2 2 0 0 Y
B 2 5 16.5 13.5 11.5 N
C 2 5.5 5.5 2 0 Y
D 2 5.3 5.5 2.2 0.2 N
E 5.5 6.5 6.5 5.5 0 Y
F 6.5 16.5 16.5 6.5 0 Y
G 0 3 16.5 13.5 13.5 N
H 16.5 22.5 22.5 16.5 0 Y
I 22.5 24.5 24.5 22.5 0 Y
J 24.5 29.3 29.3 24.5 0 Y
Activities A, C, E, F, H, I are critical. Expected time before the national campaign can be
launched is 29.3 days.
(c) The variance of the critical path is 0.6944 + 0.25 = 0.9444
(Note: Activity D is not on the critical path), so the standard deviation is 0.9718
To find the probability that expected time is greater than 32 days we need to calculate the Z
value which is
32 29.3
=
0.9718
= 2.77
From tables this give a probability of 0.0028 So there is a 0.28% chance that the expected
time will exceed 32 days, which means that there is a probability of 100 0.28 = 99.72% that
the launch can take place within 32 days.
55
1. EOQ = 63.25 Time between orders = 18.9 days
2.
For less than 600 units the stock holding cost, h = .270 = 14.0
2 9 4800
=
14
= 79
Cost of this is product cost + order cost + holding cost
=480070 + 94800/79 + 1479/2
= 337,100
If order 600
h= .268 = 13.60
Cost = =480068 + 94800/600 + 13.60600/2
= 330,552
So should order 600 units which gives a total annual cost of 330,552
3.
37 units (to nearest whole number)
4.
2 570 3000
=
. 01
= 18,493 So order 20,000 gallons (1 tanker load) every 6.7 days on average (20,000/3000)
when stock is down to 9000 gallons (3 3000)
Cost = 5703000/2000 + .0120000/2
56
= 185.50 per day
5. (a)
There is effectively a buffer stock of 300 tins as in lead time of 1 day the paint shop will only
use 200 tins. This buffer stock costs 300 0.25 = 75 per day to hold in store.
The normal stock holding cost is 0.25 500 = 125
An order is made every 5 days so the average daily order cost is 1005 = 20
Total cost is therefore 75 + 125 + 20 = 220 per day
2100200
The EOQ amount is = 400 (so ordering every 2 days)
0.25
Cost is now 100 a day, a saving of 120
If only order when down to 200 will not incur a buffer stock cost.
(b) At EOQ the product cost is 2000 per day giving a total cost of 2100
If order 2000 the product cost is 1960. The order and stock cost will now cost 260 per day
so total cost is 2220 which is more expensive.
(c) If x is the product cost this must be less than 1840 since x + 260 = 2100, which means
the cost per item should be less than 9.20. This represents a discount of 8%
6.
h = .210 = 2
2 20 80
=
2
= 40
Cost = 80 10 + 2080/40 + 240/2
= 880
If order 300 h = .29.80 = 1.96
57
Cost = 80 9.8+ 2080/300 + 1.96300/2
= 1083
If order 500 h = .29.70 = 1.94
Cost = 80 9.7+ 2080/500 + 1.94500/2
= 1264
7.
(a) Demand = 5040 = 2000 p.a.
Order cost = 642000/200 = 640
h = .252 = 0.50
Holding cost = .5 200/2 = 50
Total cost = 690
(b)
2 64 200
=
0.5
= 715.5
So order 2000/715.5 = 2.795 times a year
= 50/2.795 = 17.9 days on average
Holding cost = .5715.5/2 = 178.88
Order cost = 178.88
Total = 357.76 which is a saving of 332.24 p.a.
(c) If order 5000 then order cost = 64 2000/5000 = 25.6
Product unit cost = 1.80 so h = .251.80 = 0.45
58
And holding cost = .455000/2 = 1125
Total cost including cost of buying 2000 units= 20001.80 + 25.6 + 1125 = 4751
Total cost at EOQ = 20002 +357.76 = 4357.76
(d) Let discount = d%
Unit price = 2(1-d/100)
Cost of 2000 = 4000 (1-d/100)
h= .252(1-d/100)
= .5(1-d/100)
Holding cost = .5(1-d/100)5000/2
= 1250(1-d/100)
Order cost = 25.6
Total cost = 4000 (1-d/100)+25.6 + 1250(1-d/100)
When total cost is the same as the EOQ cost then we are indifferent between ordering the
EOQ amount or 5000
The EOQ cost = 4357.76
So 4000 (1-d/100)+25.6 + 1250(1-d/100) = 4357.76
Solving gives d = 17.48%
So a discount of at least 17.5% would be necessary to make it worth ordering in batches of

