Modelling with Algorithms

A Selection of Problems Related to the Course

Algorithms
1. The instructions labelled 1 to 7 describe the steps of a sorting algorithm applied to a list of six numbers.

1. Let i equal 1.
2. Repeat lines 3 to 7, stopping when i becomes 6.
3. Let j equal 1.
4. Repeat lines 5 and 6, until j becomes 7 − i.
5. If the jth number in the list is bigger than the (j + 1)th, then swap them.
6. Let the new value of j be j + 1.
7. Let the new value of i be i + 1.

a. Apply the sorting algorithm to the list of numbers shown below. Record in the table provided the state of the list after each pass.
Record the number of comparisons and the number of swaps that you make in each pass. (The result of the first pass has
already been recorded.) [5]
i Comps Swaps
1 9 7 3 11 5 13 5 3

b. Suppose now that the list is split into two sublists, {9, 11, 7} and {3, 13, 5}. The sorting algorithm is adapted to apply to three
numbers and is applied to each subsist separately. This gives {7, 9, 11} and {3, 5, 13}. How many comparisons and swaps does
this need? [2]

c. How many further swaps would the original algorithm need to sort the revised list {7, 9, 11, 3, 5, 13} into increasing order? [1]

2.

a. The following instructions operate on positive integers greater than 4.

Step 10 Choose any positive integer greater than 4, and call it n.

Step 15 Write down n.

𝑛𝑛
Step 20 If n is even then let 𝑛𝑛 = and write down the result.
2

Step 30 If n is odd then let n = 3n + 1 and write down the result.

Step 40 Go to Step 20.

i. Apply the instructions with 6 as the chosen integer, stopping when a sequence repeats itself. [2]

ii. Apply the instructions with 256 as the chosen integer, stopping when a sequence repeats itself. [3]

iii. Add an instruction to stop the process when n becomes 1. [1]

 b. Six items with weights given in the table are to be packed into boxes each of which has a capacity of 10 kg

Item A B C D E F
Weight (kg) 2 1 6 3 3 5
i. Use the first-fit algorithm to pack the items in the order given, and state how many boxes are needed. [2]

ii. Place the items in increasing order of weight, and then apply the first-fit algorithm. [2]

iii. Place the items in decreasing order of weight, and then apply the first-fit algorithm. [2]

iv. An optimal solution is one which uses the least number of boxes. Find a set of weights for which placing in decreasing order
of weight, and then applying the first fit algorithm, does not give an optimal solution. Show both the results of first-fit
decreasing and an optimal solution. [2]

v. First-fit decreasing has quadratic complexity. If it takes a person 30 seconds to apply first-fit decreasing to 6 items, about
how long would it take that person to apply it to 60 items? [2]
Network Algorithms
3. The network below represents a number of villages together with connecting roads. The numbers on the arcs represent distances (in miles).

a. Use Dijkstra’s Algorithm to find the shortest routes from A to each of the other villages. Give these shortest routes and the
corresponding distances. [6]

b. Traffic in the area travels at 30mph. An accident delays all traffic passing through C by 20 minutes. Describe how the network
can be adapted and used to find the fastest journey time from A to F. [2]

4. Six remote villages are linked by a set of roads. Two villages are connected directly if there is a road between them which does not pass
through another village. The table gives the lengths in miles of all direct connections.

A B C D E F
A 6 7 12 3
B 6 10 8
C 7 10 2
D 12 2 9 8
E 8 9
F 3 8
a. Why might it be thought surprising that the direct distance between A and D is as long as 12 miles? Give a possible reason why
the distance is longer than might have been expected. [2]

b. Use the tabular form of Prim's algorithm, starting at A, to find a minimum connector for these villages. Draw your connector and
give its total length. [6]
5. A mountain ridge separates two populated areas. Networks representing roads connecting the villages in each area are shown below. The
numbers on the arcs represent distances in kilometres. There is also a mountain road of length 15 kilometres connecting C to Z.

a. A national bus company needs a route from A to X.

• Use Dijkstra's algorithm on the complete network, including CZ, to find the shortest route from A to X. Give the route and its
corresponding distance. [6]

• Would it need fewer computations to use Dijkstra's algorithm on the network for villages A to F to find the shortest route
from A to C, and then use Dijkstra's algorithm on the network for villages U to Z to find the shortest route from Z to X? Give
a brief justification for your answer. [2]

b. The local council needs to discover which roads it should keep clear of snow during the winter to keep all the villages connected,
and the corresponding total length of road.
• Use Kruskal's algorithm on the network for villages A to F to find a minimum connector for {A, B, C, D, E, F}. Show your use
of the algorithm. Draw your minimum connector. [3]

• Use Prim's algorithm on the network for villages U to Z to find a minimum connector for {U, V, W, X, Y, Z}, starting at U.
Show your use of the algorithm. Draw your minimum connector. [3]

• What is the total length of road that the council must keep clear of snow? [2]
Further Network Problems

6. Simon has a list of tasks which he has to complete before leaving his home to go on holiday. The table lists those tasks, their durations, and
their immediate predecessors. The durations assume that only one person is working on each activity
Duration
Task Immediate predecessor(s)
(minutes)

A pack suitcases 30 –

B make up beds 10 –

C clean upstairs 20 A, B

D wash upstairs floors 10 C

E bring in outside furniture 15 –

F close down central heating 5 –

G disconnect TV system 5 –

H load car 10 A

I clean downstairs 25 D, F

J wash downstairs floors 10 I

K wash patios 15 E

L lock up 5 G, H, J, K

a. Draw an activity on arc network for these activities. [5]

b. Mark on your diagram the early time and the late time for each event. Give the minimum completion time and the critical
activities. [6]

c. Explain why Simon will require help if the tasks are all to be completed within the minimum time. [1]

d. Simon’s friend offers to help. They share the tasks between them so that each task is completed by only one person. Produce a
cascade chart to show how the tasks can be shared between Simon and his friend, and scheduled, so that the project can be
completed in the minimum time. [4]
7. The table lists tasks which are involved in adding a back door to a garage. The table also lists the duration and immediate predecessor(s)
for each task. Each task is undertaken by one person.

