the Further Mathematics network – www.fmnetwork.org.

uk V 07 1 1

REVISION SHEET – DECISION MATHS

CRITICAL PATH ANALYSIS (AQA)
The main ideas are covered in Before the exam you should know:
AQA Edexcel MEI OCR • How to draw precedence networks accurately and
D2 D1 D1 D2 neatly.
• How to perform forward and backward passes on a
precedence network to calculate early and late start
The main ideas in this topic are: times.
• How to find the critical activities.
• Drawing Activity or Precedence Networks • How to calculate the various types of float.
• Performing Forward and Backward Passes • How to draw a cascade chart and construct a resource
and Identifying Critical Activities histogram.
• Where resource levelling is required and how to make
• Drawing Cascade Charts and Resource effective use of float to improve efficiency.
Levelling • What is meant by crashing a network.

Terminology
Example:
An activity is a task which needs to be done and takes an
The table shows the activities involved in creating a small patio
amount of time/resources to complete.
in a garden.
Precedence tables show the activities that need to be done
together with their duration and their immediate predecessors.
Activity Task Time Preceding
Precedence networks show the sequence of the activities. The
Name (hrs) Activities
network must have one end node.
Activities are represented by nodes in the network. A Clear Garden 4
Precedences are shown by arcs in the network. B Measure area 1
Drawing a precedence network: It can be a good idea to do an C Design Patio 2 B
initial sketch as it’s often possible to make your diagram clearer.
Forward pass establishes the earliest times that activities can D Purchase fencing 1 B
begin. E Buy pots and plants 3 A,C
Backward pass establishes the latest time. F Plant all pots 1 E
Critical activites are those whose timing is critical if the project
is to be completed in the minimum time. The critical activities G Purchase paving 1 C
will form a path through the network. H Construct Garden 6 H, D,G
Float is the amount of time by which an activity can be delayed
or extended. The network for this precedence table
Independent float does not affect other activities.
Interfering float is shared between two or more activities.

E F
3 1
END
A 0
4
C G H
2 1 6

B
1 D
1

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching
each specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

The forward and backward pass

E F
4 3 7 7 1 10
END
A 10 0 10
0 4 4
C G H
1 2 3 3 1 4 4 6 6

B
0 1 1 D
1 1 4

This is the earliest This is the latest

The duration of the project is 10 hours start time for the finish time for
The critical activities are A, B, C, G and H activity the activity

This is the duration

Float of the activity
activity float type
D 2 hours independent In this example there are two hours of float shared between
E 2 hours Interfering (with F) activities E and F
F 2 hour Interfering (with E)

Cascade Chart and Resources levelling

0 1 2 3 4 5 6 7 8 9 10 A Cascade Chart shows each activity set against a time line.
A Float time is sometimes shown by using shading.
B Dependencies are shown by vertical lines.
C The cascade chart can be adjusted by using the float times
D to make use of resources more efficient.
E
F
G
H

4 If activity A needs two people and all the rest can be done
3 by one person, then the resource histogram looks like this
2 2 (note that 4 people are needed in the second hour).
1
0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10
A If only three people are available for the first three hours,
B but a fourth friend can then come and help for an hour, we
C could move activity D within its float time to make this
D possible.
E
F This would make the cascade chart look like this:
G
H

4
3
2 2 The resource histogram would now look like this:
1
0 1 2 3 4 5 6 7 8 9 10

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching
each specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

REVISION SHEET – DECISION MATHS

DYNAMIC PROGRAMMING
The main ideas are covered in Before the exam you should know:
AQA Edexcel MEI OCR • That dynamic programming is a technique for solving
multi-stage decision making problems by working
D2 D2 D2 backwards.
• What is meant by stage and state variables, actions and
values (or costs).
The main idea in this topic is: • The meaning of sub-optimisation.
• How to set up a dynamic programming tableau.
• Finding the shortest (or longest) • How to calculate the sub-optimal strategy at each stage.
route through a network by working • That a minimax route is one on which the maximum
route is as small as possible.
backwards from T to S
• That a maximin route is one on which the minimum
route is as large as possible.