5000
8.
2 100 50000
=
0.25
= 6324.6 or 6325 metres
59
Holding cost = .256325/2 = 790.625
Order cost = 10050000/6325 = 790.51
Total = 1581
Orders per year = 50000/6325 = 7.905 on average.
Time between orders = 50/7.905 = 6.3 weeks on average.
9.
Average usage during lead time of 7 days is 7000 m.
Standard deviation during this same period = 7 2502
= 661.4
(a) At 5% Z = 1.645
7000
= 1.645
661.4
The buffer level = X 7000 = 1088
(b) At 1% Z = 2.327
7000
= 2.327
661.4
X- 7000 = 1539
10.
h= .1 10 = 1
2 50 10000
=
1
= 1000 m
Holding cost = 1 1000/2
60
= 500
Order cost = 50 10000/1000 = 50
Product cost = 1010000 = 100,000
Total cost = 100,550
If order 5000 then order cost = 50 10000/5000 = 100
Product unit cost = 9.50 so h = .109.50 = 0.95
And holding cost = .955000/2 = 2375
Product cost = 9.5010000 = 95,000
Total cost = 97,475
So should order 5000 as this saves 3075
11.
Let discount = d%
Unit price = 10(1-d/100)
Cost of 100,000 = 100000 (1-d/100)
h= .110(1-d/100)
= (1-d/100)
Holding cost = (1-d/100)5000/2
= 2500(1-d/100)
Order cost = 100
Total cost = 100000 (1-d/100)+2500 (1-d/100)+100
When total cost is the same as the EOQ cost then we are indifferent between ordering the
EOQ amount or 5000
The EOQ cost = 100,550
61
So 100000 (1-d/100)+2500 (1-d/100)+100 = 100550
Solving gives d = 2%
So the discount would need to fall below 2% before it makes more sense to order the EOQ
amount.
62
1.
Random numbers are allocated as follows:
Inter-arrival time (IAT) Mid point Random numbers

20-<50 35 00-04
50-<100 75 05-24
100-<150 125 24-54
150-<200 175 55-99
Random number 05 is equivalent to an arrival time of 75 seconds and ends at 75+45=120
This could be set out in a table as follows
RNo IAT Clock Service Ends Waiting

starts time
05 75 75 75 120 0
20 75 150 150 195 0
30 125 275 275 320 0
85 175 450 450 495 0
22 75 525 525 570 0
21 75 600 600 645 0
04 35 635 645 690 10
67 175 810 810 855 0
00 35 845 855 900 10
03 35 880 900 945 20
Average waiting time = 40/10 = 4 seconds
Cash dispenser has been available for 945 seconds but has been used for only 450 (1045)
seconds so utilisation = 450/945 = 47.6%
63
2.
Inter-arrival time (IAT) Mid point Random numbers

0<4 2 00-29
4<6 5 30-69
6<8 7 70-89
8<10 9 90-99
Service time Mid point Random numbers

1<3 2 00-49
3<5 4 50-89
5<7 6 90-99
RNo IAT Clock RNo Service Service Ends Waiting

time starts time
04 2 2 10 2 2 4 0
59 5 7 07 2 7 9 0
38 5 12 98 6 12 18 0
01 2 14 75 4 18 22 4
48 5 19 91 6 22 28 3
04 2 21 12 2 28 30 7
Mean time queuing = 14/6 = 2.3 minutes
3.
(b)
Unloading times Mid point Random numbers

0-<30 15 00-19
30-<40 35 20-54
40-<50 45 55-76
64
50-<60 55 77-91
60-<70 65 92-99
Random number Unloading time

42 35
17 15
38 35
61 45
4.
(a)
Simulation does not effect the current system so no disruption to operations

Cheaper, particularly if the planned changes do not work
Simulation will take less time than experimentation
(b)
Develop conceptual model