Task Duration (hours) Immediate predecessor(s)

A measure 0.5 –

B manufacture frame and door 5 A

C cut hole in wall 2 A

D fit lintel and marble step 1.5 C

E fit frame 1 B, C

F fit door 1 E

G repair plaster around door 1 E

a. Draw an activity on arc network for these activities. [5]

b. Mark on your diagram the early time and the late time for each event. Give the minimum completion time and the critical
activities. [6]

c. Produce a schedule to show how two people can complete the project in the minimum time. [2]

d. Soon after starting activity D, the marble step breaks. Getting a replacement step adds 4 hours to the duration of activity D. How
does this delay affect the minimum completion time, the critical activities and the minimum time needed for two people to
complete the project? [3]
8. A village amateur dramatic society is planning its annual pantomime. Three rooms in the village hall have been booked for one evening per
week for 12 weeks. The following activities must take place. Their durations are shown.
• Each activity needs a room except for activities G, I and J.
• Choosing actors and dancers can be done in the same week. Rehearsals can begin after this.
• Rehearsing the dancers cannot begin until the music has been prepared.
• The scenery must be installed after rehearsals, but before dress rehearsals.
• Making the costumes cannot start until after the actors and dancers have been chosen.
• Everything must be ready for the dress rehearsals in the final two weeks of the 12-week preparation period.

Duration Immediate Predecessors

Activity
(weeks)

A choose lead actors 1

B choose rest of actors 1

C choose dancers 1

D rehearse lead actors 8

E rehearse rest of actors 6

F rehearse dancers 6

G prepare scenery 6

H install scenery 1

I prepare music 2

J make costumes 4

K dress rehearsals 2
a. Complete the table above showing the immediate predecessors for each activity [4]

b. Draw an activity on arc network for these activities [5]

c. Mark on your network the early time and the late time for each event. Give the critical activities. [5]
d. It is discovered that there is a double booking and that one of the rooms will not be available after week 6. Using the space
provided, produce a schedule showing how the pantomime can be ready in time for its first performance. [2]
Free Space

1
Room

1 2 3 4 5 6 7 8 9 10 11 12
Week
9. The figure below represents a system of one-way roads through which traffic flows from a source, S, to a sink, T. The weights on the arcs
show the capacities of the roads in cars per minute.

a. Explain why the maximum possible flow across FG must be less than 23 cars per minute [1]

b. The cut α partition the vertices into sets {S, C} and {A, B, D, E, F, G, T}. Calculate the capacity of cut α [1]

c. The cut β partition the vertices into sets {S, A, C, F, G} and {B, D, E, T}. Calculate the capacity of cut β [1]

d. Given that one of α or β is a minimum cut, write down the maximum possible flow from S to T. [1]

e. On a particular morning
• Road GE is closed for repairs,
• Roads FB, AD and DE are full to capacity

Use the diagram below to show that the maximum possible flow from part (d) is still achievable when GE is not used and FB, AD
and DE are full to capacity [2]
10. The figure below represents a system of pipes through which fluid can flow from two sources, S1 and S2, to three sinks, T1, T2 and T3. The
weights on the arcs show the capacities of the pipes in gallons per hour.

a. Add a supersource and a supersink to the network shown above. Give appropriate weightings and directions to the connected
arcs. [2]
b. Calculate the capacity of the cut X = {S1, S2, A, D}, Y = {B, C, E, F, T1, T2, T3}. [1]

c. Explain why the arcs FT1 and FT2 cannot both be flowing at capacity. [1]

d. On the diagram below, show a flow through the network of 230 gallons per hour. [2]

e. Prove that this flow is maximal [2]

.Linear Programming
11. Angelo manages a winter sports shop in a ski resort. He needs to decide how many snowboards and how many pairs of skis to purchase for
the coming season to maximise his profit from hiring them out.
• He has space for at most 250 snowboards and 500 pairs of skis.
• Because there are more skiers than snowboarders Angelo will purchase at least 10% more pairs of skis than snowboards.
• Both snowboards and skis need servicing, and his servicing facility can cope with no more than 600 units (ie snowboards or
pairs of skis).
• His expected profit from buying and renting out a snowboard is €40 for the season, and his expected profit from buying and
renting out a pair of skis is €50 for the season.
a. Define appropriate variables, construct inequality constraints, and draw a graph representing the feasible region for Angelo’s
decision problem. [10]

b. Give the objective function and find the solution which will give the maximum profit. [3]

c. Angelo considers increasing the cost of snowboard hire so that snowboard profits rise enough to change the optimal solution. By
how much will snowboard profits have to rise to change the optimal solution? [1]

d. Angelo increases the cost of snowboard hire and creates extra storage space for snowboards. What is the greatest number of
extra snowboards it is worth Angelo accommodating? [2]

12. Ian the chef is to make vegetable stew and vegetable soup for distribution to a small chain of vegetarian restaurants. The recipes for both of
these require carrots, beans and tomatoes.
• 10 litres of stew requires 1.5 kg of carrots, 1 kg of beans and 1.5 kg of tomatoes.
• 10 litres of soup requires 1 kg of carrots, 0.75 kg of beans and 1.5 kg of tomatoes.
• Ian has available 100 kg of carrots, 70 kg of beans and 110 kg of tomatoes
a. Identify appropriate variables and write down three inequalities corresponding to the availabilities of carrots, beans and tomatoes.
[5]

b. Graph your inequalities and identify the region corresponding to feasible production plans. [5]

c. The profit on a litre of stew is £5, and the profit on a litre of soup is £4. Find the most profitable production plan, showing your
working. Give the maximum profit. [3]

d. Ian can buy in extra tomatoes at £2.50 per kg. What extra quantity of tomatoes should Ian buy? How much extra profit would be
generated by the extra expenditure? [3]
13. Two products are to be made from material that is supplied in a single roll, 20 m long and 1 m wide. The two products require widths of 47
cm and 32 cm respectively.
Two ways of cutting lengths of material are shown in the plans below.

a. Given that there should be no unnecessary waste, draw one other cutting plan that might be used for a cut of length z metres. [2]

b. Write down an expression for the total area that is wasted in terms of x, y and z. [2]

c. All of the roll is to be cut, so x + y + z = 20. There needs to be a total length of at least 20 metres of the material for the first
product, the one requiring width 47 cm. Write this as a linear constraint on the variables [1]

d. Formulate an LP in terms of x and y to minimize the area that is wasted. You will need to use the relationship x + y + z = 20,
together with your answers to parts (b), (c) and (d). [3]

e. Solve your LP graphically, and interpret the solution [7]

