ity

rs
ve

y
CAMBRIDGE INTERNATIONAL AS & A LEVEL IT: COURSEBOOK

op
ni
U

C
ge

w
BEFORE YOU START

ie
id

ev
br
• Do you know how to use a spreadsheet?
am

-R
• Do you understand that spreadsheets can be used for modelling?
-C

• Are you able to test formulas within a spreadsheet?

s
es
• Are you able to test validation rules within a spreadsheet?
y

Pr
op

ity
C

Computer Aided Design (CAD) enables designers to

Introduction
rs
w

produce a model of a physical object. This could include

ie

ve
a kitchen, a building, a motor vehicle or an aeroplane.

y
A computer model is a representation of a real-world
ev

CAD software can include features such as viewing

op
ni

process. A model is created through mathematical

an object in two dimensions (2D) or three dimensions
R

analysis of the real-world process. A simulation is the

C
(3D), manipulating objects within the model, adding
use of a computerised model to predict how a real-life
ge

or removing objects, viewing the model from different

w
system might behave. Details in a computer model can
angles, applying different effects such as colour and

ie
id

be easily changed to see what effects they would have.

lighting, and focusing on specific features such as the

ev
br

This is often cheaper and safer than trying out changes

electrical cabling or heating system within a building.
on the real-world process.
am

-R
Modelling software will include some essential features:
-C

KEY WORDS • the ability to change variables within the software

s
es

model: a representation of a process • asking what-if questions to see what the result of
y

Pr

changing variables might be

op

simulation: using a model to predict real-

• formulas and functions to carry out the
ity
C

life behaviour
mathematical calculations that form the basis
rs
w

of the model
ie

ve

automatic recalculation of formulas and functions

y
•
ev

op
ni

9.1 Characteristics of • rules that define how the model behaves

R

• layers of abstraction so that different parts of the

ge

modelling software
w

model can be viewed and analysed separately.

ie
id

Modelling software is used to create a model. An example of the use of modelling is to create a model
ev
br

Spreadsheets can be used to create computerised of a roller coaster. Variables can include the height of
am

-R

models, but there are also custom-written solutions that each drop, the radius of loops, the starting speed of the
are used to model specific processes. carriage, length of each section and the weight of each
-C

carriage. Calculations will be used to define rules such as

s

For example, a network simulator such as Cloonix can

es

the amount of friction and how that will slow carriages

be used to model a computer network. The software
y

down, the effect of gravity on carriages as they move up

Pr
op

produces a diagrammatic view of the devices connected and down and the g-force that will be experienced by
to the computer network. It will be possible to identify passengers. What-if questions could be asked such as
ity
C

the Internet Protocol (IP) addresses used on the network ‘What would happen if we increased the starting speed
rs
w

and how they are assigned through a Dynamic Host by 2 km/h?’ or ‘What would happen if we increased
ie

Configuration Protocol (DHCP) server. Wireless

ve

the initial drop by 5 m?’ The effect of these changes in

y

networks can be added and security can be configured

ev

variables can then be modelled.

op
ni

to see what the effects will be. Devices can be connected

R

to specific switches and the throughput of data traffic

C

can be analysed.
e

w
g

ie
id

ev
br
am

-R

178
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
9 Modelling

op
ni
U

C
ge
Constructing buildings, kitchens, gardens, motor

w
PRACTICAL ACTIVITY 9.01 vehicles and other objects can be a very costly process. It

ie
id
Experiment with a digital single-lens reflex is important to get it right first time as any adjustments

ev
br
(DSLR) camera simulator by changing the
am that have to be made in the real world will incur

-R
variables and seeing how this affects the end financial costs. Models help to experiment with different
results. You can find an example simulator on the designs to see which look is most aesthetically pleasing
-C

and which ones react best to environmental conditions.

s
Cameraism website.

es
When designing a kitchen using modelling software,
y

variables such as lighting, worktops, cupboard doors,

Pr
op

position of units, flooring and tiling can be changed

9.2 The need for to see which configuration looks the best. Cupboards,

ity
C

drawers, dishwashers and cookers can be opened to see

rs
w

if they use the space efficiently. The angle of view can

computer models
ie

ve
be changed in a 3D environment to see what the kitchen

y
ev

might look like from a variety of angles.

op
ni
There are a variety of reasons why models might be
R

used, including training, forecasting and construction.