Dynamic programming is used to solve some optimisation problems modelled A D

By networks, though the solution is usually presented in a table. In a dynamic
programming network, the nodes are referred to as states, directed arcs are
called actions and the transition from one state to the next is a stage.
S B E
T
The method starts at T (state 0) and considers all the nodes joined directly to
T (D, E and F in the diagram), which are called stage 1 nodes.
We find the best route from each of these to T then move on to the stage 2
nodes (in this case A, B and C). The optimal route from each of these to T is C F
found using the best routes from stage 1 and so on until we reach S.
Stage 3 Stage 2 Stage 1
At each stage you work out the best strategy from that point. This is called the
sub-optimal strategy.
Example 1: The network below shows the weight limit,
in tonnes, on vehicles that use the road. A transport The problem is about finding the route with the greatest
company making regular deliveries from S to T wants to minimum weight from S to T so it is an example of a
use the largest possible vehicle in order to minimise their maximin problem. For each node we are choosing the
outgoings. What is the heaviest vehicle they can use? minimum arc on the current route to that node.

A Solution
4 D
stage State Action Value Current
4 5 (node) (into node) (Route min) max
6 3 1 0 (D) 0 3* 3
1 (E) 0 4* 4
S 3 B 2 E 4 2 (F) 0 6* 6
T
2 0 (A) 0 min (4,3)=3 4
1 min (4,4)=4*
1 (B) 0 min (5,3)=3 4
6 4 6 1 min (2,4)=2
3
2 min (4,6)=4*
C 5 F 2 (C) 0 min (3,4)=3 5
1 min (5,6)=5*
Optimal route S C F T 3 0 (S) 0 min (6,4)=4 5
Maximum weight of lorry is 5tonnes 1 min (3,4)=3
2 min (6,5)=5*

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

Example 2
A company is planning to build three new houses A, B and C at the rate of one every three months. They can build the
houses in any order but the profit will be affected by the workers available and the supply costs, which are subject to
variation. The expected profits in thousands of pounds are shown in the table

Time period Already completed Profits (£1000) (a) Represent the information on a
A B C network such that the optimal strategy
1 -- 54 50 48 will correspond to the longest path
2 A - 60 56 through the network.
B 73 - 56 (b) Use dynamic programming to
C 70 65 - determine the order in which they
3 A and B - - 66 should build the houses in order to
A and C - 69 -
maximise profit
B and C 71 - -

Solution

The letters in the nodes

60
represent the houses A A, B
already built at that stage
of the project 54
56 73
67

50 69
none B A, C A, B, C

56 70
48 71

C B, C
65
(b)
Stage State Action (house to Value max
(time period) (houses built) be built) (profit)
3 0 (A, B) 0 (C) 67* 66 It can be very helpful to write
1 (A, C) 0 (B) 69* 69 what is happening in the state
2 (B, C) 0 (A) 71* 71 and action columns.
In this case the state is the
2 0 (A) 0 (B) 60+67=127* 127
houses that have already been
1 (C) 56+69125
built and the action is the
1 (B) 0 (A) 73+67=140* 140
house that is about to be built.
1 (C) 56+71=128
2 (C) 0 (A) 70+69=139* 139
1 (B) 65+71=136
1 0 0 (A) 54+127=181 190
0 (B) 50+140=190*
0 (C) 48+139=187

The houses should be built in the order B then A then C

The maximum profit is £190000

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1
REVISION SHEET – DECISION MATHS
GAME THEORY
In a two person, zero sum game, one player’s gain is the same
The main ideas are covered in
AQA Edexcel MEI OCR
Before the exam you should know:
D2 D2 D2
• What is meant by a two person, zero sum game.
• How to interpret a payoff matrix.
• How to find a playsafe strategy.
The main ideas in this topic are: • What is meant by a stable solution.
• Finding out whether a two person zero sum • How to find the value of a game.
game has a stable solution. • How to simplify games using dominance.
• If a stable solution does not exist, using a • How to find an optimal mixed strategy for a game
graphical method to find an optimal mixed with no stable solution.
strategy. • How to convert a two person zero sum game into
a linear programming problem.
as the other player’s loss.
A pay-off matrix represents the gain for one of the players for each combination of strategies for the two players in a
two-person game.
Finding a play-safe strategy for a zero-sum game
In the pay-off matrix for A, find the minimum entry in each row; A will use the strategy that involves the greatnesses of
these values (maximin).Then find the maximum entry in each column, B will use the strategy that involves the lowest
of these values (minimax).
A stable solution occurs if and only if the maximum(row minimum) = the minimum(column maximum).
A assumes B uses his play-safe strategy and cannot do better by using an alternative strategy and B assumes A uses his
play-safe strategy and cannot do better by using an alternative strategy. A stable solution is sometimes called a saddle
point.