Simplex
14. Colin has a hobby from which he makes a small income. He makes bowls, candle holders and key fobs.
The materials he uses include wood, metal parts, polish and sandpaper. They cost, on average, £15 per bowl, £6 per candle holder and £2
per key fob. Colin has a monthly budget of £100 for materials.
Colin spends no more than 30 hours per month on manufacturing these objects. Each bowl takes 4 hours, each candle holder takes 2
hours and each key fob takes half an hour.
a. Let b be the number of bowls Colin makes in a month, c the number of candle holders and f the number of key fobs. Write out, in
terms of these variables, two constraints corresponding to the limit on monthly expenditure on materials, and to the limit on
Colin's time [2]

b. Colin sells the objects at craft fairs. He charges £30 for a bowl, £15 for a candle holder and £3 for a key fob.
Set up an initial simplex tableau for the problem of maximizing Colin's monthly income subject to your constraints from part (a),
assuming that he sells all that he produces [2]

P b c f s1 s2 RHS

c. Use the simplex algorithm to solve your LP, and interpret the solution from the simplex algorithm [8]

P b c f s1 s2 RHS

P b c f s1 s2 RHS

P b c f s1 s2 RHS

d. Over a spell of several months Colin finds it difficult to sell bowls so he stops making them. Modify and solve your LP, using
simplex, to find how many candle holders and how many key fobs he should make, and interpret your solution. [3]

e. At the next craft fair Colin takes an order for 4 bowls. He promises to make exactly 4 bowls in the next month. Set up this
modified problem as an application of two-stage simplex. [3]

f. The solution now is for Colin to produce 4 bowls, 6⅔ candle holders and no key fobs. What is Colin's best integer solution to the
problem? [1]

g. Your answer to part (f) is not necessarily the integer solution giving the maximum profit for Colin. Explain why. [1]
15. A school awards three types of prize: junior, intermediate and senior. The school has the following constraints on the number of prizes it can
award.
•      At most 50 prizes may be awarded
•      At most twice as many senior prizes as junior prizes may be awarded
•      At least 10 % of the prizes must be intermediate
Each junior prize is worth £10, each intermediate prize is worth £15 and each senior prize is worth £25. The school wants to maximise the
total value of the prizes awarded.
Let x represent the number of junior prizes awarded, let y represent the number of intermediate prizes awarded and let z represent the
number of senior prizes awarded.
a. Formulate this information as an LP problem by
• Stating the objective function
• Listing the constraints as simplified inequalities with integer coefficients. [5]

b. The simplex algorithm is used to solve this LP problem. After a first iteration the tableau below is produced.

i. Explain how the tableau shows that the solution obtained after the first iteration is not optimal. [1]
ii. Perform a second iteration [3]

c. The final tableau after a third iteration of the simplex algorithm is given below

i. State the number of each type of prize that should be awarded. [1]

ii. State the corresponding total value of these prizes [1]

d. The school decides, as an additional constraint, to award at least 10 intermediate prizes. Explain why the simplex algorithm cannot
be used to solve the LP problem with this additional constraint. [1]

e. The two-stage simplex method is to be used to solve this modified LP problem. Formulate the additional constraint and define
the new objective function. Define each variable that you use. [3]
16. Three products, A, B and C are to be made.
Three supplements are included in each product. Product A has 10 g per kg of supplement X, 5 g per kg of supplement Y and 5 g per kg of
supplement Z.
Product B has 5 g per kg of supplement X, 5 g per kg of supplement Y and 3 g per kg of supplement Z.
Product C has 12 g per kg of supplement X, 7 g per kg of supplement Y and 5 g per kg of supplement Z.
There are 12 kg of supplement X available, 12 kg of supplement Y, and 9 kg of supplement Z.
Product A will sell at £7 per kg and costs £3 per kg to produce. Product B will sell at £5 per kg and costs £2 per kg to produce. Product C
will sell at £4 per kg and costs £3 per kg to produce.
The profit is to be maximised.
a. Explain how the initial feasible tableau shown in Fig. 3 models this problem [6]

b. Use the simplex algorithm to solve this problem, and interpret the solution [8]

c. In the solution, one of the basic variables appears at a value of 0. Explain what this means. [1]

d. There is a contractual requirement to provide at least 500 kg of product A. Show how to incorporate this constraint into the initial
tableau ready for an application of the two-stage simplex method. Briefly describe how the method works. You are not required to
perform the iterations. [5]
17. The following LP problem is to be solved.

Maximise P = 3x + y + 2z
subject to x + 2y ≤ 30
2x + y + z ≥ 14
2x + z ≤ 20
x, y, z ≥ 0
a. Complete the initial tableau below so that the two-stage simplex method may be used to solve this problem.
Show how the constraints have been made into equations using slack variables, surplus variables and artificial variables.
Show how the rows for the two objective functions are formed. [7]

b. Perform one iteration of the two-stage simplex method and explain why your answer shows that the first stage has been
completed. [5]

c. Using the tableau in above as the starting point for the second stage. Reduce the tableau so that the second stage can be started
and carry out the first iteration of the second stage of the two-stage simplex method, using an entry in the s2 column as the pivot
element. [3]

d. State which variables are now basic after this first iteration of the second stage of the two-stage simplex method. [1]
e. The final tableau is shown below:

Give the maximum value of P, and the corresponding values of x, y and z. [2]

f. The original LP problem is modified by replacing the constraint 2x + y + z ≥ 14 with 2x + y + z ≥ k.

Determine the maximum possible value of k such that the solutions to both the modified and original LP problems would be the
same. [2]
Answers
Questio
Answer/Indicative content Marks Guidance
n

1 i B1 i = 2 row OK

i B1 i = 3 row OK FT

i B1 i = 4 and 5 rows OK cao

i B1 comparisons

swaps

Examiner's Comments

Candidates are required to know

bubble sort. Those who recognised
it had an easier found the question
easier to deal with, but were at risk
of losing marks by not continuing
to the end, as the algorithm
specifies. Leaving the last line
blank cost one of the three marks
i B1
for the procedure, the mark for the
numbers of comparisons and the
mark for the numbers of swaps.