U

Models can be used to predict population growth.

C
Whatever the reason, a model must be able to answer It’s important for governments to know what future
ge

w
what-if questions. populations will be so they can plan for housing,

ie
id

Models can be used for the purpose of training people transport infrastructure and public services. Scientists

ev
br

to use equipment. This could range from learning to model population growth within natural habitats. They
can identify the existing population and reproduction
am

drive a forklift truck to flying an aeroplane or operating

-R
a nuclear power plant. Models are used for training rates and predict what future populations will be. There
will come a point when the population will outgrow
-C

because there is less risk of injury to the trainees and

s

instructors than learning in a real environment. When its carrying capacity (number of inhabitants that can
es

learning to drive a forklift truck, the trainee may make survive on limited resources) and the growth rate will
y

Pr

adjust. This is known as density-dependent growth. If

op

a mistake by lifting a load too high and toppling the

truck. In real life this could cause serious injury as well there are no limiting resources, then the growth rate
ity
C

as costing money in terms of repairs, but using a model would be expected to increase exponentially. This is
known as density-independent growth.
rs
w

removes both of these risks. Costs are saved because

ie

ve

real equipment does not suffer wear and tear, fuel is not
y
ev

required to operate machinery and instructors do not PRACTICAL ACTIVITY 9.02

op
ni

need to be present all of the time. It is also possible to test

R

the trainee in unpredictable situations such as turbulence Experiment with the Wolfram population growth
C

model, which you can find on the Wolfram

ge

for an aeroplane or driving a heavy goods vehicle on ice.

w

Unpredictable situations that would happen in real life Demonstrations Project website, by changing
ie
id

can also be removed for trainees when they first start the initial population, growth rate and carrying
ev
br

learning so they can focus on basic controls. capacity variables.

am

-R

Models can also be used for forecasting. One of the most

common models for this purpose is weather forecasting. Climate change models consider the interactions
-C

Patterns that have happened in the past can be analysed between the atmosphere, oceans, land, ice and the sun
es

along with current data to predict what the weather might to estimate climate trends. Climate change models are
y

be in the future. Businesses and governments use financial

Pr

used for long-term predictions and it can be decades

op

modelling to predict what might happen to profits or the before it will be known if the model is accurate. Models
ity
C

economy. It is too risky for a business to make sudden will include several scientific rules such as the first law
changes in the marketplace without testing them out of thermodynamics, the Stefan-Boltzmann Law, the
rs
w

with a model first. A variety of what-if questions can be Clausius-Clapeyron equation and the Navier-Stokes
ie

ve

asked using a model to determine how to make the most equations, but they include so many uncertainties
y
ev

op
ni

profit. Variables that could be changed include selling and are so complex that there is no known exact
R

prices, adjusting the quantity of products to supply at any

U

solution. These complex models require the use of

C

given time, times of year to sell products and the effect of supercomputers. Climate change models differ to
e

weather on seasonal sales. weather forecasting in that they predict average weather
g

ie
id

ev
br
am

-R

179
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
CAMBRIDGE INTERNATIONAL AS & A LEVEL IT: COURSEBOOK

op
ni
U

C
ge
conditions over a long period of time rather than

w
specific weather conditions on a given day. Models from PRACTICAL ACTIVITY 9.04

ie
id
the past can be tested against what really happened to Go to the Onramp traffic simulation website and

ev
br
see if the model was accurate, enabling the models to be
am try out their model of road traffic flow.