Example: Anna and Barry play a zero sum game. The Solution
game can be represented by the following payoff matrix
Row
for Anna B1 B2 B3 minimax
max
B B B

B1 B B2B B3B
A1 3 2 5 5
A1 3 2 5 A2 -1 4 -2 4
A2 -1 4 -2 3
A3 2 1 3 3
A3 2 1 3
Col min -1 1 -2
1 ≠ 3 so solution
(a) Show that there is not stable solution maximin 1 is not stable

Dominance: A row/column can be eliminated if all the entries in that row/column are less than or equal to the corresponding
entries of another row/column because the player would never choose that row/column.
Finding a mixed strategy using a graph: If player A has two possible strategies, assume he adopts the first
strategy with probability p and the second with probability 1-p. The expected payoff for A will depend on
which strategy B chooses so plot A’s expected payoff for each of B’s strategies as lines on a graph. A’s best
strategy is to play with the value of p which gives the highest point at the intersection of the lines. This will
maximise A’s minimum return (maximin). The value of the game can be B1 B2 B3 B B B

calculated using this value of p. A1 3 2 5

Example (cont) A2 -1 4 -2
(b) Explain why Anna should never play strategy A3 A3 2 1 3
(c) Find the best mixed strategy for Anna and give the value of the game for her.

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1
Solution
(b) Anna should not play A3 since row A3 is dominated by row A1 since 3 > 2, 2 > 1 and 5 > 3.
(c) Let Anna play
B1 B2 B3
A1 with probability p
B B B

A1 3 2 5
A2 -1 4 -2 A2 with probability (1 − p )

Anna’s optimal strategy

Anna’s expected payoff if Barry plays strategy:

B1: 3 p − 1(1 − p ) = 4 p − 1
B

B2: 2 p + 4(1 − p ) = 4 − 2 p
B

B3: 5 p − 2 (1 − p ) = 7 p − 2
B

Solving for p :
4 p −1 = 4 − 2 p
⇔ 6p = 5
From the graph you can see that the optimal expected return for
5 Anna occurs at the intersection of the lines for Barry playing
⇔ p=
6 strategies B1 and B2.

5 1
So Anna should play A1 with probability and A2 with probability
6 6
⎛5⎞ 20 14 1
Value of the game: 4 ⎜ ⎟ − 1 = −1 = =2
⎝6⎠ 6 6 3

Finding a mixed strategy using the Simplex algorithm

The graphical method cannot be used if player A has more than two strategies so we must use the simplex
method.

Let the probabilities of each of A’s strategies be p1 , p2 , p3 etc.

If there are any negative elements in the payoff matrix, add k to make them all positive.

Model the situation as a linear programming problem:

Maximise V-k (this is the expected payoff from the original game)
Subject to V ≤ expected payoff from B playing strategy 1
V ≤ expected payoff from B playing strategy 2 etc
p1 + p 2 + p3 + ... ≤ 1 (because the sum of the probabilities cannot exceed 1)
p1 , p2 , p3 ≥0 (no negative probabilities)

This information can now be put into an initial simplex tableau and solved in the usual way.

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

REVISION SHEET – DECISION MATHS

ALLOCATION PROBLEMS AND THE HUNGARIAN ALGORITHM

The main ideas are covered in Before the exam you should know:
AQA Edexcel MEI OCR • That you will be working on a matrix or array of
D2 D2 D2 numbers called the payoff matrix.
• How to reduce a matrix to one containing zeros and
interpret this to give an optimal allocation.
• How to apply the Hungarian algorithm to minimise
The main ideas in this topic are: costs.
• Extending the ideas introduced in • That the matrix of numbers must be square before
matchings to find a maximum matching you can apply the Hungarian algorithm.
and its associated cost. • How to deal with maximisation problems by
subtracting all the entries in the matrix from a
• Using the Hungarian algorithm to find an constant.
optimal, minimum cost solution.

Allocation problems are solved by reducing the payoff matrix by subtracting the least value in each row (or column)
form all the entries in that row (or column). You now have a matrix showing relative costs and the least entry in each
row (column) will now be zero. These zeros will represent a minimum “cost” solution. A good way to see whether you
have a maximum matching is to find the minimum number of lines needed to cover all the zeros in the matrix.