Some candidates clearly spent an

inordinately long time on parts (ii)
and (iii), with only two marks and
one mark available respectively.
Candidates need to move on if
their approach is taking longer than
is commensurate with the marks
allocated.

ii comparisons 6 B1 cao (OK if in 2 parts)

ii swaps 3 B1 cao (OK if in 2 parts)

ii
further swaps 6 B1 cao
i

Total 8

Examiner's Comments

2 i B1
Most succeeded in answering the
question, which included a
requirement to state how many
boxes had been used.
i 3 boxes B1

sorted increasing

Examiner's Comments

Most candidates continued to

score well on these two packing
questions, but some forfeited all
ii B1
marks by getting their “increasing”
and “decreasing” confused. It’s not
always the case that questions can
have a good storyline, but most do,
including this question. Candidates
sensitive to that might have
spotted such an error.

ii B1

placing a “3” or D or E into box 1

Examiner's Comments

Most candidates continued to

score well on these two packing
ii questions, but some forfeited all
M1
i marks by getting their “increasing”
and “decreasing” confused. It’s not
always the case that questions can
have a good storyline, but most do,
including this question. Candidates
sensitive to that might have
spotted such an error.

ii
A1
i

e.g. (for fitting into boxes of size 10)

6332222

i Reducing order / first fit:

M1 valid example
v

Optimal:

correctly doing it

Examiner's Comments

i
A1 Most candidates found this
v
challenging. It is to their credit that
they realisedwhat was needed, but
many generated large random sets
of weights, and hoped. Few
 thought it through to minimalist
sets such as 3, 3, 3, 3, 2, 2, 2, 2.

multiplying 30 by a squared value

Examiner's Comments

Fewer correct solu were those who

scaled up in proportion, but there
v 30 × (60/6)2 = 3000 secs … 50 minutes M1 were several who modelled quartic
complexity instead of quadratic,
and many others who confused
items and times in trying to
compute scale factors.tions were
seen than might have been hoped
for. Of course, there

v A1

Total 10

3 i B1 Dijkstra – C correct

i B1 other working values

i B1 order of labelling

i B1 labels

Note that D and G could be

i
labelled in the reverse order.

i B1 first 4 pairs

second 4 pairs

Examiner's Comments

i B1 Part (i) of this question was

answered well by candidates at all
attainment levels. Everyone
understood what is required in an
application of Dijkstra.

ii Turn distances to times throughout the network. E1 Explanations needed, not answers

ii Add 10 mins to every arc incident upon C.

any correct logic

(or do Dijkstra twice, once with C deleted, and compare with the
ii E1
adjusted time through C)
Examiner's Comments
 Part (ii) turned out to be very
challenging. Most candidates
scored zero, a few scored one, and
very few completely correct
solutions were seen. It had been
expected that most candidates
would score at least one, since the
first mark was for realising that the
network needed adapting to show
times rather than distances. The
majority of candidates did not try to
modify their networks. They
attempted to find the new solution
– emphatically not what had been
asked. Many who did attempt an
adaptation turned the delay into
an equivalent extra mileage, but
then gave no subsequent
indication that they needed to
convert their shortest route into a
fastest journey time, as had been
requested.

Dealing with the delay was a

challenge. Adding 20 minutes to
each arc through C gives a 40
minute delay on routes through C.
One cannot add 20 to the “ins” or
20 to the “outs” since one does not
know which will be which. 10
minutes extra on each arc incident
on C is what is required.

There is an alternative - add 20

minutes on the time taken for the
shortest route from part (i), and
then apply Dijkstra to the time-
weighted network with C and its
arcs deleted. Then choose the
better of the two.

It was distressing to read, so many

times, candidates who wished to
add 20 minutes or 10 miles to C,
thus showing a complete lack of
understanding of the nature of a
network.

Total 8

needs numerical justification

ACD is 7 + 2 = 9 (< 12)

4 i B1 Examiner's Comments
or AFD is 3 + 8 = 11 (< 12)

There were two marks for this, and

 two things to do to get them. Few
candidates failed both to give a
shorter indirect route, and to make
an appropriate comment regarding
the triangle inequality. Candidates
were not expected to quote the
triangle inequality, but the essential
maths is that for straight
connectionsAD ≤ AC + CD (or AD
≤ AF + FD). But then roads are not
often straight! That observation
would have been enough. Why
these particular roads were not
straight was a matter of
speculation, and in some answers,
fertile imaginations.

AD could by via some point of interest, or over difficult terrain, or …

i B1
The triangle inequality applies to triangles!

starting at and crossing row A (i.e.

ii M1
no selection in row)

selecting FA and BA (or first two

ii M1
arcs following wrong start)

numbering columns A, F and B

ii M1
(similarly)

all correct
ii A1 (dependent on 3 Ms)
(can cross all rows)

ii B1 cao (weights not needed)

26 (miles)

cao

Examiner's Comments
ii B1

This was answered well by a

substantial majority of candidates.
 The most common failing was the
failure to number columns as they
were included in the set of columns
under consideration.

Total 8

5 i Order of choice … AB, BD ↔ DE, DC, EF M1 Kruskal (first 2 arcs

i A1 identified – OK if by length)

i B1 min connector

ii Order of inclusion … U, W, X, V ↔ Y, Z M1 Prim (first 2 vertices

identified - OK to say UW, WX, etc,

ii A1 but in that order. Not OK to identify
by lengths)

ii B1 min connector

ii
Total length = 72 km. B1cao 27 + 30
i

ii
B1 + 15 for the pass
i

Examiner's Comments

Examiners needed to be convinced

ii that Kruskal was being used in (i)
i and Prim in (ii). The safest way to
do this is to show the order of
including arcs for Kruskal, and the
order of including vertices for Prim.

Total 8

6 i M1 activity on arc

i A1 single start and end

i A1 A, B, C OK

i A1 D, F, I OK

i A1 rest OK

forward pass (must have at least

i M1
one join correct)

i A1 FT

backward pass (must have at least

i M1
one burst correct)
i A1 FT

i Minimum completion time = 100 minutes B1 cao

cao

Examiner's Comments

Parts (i) and (ii) produced good

scores for nearly all candidates,
although there were more
examples of activity-on-node than
have been seen of late. One or two
candidates had networks with arcs
labelled with activities, but then
with two unspecified times
attached to each activity, rather
than two times attached to each
event. Their times earned no
marks. Activity “L” was often
omitted.

There was only one “burst” at

which the method mark for the
backward pass could be awarded,
this also accessing an accuracy
mark. This was at A’s “j” event /
H’s “i” event. The late event time
should have been 30, but 85 was
often seen.
i Critical activities are A, C, D, I, J and L B1

There were many instances of

candidates inserting unnecessary
dummies, e.g. between the end of
G and the beginning of L. This
does not incur any penalty, but
there is then an extra event
needed, with its early time and its
late time. The late time should be
the same as the late time for L’s “i”
event, but many candidates lost
the backward pass accuracy mark
by subtracting the activity duration,
i.e. by subtracting G’s duration in
this example.

Whilst some good scores were

seen, the quality seen differed
greatly. The worst case scenarios
had diagrams wandering over the
page with connections which
looked like tangled fishing lines.
The best were neat, crisp and clear
– the above point about clarity of
thought and clarity of expression
being clear to see.
ii M1 activity on arc

ii A1 single start and end

ii A1 A, B, C OK

ii A1 D, F, I OK

ii A1[5] rest OK

forward pass (must have at least

ii M1
one join correct

ii A1 FT

backward pass (must have at least

ii M1
one burst correct)

ii A1 FT

ii Minimum completion time = 100 minutes B1 cao

ii Critical activities are A, C, D, I, J and L B1 cao

Needs a comparison of times,

possibly implied.