-R
improved for the future.
Identify the variables that can be changed and
-C

what the effects are of changing these.

s
PRACTICAL ACTIVITY 9.03

es
What questions could you ask of the model?
y

Search on Youtube for the video of the

Pr
op

installation of the final phase of the Met Office

ity
C

Supercomputer to see the scale of just one of

The Met Office Hadley Centre’s three Cray SC40
9.3 What-if analysis
rs
w

supercomputers which are capable of 14,000

ie

ve
trillion calculations a second. What-if analysis is the process of asking the question

y
ev

op
ni
‘what would happen if … ?’.
R

C
Models can be used to help with queue management.
ge

KEY WORD

w
This can include queues at a supermarket, telephone

ie
id

helpline queues and amusement parks. Being able to what-if-analysis: experimenting with changing

ev
br

predict the number of customers and having rules variables to see what would happen to the
that define how long it takes to serve each customer
am

output if those variables changed

-R
can help to determine how many staff are needed to
serve the customers. Queues can be quite simple as in
-C

a supermarket or they can be complex where different Some examples of what-if questions in a financial
es

customers are given different priorities. These rules such spreadsheet model might include:
y

Pr

as ‘first in first out’ or ‘shortest job first’ will form part

op

• What happens to the total income for the month if

of the model. The behaviour of customers can also form we increase the selling price by 20%?
ity
C

part of a model, including balking (deciding not to join

a queue if it is too long), jockeying (switching between • What happens to our costs if we use a different
rs
w

queues) and reneging (leaving a queue). Variables will supplier for one of our parts?
ie

ve

include number of customers, number of servers, time to

y
ev

• What happens to our total expenditure for the next

op
ni

serve a customer and expected number of customers at

five years if we move to new premises?
R

different times of the day.

C

Part of what-if analysis includes the use of goal seek.

ge

Models can also be used for traffic flow. This can

w

It is possible to set a goal and find out what needs to be

help with planning new roads, improving existing
ie
id

done to achieve that goal. Some examples of goal-seek

roads or building new shopping centres or residential
ev

questions might include:

br

areas. Models will take into account existing traffic

am

-R

movement and projected changes in the number of • What price do we need to sell at in order to sell
vehicles expected along a route. The model can be used 5000 items per month?
-C

to experiment with different traffic control measures

s

• How many items do we need to sell to break even

es

such as traffic lights, traffic islands or traffic calming (zero profit and zero loss)?
y

measures. Transport planners can then see what effect

Pr
op

making changes to the roads would have without

committing to projects that might not be suitable. KEY WORDS
ity
C

Variables within the model can include length of

rs
w

vehicles, vehicle speed, average acceleration of vehicles goal seek: seeing what a variable needs to
ie

ve

and congestion. Different modelling techniques are change to for a goal in terms of output to
y

be achieved
ev

used, such as microscopic models that model the

op
ni

behaviour of single vehicles, and macroscopic models

R

that model the behaviour of a stream of vehicles.

e

w
g

ie
id

ev
br
am

-R

180
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
9 Modelling

op
ni
U

C
ge

w
PRACTICAL ACTIVITY 9.05

ie
id
Open 9.01 bike hire.xlsx which is a model to calculate the profit that would be made for different numbers

ev
br
of bikes being hired.
am

-R
1 a What would happen to the profit if the number of short male mountain bikes booked was
-C

changed to 70?

s
es
b What if there were no short male mountain bikes booked? Did any other data change when you
y

Pr
changed the quantity booked?
op

2 What would happen to the profit if the junior hourly rate was increased to £12.50?

ity
C

rs
w

3 Experiment with changing other variables within the spreadsheet model and watch what happens to
ie

the outputs.
ve

y
ev

op
ni
R

C
ge

WORKED EXAMPLE 9.01

w
ie
id

9.01 bike hire.xlsx calculates the profits for a bike hire company. In this example, the company wants to see how

ev
br

many bikes of a particular type they need to hire out to increase their profit to £2500:
am

-R
-C

s
es
y

Pr
op

ity
C

rs
w
ie

ve

y
ev

op
ni
R

Figure 9.1: Goal seek configuration.

ge

Figure 9.2 shows the result of the goal seek being 65.63333 bikes, which would of course need to be rounded up to
ie
id

70 bikes:
ev
br
am

-R
-C

s
es
y

Pr
op

ity
C

rs
w
ie

ve

y
ev

op
ni
R

Figure 9.2: Goal seek result.