Solution: Firstly reduce the rows

Example
P Q R S
A 0 0 8 7 Subtract 20 from each entry
XYZ taxis have four drivers, Alan, Betty, Chris and
B 3 0 12 4 Subtract 20 from each entry
Dave. One morning they have four bookings to
collect people (P, Q, R, S) and take them to the C 10 9 0 6 Subtract 17 from each entry

station for the 0800 train. The table shows the time D 1 0 1 2 Subtract 24 from each entry

in minutes it will take for each driver to drive form

their house to the customer’s house and then This is not a maximum P Q R S

transport each customer to the station. matching since all the A 0 0 8 7

zeros can be covered with B 3 0 12 4
P Q R S
3 lines. C 10 9 0 6
A 20 20 28 27
D 1 0 1 2
B 23 20 32 24
Reduce columns
C 27 26 17 23
D 25 24 25 26 P Q R S This is now a maximum
A 0 0 8 5 matching.
The manager of XYZ taxis wants to minimise the B 3 0 12 2 Alan – P, Betty – Q
Chris – R, Dave – S
total time taken, who should he allocate to each C 10 9 0 4 Total time = 20 + 20 + 17 + 26
journey? D 1 0 1 0 = 83 minutes

The Hungarian algorithm

1. If the payoff matrix is not square, add in dummy row(s) or column(s) of equal numbers to make it square.
2. Subtract the minimum entry from each row from all the entries in the row.
3. If necessary repeat step 2 for the columns.
4. Draw the minimum number of horizontal and/or vertical lines to cover all the zeroes.
5. If the number of lines is equal to the number of columns in the matrix, the positions of the zeroes indicate an optimal
matching; if not, go to step 6.
6. Augment the matrix: find the smallest element not covered; subtract it from the non-covered elements and add it to
any elements covered by 2 lines.
Go to step 2.

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

Note: If you need to maximise, subtract every number in the original payoff matrix from the largest number in the
matrix before applying the algorithm. The final payoff is then found by adding the original payoffs in the cells used in
the optimal matching.

Example: The average scores for five members of a quiz team are shown in the table:

music sport geography history General

knowledge
Alan 17 19 18 15 16
Brenda 20 18 15 19 17
Cally 13 17 17 16 14
David 12 16 18 15 14
Edwin 14 16 15 16 15

A different person must be chosen to answer questions in each of the five rounds in the final. Using their past
performance, who should do each round in order to maximise score?

(a) Explain why you should replace each entry x by 20 – x before using the Hungarian Algorithm
(b) Form a new table by subtracting each number from 20. Use the Hungarian algorithm to allocate the sports team
members.
(c) State the expected score for the team based on their practice scores.

Solution: (a) Because the problem is a maximisation problem and 20 is the largest number in the matrix.
(b)
music sport geography history General
knowledge
Alan 3 1 2 5 4
Brenda 0 2 5 1 3
Cally 7 3 3 4 6
David 8 4 2 5 6
Edwin 6 4 5 4 5

Reduce rows Reduce columns

music sport geography history General music sport geography history General
knowledge knowledge
Alan 2 0 1 4 3 Alan 2 0 1 4 2
Brenda 0 2 5 1 3 Brenda 0 2 5 1 2
Cally 4 0 0 1 3 Cally 4 0 0 1 2
David 6 2 0 3 4 David 6 2 0 3 3
Edwin 2 0 1 0 1 Edwin 2 0 1 0 0

Solution is not optimal as all zeros can be covered with 4 lines

Augment by 1

music sport geography history General

knowledge
Alan 2 0 1 3 1
Brenda 0 2 5 0 1
Cally 4 0 0 0 1
David 6 2 0 2 2
Edwin 3 2 2 0 0

It now takes 5 lines to cover the zeros so matching is optimal

Alan – sport, Brenda – music, Cally – History, David – Geography, Edwin – General Knowledge

(c) Expected scores: 19 + 20 + 16 + 18 + 15 = 88

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

REVISION SHEET – DECISION MATHS

NETWORK FLOWS
Before the exam you should know:
The main ideas are covered in
• What is meant by source, sink and capacity.
AQA Edexcel MEI OCR
D2 D1 D2 • What a cut is.
• How to find an initial flow.
• The meaning of a flow augmenting path and how to
The main ideas in this topic are: find them.
• Modelling flows using bipartite graphs.
• How to use the labeling procedure.
• Finding the maximum flow through a network.
• What is meant by excess capacity and back capacity.
• The maximum flow-minimum cut theorem
• What is meant by a saturated arc.
• The maximum flow – minimum cut theorem.
• How to insert a super-source and super-sink into a
network.