Examiner's Comments

The mark in part (iii) was earned

by few candidates. The vast
majority of answers could be
ii e.g. Critical activities (100 mins) + others. paraphrased either by “Because he
B1
i e.g. B has to be done whilst A is underway. won’t complete within the minimum
time without help” or by “Because
two activities both have to be done
at the same time”. Both leave the
question unanswered. Arguably
the best answer, that there are 160
minutes of work to be done in 100
minutes of elapsed time, was seen
from only a very few candidates.

(If L omitted in (i) ignore omission here.) diagram like this or attempted
i
M1 cascade … no more than 1 omitted
v
activity

i nowhere needing more than 2

A1
v people

i
A1 precedences correct
v

fully correct, inc who does what

i Examiner's Comments
A1
v
There were candidates, when
answering part (iv,) who tried to
 schedule all non-critical activities to
start at time 0. There were those
who, whilst scheduling correctly,
failed to show who was doing
what.

Total 16

M1 Activity on arc

A1 Single start and end

e.g.
A1 A, B, C, D (precedences)

7 i A1 E (precedences)

A1 F and G (all correct)

M1 A1√ forward pass

M1 A1√ backward pass

i minimum completion time = 7.5 hours B1 time (cao)

critical activities (cao)

Examiner's Comments

critical activities – A, B, E, F, G
i B1 The CPA network was generally
(or ABEG + ABEF)
well constructed. The most
common error was to have
activities F and G share the same i
node and same j node.

M1 Activity on arc

A1 Single start and end

e.g.
A1 A, B, C, D (precedences)

ii A1 E (precedences)

A1 F and G (all correct)

M1 A1√ forward pass

M1 A1√ backward pass

ii minimum completion time = 7.5 hours B1 time (cao)

critical activities (cao)

Examiner's Comments
critical activities – A, B, E, F, G
ii B1
(or ABEG + ABEF) The forward and backward passes
were also well done. The common
error was to have a late time of 6
at C’s j-node / D’s i-node.
 not ft

must be labelled or to scale (e.g.

on the squares provided)
Can be written out instead.
e.g.

Examiner's Comments
B1
ii
Some candidates failed to answer
i
B1 the question in this part. Schedules
Needs to be clear what is done by whom.
and cascade diagrams were seen,
This doesn’t necessarily require people
but without indication of who was
being labelled ... but might.
doing what. In some cases that
might have been due to candidates
attempting to use colouring, or
other shading, which could not be
seen on scanned scripts.

i
8.0 hours or delay 0.5 hours B1 cao ISW if needed
v

i
A, C, D B1 cao
v

cao ISW if needed

Examiner's Comments

This part gave another opportunity

for the unfocused to engage in
descriptive flights of fancy. The
question could have been phrased
as “Give the new minimum
i completion time, critical activities
8.5 hours or delay of 1 hour B1
v and …”, and it is probable that a
better response would have
followed. But there are no
apologies for having asked “How
does this delay affect the minimum
completion time, critical activities
and … ?” There is a little bit of
modelling implicit in that, and it
certainly found somecandidates
wanting.

Total 16

A, B, C, G and I
B1 −1 for 1 or 2 wrong rows
8 i
B3 −2 for 3 or 4 wrong rows
−3 for 5 or 6 wrong rows
 activity on arc
M1 immediate predecessors A, B, C, I,
ii A1 G
A1 immediate predecessors D (A),
E(B), F(C,I,)

immediate predecessors J(A,B,C),

ii A1
H(D,E,F,G)

immediate predecessors K(J,H) +

ii A1
rest

ii M1 forward pass

ii A1√

ii M1 backward pass

ii A1√

ii Critical activities … A, D, H, K B1cao

ii
Same as for (ii)
i

eg

i B1cao F (ES2, LF9)

v B1cao rest

Examiner's Comments

This question started by requiring

candidates to model written
precedence expressions by a list of
immediate predecessors. This
made it more difficult than usual to
produce the activity-on-arc network
i in part (ii). Creditable attempts
v were seen, but many candidates
lost marks in their use of dummy
activities. For instance, many
candidates did not realise that the
following logic is OK for F
(immediate predecessors of C and
I) but not for J (A, B and C), since it
also has I as one of its immediate
predecessors.
 Total 16

Examiner’s Comments

Nearly all candidates gave the

correct answer that the only flow
into FG came from CF and as the
maximum capacity of CF was 18 it
B1
was therefore not possible to have
(AO1.1)
The only flow into node F is from arc CF which has a maximum a maximum flow of 23 cars per
9 a
capacity of 18 therefore the flow along FG must be less than 23 minute along FG. It should be
noted that in questions like this a
[1]
specific answer relating to the
given network and values given in
the question is required and not a
general statement about why a flow
may not be possible. A number of
candidates thought that the number
on arc FG implied that the flow
(and not the capacity) was 23 and
so they gave an argument based
on the fact that it could not be less
than 23 as it had to be 23.

B1
(AO1.1)
b
(i) [Capacity of cut α is] 97 Examiner’s Comments
[1]
This part was answered extremely
well by the majority of candidates
with nearly all giving the correct
answer of 97.

B1
(AO1.1)
b
(ii) [Capacity of cut β is] 161
[1]
Examiner’s Comments
 Understandably this part was not
answered as well as (b)(i) with
many candidates not including all
the arcs that went in the direction
from the source set of nodes to the
sink set (which were AD(23),
AE(18), SB(26), CB(28), FB(17),
GE(18) and GT(31) giving the
correct answer of 161). Some
candidates incorrectly included arc
BG. As this arc was directed from
the sink to source, it should not
have been included in the
calculation of the capacity of that
particular cut.

Follow
through their
least from
(b)(i) and (ii)

B1ft
c Maximum possible flow is 97 (AO2.2a Examiner’s Comments
)
Many candidates correctly applied
the maximum flow-minimum cut
theorem in this part and realised
that the least value from parts (b)(i)
and (ii) would be the maximum
possible flow through the network.