C
e

w
g

ie
id

ev
br
am

-R

181
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
CAMBRIDGE INTERNATIONAL AS & A LEVEL IT: COURSEBOOK

op
ni
U

C
ge

w
PRACTICAL ACTIVITY 9.06

ie
id
Open 9.02 student marks.xlsx which shows exam and coursework marks for students in a class.

ev
br
am
Use goal seek to answer these questions.

-R
1 What coursework mark would John need to increase his total mark to 390?
-C

s
es
2 John resits exam 3. What mark does he need to increase his total to 425?
y

Pr
op

3 Sameena resits her coursework. What mark does she need to get a total of 380?

ity
C

4 Fathima hasn’t taken exam 3 yet. What mark does she need to get a total of 250?

rs
w

5 Experiment with setting your own goals.

ie

ve

y
ev

op
ni
R

C
9.4 Test a spreadsheet model
ge

w
In Chapter 8 you learned how to choose test data and to test formulas, functions and validation rules within a

ie
id

spreadsheet. When testing a spreadsheet model, it is important to test all aspects of the model. This could include:

ev
br

• formulas and functions • graphs and charts

am

-R
• validation rules • cell and worksheet protection
-C

• conditional formatting • macros.

es
y

• form controls
Pr
op

A test plan needs to be created to test all these aspects of the spreadsheet model. It needs to include clearly identified
ity
C

input values and the expected output values.

rs
w
ie

ve

WORKED EXAMPLE 9.02

y
ev

op
ni
R

9.03 Block Paving.xlsm produces a quote for block paving a driveway. Here is an extract of a test plan for the model:
U

C
ge

Number Description Type Input Value Expected Result

w
ie
id

1 Area calculation Valid Input Screen: Area on Invoice = 84

ev
br

Length = 12, Width = 7

am

-R

2 Customer Name Valid Input Screen: Invoice customer name = ‘Mr. Marzi’
on Invoice Customer Name = ‘Mr. Marzi’
-C

s
es

3 Length Invalid Input Screen: Error Message = ‘Only use real

validation numbers (e.g. 12.3)’
y

Length = ‘twelve’
Pr
op

4 View Invoice Button Input Screen: Invoice displays in print preview

ity
C

Button Click on ‘View Invoice Button’ mode with same data from input
rs

screen
w
ie

ve

y
ev

op
ni
R

C
e

w
g

ie
id

ev
br
am

-R

182
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
9 Modelling

op
ni
U

C
ge
• goal-seek to find out what variables need to be

w
PRACTICAL ACTIVITY 9.07
changed to achieve a given outcome.

ie
id
Open 9.03 Block Paving.xlsm

ev
br
Other advantages of spreadsheet models include the
am
Follow the test plan in Worked Example ability to share the spreadsheet with colleagues easily

-R
1
9.02 and see if the actual results match the so that many people can experiment with the model,
and the fact that most organisations already own
-C

expected results.

s
spreadsheet software which reduces the need for training

es
2 Create another four entries for the test plan and purchase costs.
y

Pr
in Worked Example 9.02
op

Spreadsheets do have their limitations though. They are

3 Ask somebody else to follow your test plan only as accurate as the formulas and functions that are

ity
C

entries exactly as they are written and see the rules that represent the real world. Unless extremely

rs
w

if there is anything you could have put complex rules are used, then a spreadsheet model will
ie

ve
more clearly. never be an accurate representation. These complex

y
ev

rules require mathematical and computer expertise in

op
ni

Extension 1: add cell protection, additional order to set them up and it can take a very long time
R

form controls and conditional formatting to the

C
to create models that are truly representative of the
spreadsheet. Extend your test plan to cover
ge

real world. Spreadsheets can only be used to simulate

w
these new features of the model. numbers, but cannot simulate the effect on objects.

ie
id

ev
Extension 2: open 9.04 Theme Park Queue
br

Model.xlsm and create a test plan to test the Questions

am

-R
theme park queue model. You will need to spend
a bit of time understanding how the model works 1 Identify three characteristics of modelling software.
-C

and how the graph and conditional formatting 2 Describe, using examples, how a spreadsheet
es

respond to input values. could be used to model the costs or savings to an

y

organisation of moving premises.