The algorithm for finding a maximum flow

1. Always start with an initial feasible flow, found by inspection.
2. Label each arc with
• the flow along it, shown by an arrow pointing back towards the source
• the excess capacity, which is the amount by which the flow could be increased, shown by an arrow
pointing forward towards the sink
3. Systematically look for flow augmenting paths and mark these on your network using the labelling procedure
4. When all paths are blocked by saturated arcs you have found the maximum flow.

Example A 3 C
In this the directed network 5 4
a) What is the maximum flow along the path SACT? 2
3
S T
b) Find an initial flow of value 7. 4
8
c) Find the maximum flow in the network 6

d) What are the capacities of these cuts B 4 D

Solution 3
A 3 C
3
a) The maximum flow along SACT is 3
5 0 4
(this is determined by the arc of least capacity on the path). 3 2
2
b) A flow of 7 (shown on the diagram) is 3
S 1 T
SACT with capacity 3 2 4 4
8
SBDT with capacity 4 4 6
0
4 4
B D
4

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching
each specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1
3
3 4
c) Flow augmenting paths A C 3
0 4
5 01
SBCT with capacity 1 4 3 21
2
1 2 2
SBCDT with capacity 1 3 1
SABCDT with capacity 1 2 10
S 3 T
21 1 4 3 4
1 2
0 2 8
6 54 3 Diagrams showing the flow
6
0 augmenting paths can be
This gives a maximum total flow very messy. Try to keep
of 10. The flow is shown on this 4 yours as tidy as possible
diagram, along with the saturated arcs. 4 5 and always list the flow
B D 6
4 augmenting paths you have
A 3 C used

4 4
1
3 Cuts
S T A cut partitions the vertices into two sets, one containing the
2
6 source and one containing the sink.
6
The capacity of a cut is the total of all the cut edges with
4 D
direction going from source to sink
B
A 3 C
Find the capacity of the cuts shown on the directed network:
5 4
Note that only three cuts have been shown here, but there are many
more cuts in this network. 2
3
S T
C1 is the cut {S}, {A, B, C, D, T} 4
8
It has capacity 5 + 6 = 11 6

C2 is the cut {S, B}, {A, C, D, T} Note that we do not add the C1 B 4 D
capacity of arc AB as it is directed C2
It has capacity 5 + 0 + 3 + 4 = 12 from the sink side of the cut to the
C3
C3 is the cut {S, A, B}, {C, D, T} source side

It has capacity 3 + 3 + 4 = 10

Maximum flow- minimum cut theorem.

The theorem states that the maximum flow in a directed network is equal to the capacity of the minimum cut.
In the example above the cut C2 is the minimum cut and it has a value 10. This confirms that the flow of 10 found in (c)
above is the maximum flow.
A 3 C
S1 5 4
Networks with many sources and sinks
T1
2
9 4 8
If there is more than one source (S1 and S2 on 4 3
the diagram) or sink (T1 and T2 on the diagram) S T
4
you must introduce supersource (S) and/or 10 8 13
supersink (T). T2
5
S2 6
SS1 must have a capacity 5 + 4 = 9 B 4 D
SS2 must have capacity 4 + 6 = 10
T1T must have capacity 4 + 4 = 8
T2T must have capacity 8 + 5 = 13

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching
each specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

REVISION SHEET – DECISION MATHS 2

LINEAR PROGRAMMING: THE SIMPLEX ALGORITHM 1
Before the exam you should be able to:
The main ideas are covered in
• Formulate a linear programming problem to
AQA Edexcel MEI OCR maximise the objective function, subject to the given
constraints
D2 D1 D2 D1 • Use slack variables to convert inequality constraints
into equations
• Set up the initial simplex tableau
The main idea in this chapter is: • Perform the Simple algorithm for maximising an
objective function
Using the Simplex algorithm to deal with linear • Identify initial, intermediate and final tableaux and
programming problems with more than two variables. know when the solution is optimal
• Interpret the values of the variables and the objective
function at any stage in the Simplex method
• Clearly state the solution to the original problem

Simplex Method for Maximisation Problems

Getting started: Formulation
Translating a real life problem into a linear programming problem is called formulating the problem and is an example of
mathematical modelling. Each problem must have clearly defined variables, an objective function and is subject to
certain constraints
Slack Variables
In order to enable problems to be converted into a format that can be dealt with by computer, slack variables are
introduced to change the constraint inequalities into equalities. Each vertex of the feasible region would then be defined
by the intersection of lines where some of these variables equal zero.
The Simplex Method
The Simplex Method starts at one vertex and systematically moves round all the vertices of the feasible region,
increasing the objective function as it goes, until it reaches the one with the optimal solution. This is easy to visualise on
a 2 dimensional problem, but can be generalised to include more variables. Once there are more than two variables, a
graphical approach is no longer appropriate, so we use the simplex tableau, a tabular form of the algorithm which uses
row reduction to solve the problem.