Consistent
flow pattern
(flow in = flow
M1 out at each
(AO2.1) node) – one
number on
each arc GE
blank/0,
FB = 17, AD =
23, DE = 16
Condone
incorrect flow
d across one
vertex for the
M mark
Condone GE
blank
A1
(AO2.2a
)
Examiner’s Comments
[2]
The most common
misunderstanding in this part was
by some candidates who only
 showed the flow in one particular
path from S to T rather than the
flow through the entire network
(subject to the conditions given in
the question). There are two main
points that candidates must
consider when tackling this type of
question. They are the feasibility
condition (which says that the flow
along each arc must not exceed
the capacity of that arc) and the
conservation condition (which says
that apart from the source (S) and
sink (T) nodes the total flow into a
node must equal the total flow out
of that node).

Misconception An arc that

has no flow passing through it,
such as GE in this part, should be
labelled with a 0 and not simply left
blank.

Total 6

B1 (AO
1.1)
A supersource with arcs SS1 and SS2.

A supersink with arcs T1T, T2T and T3T. B1 (AO

1
a 1.1)
0
Arcs labelled with directions and the following weights (or greater):

SS1 = 155, SS2 = 90, T1T = 80, T2T = 70 and T3T = 85

[2]

B1 (AO
Capacity of cut = 15 + 35 + 60 + 40 + 60 + 10 + 10 + 10 + 35 1.1)
b
= 275 gallons per hour
[1]

B1 (AO
2.4)
The maximum flow out of F is 95 (along arcs FT1, FT2 and FC).

c The total capacity of arcs AF, BF and GF is 35 + 40 + 15 = 90 and

so the maximum flow into F is 90, therefore FT1 and FT2 cannot
both be full to capacity

[1]
 M1 (AO
Consistent
2.1) flow pattern
(flow in = flow
out at each
node) – one
number on
each arc.
Furthermore
as the
capacity of
the arcs
leading into
the three
d
sinks is 235
and both FT1
and FT2
cannot both
be saturated
one of these
two arcs must
be reduced
for this mark
to be
A1 (AO awarded.
2.2a)
A correct
answer with
flow = 230
[2]

E1 (AO
3.1b)
The capacity of the cut which partitions the vertices into the sets
{S1, S2, A, B, C, E}, {D, F, G, T1, T2, T3} is 30 + 35 + 40 + 20 + 65
+ 40 = 230 ∴ min. cut ≤ 230

e E1 (AO
By the maximum flow–minimum cut theorem the maximum flow is
2.1)
equal to the minimum cut and so therefore the maximum flow
through the system is 230 gallons per hour

[2]

Total 8

1
i e.g.
1

i Let x be the number of snowboards B1

i Let y be the number of (pairs of) skis B1 or vice-versa of course

i x + y ≤ 600 B1

i x ≤ 250 and y ≤ 500 B1 both

i 1.1x ≤ y B1

i B1 FT horizontal line

i B1 FT vertical line
i B1 FT positive slope line

i B1 x + y = 600

Note … error tolerance of +/− half

i a small square within feasible
region

shading … follow any pentagon

bounded by the y-axis, a horizontal
line, a vertical line, a negatively
inclined line and a positively
inclined line

Examiner's Comments

It would be unusual if the report on

the LP question did not start by
stressing the importance of
variable definition. The usual
requirement for ’number of’ is not
an arbitrary ‘hobby-horse’, but
reflects the need that variables be
unambiguously defined. This was
well illustrated in this paper where
there was the added complication
of having both snowboards and
i B1
pairs of skis. Suppose we had a
report giving information about
Marco's purchase of snowboards
... we are told how many, the cost
and the delivered weight. So
"x=snowboards" is not sufficiently
specific ... it could be the number,
the cost or the weight. The same
report also has information about
skis ... number of pairs, the cost
and the delivered weight. In this
case "y=pairs of skis" is
unambiguous, and it was therefore
allowed.

The ’10% more’ constraint posed

some difficulties, though not so
many as might have been
expected.
 The graph was often drawn well,
although the gradient of the “10%
more” constraint did not always
match the candidate’s algebra.

Part (ii), the optimisation, was done

well. Parts (iii) and (iv), the post-
optimal analysis, proved to be
more of a challenge. In particular,
in part (iii) very many candidates
produced the answer €6 from the
computation (29000 –27500)/250
... the reader might try to see why.

ii Objective = 40x + 50y B1 objective

considering profits at the two

ii M1 indicated points of their pentagon
(or using a profit line)

ii 29000 at (100,500) A1 cao www

ii 27500 at (250,350)

ii Solution … 100 snowboards and 500 pairs of skis

ii
or more B1 cao (allow etc)
i

moving to appropriate new feasible

i
35 snowboards M1 point on their negatively inclined
v
line

i cao… integer! (allowing 30 to 40

A1
v for graphical inaccuracy)

Total 16

1 Let x be the number of (10s of) litres of stew and y be the number of “number of”, referring to soup &
i B1
2 (10s of) litres of soup that Ian makes. stew

identification of soup and stew

i B1
variables

i B1 −1 each scaling or
Carrots: 0.15x + 0.1y < 100, i.e. 3x + 2y < 2000

i B1 systematic error, e.g.

Beans: 0.1x + 0.075y < 70, i.e. 4x + 3y < 2800

equalities

Examiner's Comments

i B1
Tomatoes: 0.15x + 0.15y < 110, i.e. 3x + 3y < 2200 Many candidates failed to get to
grips with the underlying variables,
confusing stew and soup with
carrots, beans and tomatoes. To
 add to their confusion there were
both litres and tens of litres to
handle … too much for many to
make sense of.

Through it all shone those who had

the clarity of thought to identify and
define their variables clearly and
unambiguously. As in every report,
it is repeated here that the
identification should start with “Let
… be the number of …”

Intercepts are (666.7,0) and (0,1000)

ii
(700,0) and (0,933.3)
(733.3,0) and (0,733.3)

axes consistently labelled and

scaled

line 1
line 2
line 3 all √ subject to negative
gradients
B1
shading giving feasible
quadrilateral bounded by axes …
B1
ii or identified by vertices
B1
B1
Examiner's Comments
B1

Many candidates who stumbled

through part (i) managed to score
well in part (ii). Quite a lot failed to
label and scale their axes. A few
produced broken scales, and
scored zero.