Pr
op

ity
C

9.5 Evaluating the 9.6 Simulations

rs
w
ie

ve

y
effectiveness of
ev

Advantages and disadvantages

op
ni
R

spreadsheet models of using a model to create and

ge

There is a variety of tools within a spreadsheet that can run simulations

ie
id

be used to help with modelling:

ev
br

A simulation is the use of a computerised model

• variables that can be changed to ask what-
am

-R

to predict how a real-life system might behave. As

if questions with modelling, simulations can be used for training,
-C

• formulas and functions which define the rules forecasting and construction.
s
es

of the model Advantages include:

y

Pr

• graphs and charts which can show a graphical

op

• expensive prototypes or the real thing do not need

representation of the forecast to be created in order to experiment with different
ity
C

• instant, automatic recalculation of formulas, variations and answer what-if questions

rs
w

functions and graphs/charts to answer what-if • changes to the model can be made very quickly and
ie

ve

questions immediately the effect can be seen just as quickly

y
ev

op
ni

• conditional formatting to highlight the effects • alternative models and designs can be used to see
R

how they react differently

e

w
g

ie
id

ev
br
am

-R

183
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
CAMBRIDGE INTERNATIONAL AS & A LEVEL IT: COURSEBOOK

op
ni
U

C
ge
• unusual events (for example, earthquakes) can be wind direction or wind speed and emotional reactions

w
tested without the need to wait for them to happen to being evacuated. However, planners can experiment

ie
id
in real life with lots of what-if questions in order to plan for a

ev
br
large variety of circumstances, which means that when
• equipment does not suffer from wear and tear or
am

-R
a natural disaster does occur people can be better
damage when being tested or experimented with
equipped to deal with it.
-C

• dangerous situations such as aeroplane equipment

s
es
failure can be simulated without putting Pilot training
y

people in danger

Pr
When it comes to large aircraft, it can cost thousands
op

• simulations can be ‘sped up’ so that effects can be of pounds just to take off, fly and land. This cannot be

ity
C

analysed over a long period of time without having repeated too often as it will become too costly to train
to wait for that period of time to elapse.

rs
pilots. Flight simulators can help by removing the fuel
w

costs associated with flying. They can also remove the

ie

ve
Disadvantages include:

y
danger that a trainee pilot might pose if they make a
ev

op
ni
• the way a simulation reacts is only as good as the mistake while in the air.
R

model it is based upon

C
Flight simulators not only include software, but very
ge

• simulation software and equipment can be very specialised equipment which can cost hundreds of

w
expensive to purchase thousands of pounds. The equipment will be designed

ie
id

• people need to be trained to use simulation to react in a similar way to a real aircraft, so that the

ev
br

equipment and software pilot can feel the physical effects of any movements
am

-R that they make (including take off, landing, turbulence

• complex models take many years to develop, or even a crash landing) in a way that is as close to real
-C

especially if they are designed to react like the life as possible. The software will include the rules of
s

real process
es

the model that define how the aircraft should react in a

y

• it is impossible for a simulation to be a perfect variety of circumstances.

Pr
op

representation of the real-world process. Pilots can also practise landing and taking off at
ity
C

airports they have not visited before, including some

of the most dangerous airports in the world such as
rs
w

The use of simulation Toncontin Airport in Honduras. Simulations give pilots

ie

ve

y
the opportunity to respond to malfunctions such as an
ev

Natural disaster planning

op
ni

engine failure, cabin pressure failure or landing gear

R

When planning for natural disasters, people need to failure. These would be far too dangerous to attempt in
C

know what the effects of a natural disaster might be. real life.
ge

It is impossible to do this in real life without the actual

ie
id

disaster happening and, as they are natural disasters, it

ev
br

is impossible to force them to happen. If the planners

am

-R

wait for the natural disaster to happen, then it is too

late to plan.
-C

Simulations based on models of natural disasters such

es

as earthquakes, volcanic eruptions, hurricanes, bush fires

y

Pr

and tsunamis can be used to see what the effects might

op

be. Planners can experiment with different variables such

ity
C

as wind speed and direction to see how quickly a fire

might spread, which will help them to plan evacuations
rs
w

and firefighting.
ie

ve

y
ev

As with all simulations, these rely upon the accuracy

op
ni

of the model. There will be many things that the model

R

Figure 9.3: Toncontin Airport.