The Simplex Algorithm

1. Represent the problem in a tableau Note on finding pivot column (step 2)
2. Use the objective row to find the pivot column You can choose any variable in objective row with negative
3. Use the ratio test to find the pivot element coefficient, but it is usual to pick the most negative. Give priority
4. Divide through the pivot row by the pivot element is to the original rather than slack variables.
5. Add/subtract multiples of the transformed pivot row Note on ratio test (step 3)
to/from the other rows to create zeros in the pivot Divide each R.H.S. value by the corresponding element in the
column pivot column, ignore negative ratios and division by zero.
6. Repeat until no negatives in objective row Choose row with the smallest ratio as the pivot row.
7. Read the solution from the table

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.
 the Further Mathematics network – www.fmnetwork.org.uk V 07 1 1

Example:
A manufacturer makes three products x, y and z which give profits of £160, Staff time Storage
£120 and £120 per tonne respectively. Production is constrained by (hours /tonne) (m3 /tonne)
availability of staff and storage as summarised in this table: x 5 5
y 5 3
Formulate a linear programming problem. z 6 4
availability 30 20
Objective function Constraints
Objective function is maximise P = 160x + 120y + 120z Subject to: 5x + 5y + 6z + s1 = 30 (staff time)
⇒ P − 160x − 120y − 120z = 0 5x + 3y + 4z + s2 = 20 (storage)
x ≥ 0, y ≥ 0, z ≥ 0 s1 ≥ 0, s2 ≥ 0

Solving the problem

Considering the problem in the example on the previous page, we must now set up an initial tableau
Setting up the initial tableau
P x y z s1 s2 RHS Ratio test Look for the largest negative value in the
objective row. This is in the x column, so we
1 -160 -120 -120 0 0 0 will choose to increase x first.
0 5 5 6 1 0 30 30/5 = 6 Since 4 < 5 we choose 5 as the pivot element.
The idea now is to make the pivot 1 so divide
0 5 3 4 0 1 20 20/5 = 4 the pivot row by 5.

First iteration
Look for the largest negative value in the
P x y z s1 s2 RHS Ratio test objective row. This is in the y column, so we
row 1 + 160 × row 3 1 0 -24 8 0 32 640 will choose to increase y.
Since 5 < 10 we choose 2 as the pivot
row 2 − 5 × row 3 0 0 2 2 1 -1 10 10/2 = 5 element.
The idea is to make the pivot 1 so divide the
row 3 0 1 0.6 0.8 0 0.2 4 4/0.4 = 10 pivot row by 2.

Second iteration
P x y z s1 s2 RHS The solution is optimal since there are no
row 1 + 24 × row 2 1 0 0 32 12 20 760 negative values in the objective row.

row 2 0 0 1 1 0.5 -0.5 5

row 3 – 0.6 × row2 0 1 0 0.2 -0.3 0.5 1

Reading the tableau

The final tableau represents the following set of equations
Row 1: P + 32z + s1 + s2 = 760
Row 2: y + z + 0.5 s1 – 0.5s2 = 5
Row 3: x + 0.2z – 0.3 s1 + 0.2s2 = 1
You can check your solution by
substituting the values obtained for x, y
The most obvious solution to this is obtained by setting the “basic” variables
and z into the original objective function
(columns with zeros and a single 1) equal to the RHS and setting the “non- to check that the profit is correct:
basic” variables (columns with more than one non-zero entry) equal to 0. P = (160 × 1) + (120 × 5) + (120 × 0)
This gives the solution P = 760, x = 1, y = 5, z = 0, s1 = 0, s2 = 0 = 160 + 600 + 0
= 760
Interpreting the solution
In order to maximise his profit the manufacturer should make one tonne of product x, five tones of product y and no
product z. this would use all the available resources and would generate a profit of £760.00

Disclaimer: Every effort has gone into ensuring the accuracy of this document. However, the FM Network can accept no responsibility for its content matching each
specification exactly.