Line 2 irrelevant. Comparing at (0, 733.3), (533.3 ± 10, 200 ±

10)and (666.7, 0)
comparing 3 vertices (not origin) or
ii (accuracy quoted is for graphical solutions).
M1 profit line with approximately
i Max profit at intersection of lines 1 and 3 (533.33,200) with profit
correct gradient (−5/4)
£3466.67 (accuracy from 3375 to 3560)
(cf £3333.33 and £2933.33)

ii
So make 533.33 litres of stew and 200 litres of soup, A1 stew and soup (cao)
i

profit (cao)

Examiner's Comments

ii
giving a profit of £3466.67 (3375 – 3560). A1 The usual optimisation requirement
i
had a sting in the tail for those who
had chosen their variables to
represent tens of litres of the
products … they had to do some
 interpreting … not 20 litres of soup,
but 200. Keeping in touch with
reality might have helped!

i
Best solution now at (0, 933.3)…profit £3733.33 (£373.33) M1
v

i
So best new solution uses 30 kg extra tomatoes (140 kg total) A1 30kg (allow 140 new total) cao
v

(allow £3658.33 new total) cao

i Examiner's Comments
Extra profit is £(3733.33 – 3466.67 – 30*2.5) = £191.67 A1
v
A challenging last part question. A
few candidates succeeded with it.

Total 16

M1
1
i 3 widths + waste
3
A1

ii 0.06x + 0.21y + 0.04z (m2) M1 3 areas

ii A1√ any units OK … ignore scaling

ii
2x + y > 20 B1
i

i
y + 3z > 24 B1√
v

v Use of z = 20 − x − y M1

Minimise 0.02x + 0.17y (constant of 0.8 not needed but OK if

v there) A1√ “minimise” not needed - given
st 2x + y > 20

v −3x − 2y > −36 or 3x + 2y < 36 A1√

 line (cao)
B1 line (cao)
B1 shading – follow two neg grad lines
v
B1 making a triangle with base on x-
i
axis
M1 objective valued at (10,0) and at
(4,12)

v
Cut 10m according to plan × and 10m according to plan z. M1 z =10
i

Gives …
v
i
20m of material with width 47cm

v
30m of material with width 32cm A1 both
i

v
1m2 of waste. A1 waste
i

Examiner's Comments

This question proved to be too

difficult for nearly all candidates.
Very few decent graphs were seen.
Most could do parts (i) and (ii), but
thereafter confusion reigned.
Candidates could not get to grips
v with the scenario, and it was
i common to see x > (or ⩾) 20 for (iii)
and y > (or ⩾) 24 for (iv).
The instruction seemed clear
enough in (v), but there were very
few attempts to use z = 20 − x − y.
Given all that, few could make any
coherent sense of where they had
got to in terms of drawing the
feasible region.

Total 16

1
i materials 15b + 6c + 2f ≤ 100 B1 cao
4
 cao

Examiner's Comments

Whilst candidates generally

ploughed through this question
mechanistically, there was a strong
i time B1 developmental thread to it.
A substantial minority of
candidates lost marks and time in
earlier parts of the question by
formulating and maximising a profit
function, whereas the question was
quite specific in referring to
income.

ii B1 objective … cao

ii B1 rest … cao

ii
B1 pivot
i

ii
M1 first iteration
i

ii
A1 cao
i

ii
B1 pivot
i

ii
M1 second iteration
i

ii
A1 cao
i

ii Non-integer solution ( bowls and candle holders) using all of

B1 solution ft
i budget and all available time, giving income of £225

ii
B1 resources and income cao
i

Examiner's Comments

The problem was degenerate -

there were two vertices of the
three-dimensional feasible region
ii
which were jointly optimal (as was
i
any point on the line joining the two
vertices). Candidates following
through the given instructions
should have first found, in part (iii),
a non-integer solution.

i Might miss out “b” col. Any valid

M1
v approach using simplex
 i
A1 solution ft
v

i Make 15 candleholders. Same income as before, but £10 materials

A1 comment cao
v remain (and integer solution this time).

Examiner's Comments

Candidates were led, in part (iv) to

i the adjacent solution, which was
v integer. Some candidates shortcut
this process, but were credited
appropriately.

v B1 new objective

v B1 bowls ≤ 4

v B1 bowls ≥ 4

v OR or

v B1 objective

Examiner's Comments

In part (v), the set-up requires two

new inequalities, that the number of
v B1 bowls should be both less than or
equal to 4 and greater than or
equal to 4. Many candidates
missed the first of these
constraints.

v B1 bowls ≥ 4

Special case … Candidates may ignore the instruction and set up

v an ordinary simplex with b excluded and with reduced resources of SC2 −1 each error
£40 and 14 hours.

4 bowls, 6 candle holders and 2 key fobs.

v
(Uses all of the budget. Leaves an hour to spare. Gives an income B1
i
of £216.)

v There might be another solution with less income, but even less
B1
ii expenditure.

Total 20

Maximise 10x + 15y + 25z B1(AO3.

1
a 3)
5
subject to x + y + z ≤ 50 B1(AO1.
 1)
y ≥ 0.1( x + y + z ) M1(AO3
.3)
Correct or z ≤ 2x but
⇒ x − 9y + z ≤ 0 structure – allow incorrect
accept any inequality for
z ≤ 2x A2(AO1. inequality
the M mark
1)
−1 each error
x ≥ 0, y ≥ 0, z ≥ 0 (AO1.1)

[5]

B1(AO1.
1) Negative
b Eg There is still a negative entry in the objective row and so entry in the
(i)
z can still be increased objective row
[1]

M1(AO3
.4)
Pivot
A1(AO1.
b 1) Any non basic
variable
(ii) column
correct
A1(AO1.
1) All correct
[3]

B1(AO3.
2a)
c
(i) 15 junior prizes, 5 intermediate prizes and 30 senior prizes
[1]

B1(AO3.
2a)
c
(ii) £975
[1]

E1(AO3.
Eg The simplex algorithm can only be used to solve constraints of
5b)
d the less than or equal to form and the new constraint is
y ≥ 10
[1]

B1(AO3.
1a)
B1(AO2.
5)
e s4 is a surplus variable, a1 is an artificial variable
B1(AO3.
A + y − s4 = 10 where A is the new objective function to be
5c)
minimised

[3]

Total 15

1
i Let a be the number of kg of A… B1 variable defs.
6
i Line 1 ⇔ max (7–3)a+(5–2)b+(4–3)c ⇔ 4a+3b+c B1 objective

i B1 (7 − 3)…

i Line 2 ⇔ 10a+5b+12c ≤ 12000 (availability of X) B1 identifying constraints

i Line 3 ⇔ 5a+5b+7c ≤ 12000 (availability of Y) B1 LHS (used)

≤ + RHS (available)

Examiner's Comments

The modal mark was 5. This was

because hardly any candidates
correctly identified the variables.
For instance, a common offering
was “Let a be the amount of A ...”,
or similar. That would allow a to
be, for example, 1000kg. But in the
algebra of number, and in the
modelling which is given, letters
represent numbers.
i Line 4 ⇔ 5a+3b+5c ≤ 9000 (availability of Z) B1
This point has been emphasised in
examiner reports every year, but
still it eludes candidates. Variable
definitions MUST be in terms of “…
number of …”, in this case “Let a
be the number of kg of A …”. This
is not being picky. Precise variable
definition is absolutely crucial to
correct modelling.