C

can’t predict completely, such as sudden changes in

e

w
g

ie
id

ev
br
am

-R

184
-C

s
es

Copyright Material - Review Only - Not for Redistribution

 ity
rs
ve

y
9 Modelling

op
ni
U

C
ge
Car driving ‘What happens if I increase the temperature?’ and the

w
outcomes can be seen. It’s also possible to speed up the

ie
id
Simulators can be used to help to learn to drive a car. In
simulation so that rather than waiting hours, days or

ev
most countries, learner drivers start immediately on the

br
even years to see what the effects might be, results can be
road having never used any of a car’s controls before.
am

-R
seen much more quickly.
They immediately have to deal with hazards such as
other cars, pedestrians, cyclists, potholes and dangerous Nuclear science requires a lot of computing power
-C

s
junctions. In addition to this, other drivers on the road in order to simulate nuclear reactions. For example,

es
are held up because of the learner driver being slow reactors convert uranium and plutonium through
y

Pr
and hesitant. nuclear fission which involves millions of collisions
op

every microsecond. Even with supercomputers, nuclear

Simulations can enable a learner driver to become

ity
C

scientists cannot represent all of these collisions in a

familiar with the controls of a car and dealing with
simulation.

rs
w

hazards before taking a car onto the road for the first
ie

ve
time. A simulator is a safe environment and could

y
PRACTICAL ACTIVITY 9.08
ev

reduce insurance premiums for driving instructors. The

op
ni

simulator could also be used when the learner driver

R

Look online for a nuclear power plant simulator.

C
needs to attempt manoeuvres such as turning in the road
An example can be found on the Dalton
ge

or reversing around a corner for the first time, rather

w
Nuclear website.
than having to do this straight away on a real road. Even

ie
id

after a learner driver has passed their test, they could

ev
br

use simulators to learn how to drive on a motorway,

Questions
am

drive in icy conditions and how to handle a car if it gets

-R
out of control, such as in a skid.
3 Describe two advantages of using a simulator for
-C

Simulations can also be used for racing drivers to pilot training.

es

experiment with different car setups to find the optimum

4 Describe one disadvantage of using a simulator for
y

Pr

configuration for a racing circuit, rather than risking

op

learning to drive.
crashing. Racing drivers would be able to take the car to
ity
C

the limit knowing that the virtual crash will not result
in an injury. However, a simulation will never behave REFLECTION
rs
w

exactly the same as the real car as the rules of the model
ie

ve

will never be perfect. Therefore it is still necessary to 1 Consider the implications of not being able
y
ev

to model the effectiveness of changing the

op
ni

practise using a real racing car.

timing of traffic lights, but instead having to
R

try it out over a period of time.

Nuclear science research
ge

Nuclear science is very dangerous. Even a small 2 Think about situations that would be too
ie
id

accident could expose a person to radiation that could dangerous for a pilot to learn how to react to
ev
br

cause serious disfigurement or death. It is therefore in a real aircraft but would be better suited
am

to a flight simulator.
-R

not possible to experiment with nuclear reactions

in the real world in the hope that something might 3 Discuss what other models and simulations
-C

work. Simulations can be used to try out different

s

are used in real life.

es

nuclear reactions by adjusting the coolant temperature,

y

changing the way the control rods are used and the rate
Pr
op

of reaction. What-if questions can be asked such as

ity
C

rs
w
ie

ve

y
ev

op
ni
R

C
e

w
g

ie
id

ev
br
am

-R

185
-C

s
es

Copyright Material - Review Only - Not for Redistribution