Build computer model
Verify that computer model is correct
Validate computer model against real data. The run length of the simulation should be
the same as the time of operation of the depot. The number of runs should be
sufficient to give reliable results (small confidence interval)
Repeat the simulation with two bays and compare the average waiting time of lorries
of both systems
(c)
Frequency distribution for the loading/unloading times
Loading Unloading
Time Mid % Cum % Random % Cum % Random
(mins) point No. Nos.
65
0 to <30 15 20 20 00-19 30 30 00-29
30 to 35 35 55 20-54 40 70 30-69
<40
40 to 45 22 77 55-76 25 95 70-94
<50
50 to 55 15 92 77-91 4 99 95-98
<60
60 to 65 8 100 92-99 1 100 99
<70
Inter-arrival time distribution
Time Mid point % Cum % Random Nos.

0 to <10 5 15 15 00-14
10 to <20 15 40 55 15-54
20 to <30 25 30 85 55-84
30 to <40 35 5 90 85-89
40 to <50 45 5 95 90-94
50 to <60 55 3 98 95-97
60 to <70 65 2 100 98,99
Type of lorry
Random Nos.
Loading (L) 70% 00-69
Unloading (U) 30% 70-99
Arrivals Service
RN IAT Clock Queue RN Type RN Time Starts Ends Wait
time
20 15 15 0 17 L 42 35 15 50 0
66
96 55 70 0 23 L 17 15 70 85 0
28 15 85 0 66 L 38 35 85 120 0
59 25 110 1 38 L 61 45 120 165 10
73 25 135 1 76 U 80 45 165 210 30
00 5 140 2 20 L 56 45 210 255 70
10 5 145 3 5 L 87 55 255 310 110
88 35 180 2 78 U 15 15 310 325 130
Average waiting time is 43.75 minutes. Maximum queue size is 3
67

Solutions Manual

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Solutions Manual

Uploaded by

Copyright:

Available Formats

Solutions Manual

Year Cost Index

Year Cost () Index (Year 1 = 100)

Resource Index (2013 = 100)

So card has increased the most (up 60%)

For 2015 based on 2013

Laspeyres' 114.75 Paasche's 188.12

7. Simple index for year 2 using year 1 as the base year

Supervisor: 14/12 x 100 = 116.7

Skilled: 10/9 x 100 = 111.1

Unskilled: 7/6 x 100 = 116.7

Supervisor: 15/12 x 100 = 125.0

Skilled: 11/9 x 100 = 122.2

Unskilled: 8/6 x 100 = 133.3

So unskilled increased wages the most.

Year 3 based on year 1

So price has dropped by 140,000 91,264 = 48,736

Year Turnover RPI Real T/O

sum 1888 2312.9 2395 2932

Laspeyres 122.51 Paasches 122.42

Laspeyres 130.65 Paasches 133.28

Month Price Index

Aug 155 100.00

(b)(i) August to January change = 104:52 - 100 = 4:52%

Average Average Wages deflated

Real drop in price = 11 179 at 2010 prices, i.e. 6.9%

Year Population Sample

Total Female Male Total Female Male

1 320 192 128 20 12 8

2 250 150 100 16 10 6

Total 800 480 320 50 30 20

to use online questionnaire.

could be asked to complete the questionnaire.

group they are in.

5 1, 2 and possibly 5 would have a sampling frame.

7. Answers depend on sample chosen.

living with parents or at university).

supporters or the different ages.

data is kept) or even all people in the local town.

leavers could then be selected from each chosen school.

sample of voters in the town.

(b) No, since all students would be on campus.

(c) Possible answers include:

Misleading or ambiguous questions. Could be overcome by a pilot study.

Asking the wrong question.

Asking personal questions.

Asking leading questions.

Making the questionnaire too long.

Difficulty in obtaining information about the student population.

may be too expensive.

Poor response rate. This can be improved by face-to-face interviews.

if only interview respondents in a particular part of the university).

Conducting survey during vacations or exam periods!

Randomly select 40 employees from the database of all 200 employees

Dept Number Proportion of total Sample size

Target population all first year students and all staff.

Sampling frame database of student and staff records.

Stratified sampling sample contains same proportion of relevant categories as

Multi-stage sampling hierarchical structure. If the students were based on different

campuses it might be necessary to pick one or two of these campuses randomly to

reduce travelling time.

(b) Possible questions are:

responses between staff and students.)