Some candidates tried to answer

this in words … possible, but in
practice this did not lead to many
marks.

ii B1 Pivot

ii M1A1

ii B1 Pivot √

ii M1
ii A1 cao

ii Make 2400 kg of B B1 must refer to kg

Examiner's Comments

ii at a profit of £7200 with 1.8kg of Z left B1 The simplex was done well.
However, the interpretation was
often incomplete or had units
missing, which were vital in this
question.

ii Either…It means that the second constraint is coincidentally exactly

i satisfied at the solution.

ii
or…It means that product A is in the solution, but at zero value. B1
i

ii Examiner's Comments
(Candidates may refer to degeneracy, which will earn the mark.)
i
Again, interpretations were often
incomplete.

i
B1 new objective
v

i
B1 surplus+artificial
v

i
B1 new constraint
v

i
Minimise Q until 0 (if feasible).
v

i
Then drop Q and f and proceed to optimum. B1
v

i
B1
v

Examiner's Comments
i
Allow up to 3 out of 5 for big M.
v The question asked for
incorporation into initial tableau, but
candidates often lost marks by
failing to do this.

Total 20
 First two
equations

Three rows
correct
following
through their
B1 equations (all
boxes
(AO1.1)
complete in
x + 2y ≤ 30 ⇒ x + 2y + s1 = 30 each row)
B1
2x + z ≤ 20 ⇒ 2x + z + s3 = 20 (AO3.1a
P = 3x + y + 2z ⇒ P – 3x – y – 2z = 0 )
Examiner’s Comments
B1
2x + y + z ≥ 14 ⇒ 2x + y + z – s2 + a1 = 14 (AO3.1a
It is vital in questions like this that
)
the candidate reads the entire
Q = a1 ⇒ Q + 2x + y + z – s2 = 14 B1
question carefully before beginning
(AO3.1a
as a number of candidates only
1 Where s1, s3 are slack variables, s2 is a surplus variable and a1 is an )
a completed the initial tableau and
7 artificial variable B1
did not show the constraints (as
(AO2.5)
equations) or show how the two
Q P x y z s1 s2 s3 a1 RHS objective functions were formed. It
was clear that candidates were
1 0 2 1 1 0 –1 0 0 14
much more confident in adding
0 1 –3 –1 –2 0 0 0 0 0 M1
slack variables to the two ‘less than
(AO3.3)
0 0 1 2 0 1 0 0 0 30 or equal to’ constraints and re-
0 0 2 1 1 0 –1 0 1 14 writing the P objective function than
A1
0 0 2 0 1 0 0 1 0 20 adding the surplus and artificial
(AO1.1)
variables to the ‘greater than or
equal to’ constraint and forming the
[7]
second (Q) objective function.

AfL When candidates are

asked to derive equations or
constraints that need to include
variables that have not been
defined in the question that it is the
responsibility of the candidate to
define the variables they use. In
this case, for example, many
candidates did not make it explicitly
clear from the letters they used
which was the artificial and which
was the surplus variable.
Candidates should also be
 reminded that when filling in an
initial tableau that all cells must be
completed (especially zeros).

Condone no
positive
values in the
Q row

B1 Examiner’s Comments
(AO1.2)
Again, in this part the number of
B1 marks on offer should have
b
(AO1.2) indicated the fact that there were
two points that needed to be made
(The solution is feasible since) Q = 0 when discussing whether the first
[2] stage had been completed. The
first was that the solution was
(The solution is optimal since) there are no non-artifical positive feasible at this stage as Q = 0 and
numbers in the Q row (so therefore the first stage is complete) the second was that the solution
was optimal since there were no
non-artificial positive numbers in
the Q row. A number of candidates
misunderstood the demand here
and gave a response based on the
pivot column or which variables
were now basic.

Pivot row and

column
correct
(condone a
slip on pivot
row)
Any non-zero
M1 or 1 column
(AO3.4) correct
cao
P x y z s1 s2 s3 RHS
1 0 –1 –1/2 0 0 3/2 30 A1
c
0 0 2 –1/2 1 0 –1/2 20 (AO1.1)
Examiner’s Comments
0 1 0 1/2 0 0 1/2 10 A1
0 0 –1 0 0 1 1 6 (AO1.1)
Candidates overall were very
proficient in carrying out the
[3]
simplex algorithm and many
identified the correct pivot value
(the 1 in the bottom row) and used
this to carry out the first iteration as
requested. The most common
errors were sign errors on certain
values or difficulty in dealing with
 fractions. There were a significant
minority were pivoting on an
incorrect value with some
candidates using a negative value
as a pivot.

Condone
inclusion of P

Examiner’s Comments

B1 Many candidates left this part blank

(AO1.1) but of those who did attempt it,
d x, s1, s2
many correctly stated the three
[1] variables that were basic after the
first iteration. In this series
examiners condoned the inclusion
of P as a basic variable but centres
are reminded that P is not a
variable in this sense and its
inclusion in future series may
possibly be penalised.

B1
(AO1.1)
Pmax = 55 and
B1 Examiner’s Comments
e
(AO1.1)
x = 0, y =15, z = 20
Nearly all candidates who
[2] attempted this part correctly stated
the maximum value of P as 55 –
many candidates however stated x
incorrectly as 21 rather than 0.

Or substitute
values from
(e) into 2x + y
+z
k = 35
SC B1 k = 35
M1
The bottom row says that the surplus variable s2 = 21 with no
(AO3.4) working
This implies that that the constraint
f 2x + y + z ≥ 14 is being underused by 21 and therefore the solution
A1
to the LP problem would be no different if this constraint was
(AO3.5a
replaced by 2x + y + z ≥ 35 Examiner’s Comments
)

Of the two possible ways of

attempting this last part, the vast
majority of candidates used the
results from part (e) and substituted
these values into 2x + y + z to
 obtain the answer of k = 35.
Another way of tackling this part
was to realise that the surplus
variable had a value of 21 which
indicated that the corresponding
constraint was being underused by
21 and so the solution to the LP
problem would be no different if 14
was added to the 21 to give k =
35. Once again, as the word
‘determine’ was given in the
question the correct answer with no
working only scored 1of the 2
marks available.

Total 17