Chapter 9: Modelling and Simulation

Modelling
A model is a computer representation of a real-world process. A model is created through

mathematical analysis of the real-world process. Modelling software is used to create a

model. Spreadsheets can be used to create computerized models, but there are also

custom-written solutions that are used to model specific processes.

Using What-If Analysis

Computer models all base their model upon a what-if analysis. This involves changing a

model to find out what the results would be, if different input values are chosen. You

need to know how to build, test and change data within models, or change the model

itself to produce different sets of results.

Simple computer models can be created, tested and used in Excel. Some models only

require a change in data.

The Need for Computer Models

There are a number of reasons why computer models are useful and necessary.

Cost
As has already been implied, one of the major reasons we need models is because Of cost.

We always have to pay for our mistakes, if engineers are designing structures such as

bridges and high-rise buildings, or designing cars and planes, it is very expensive to correct

mistakes. If a computer model is used, then the costs of any mistakes practically

disappear.

Testing
Another reason is the ability to test the design. It would be impossible to build several

different types of building just to see which is most likely to survive heavy flooding.

However, it can be done relatively easily with a computer model,

 Prediction
It is impossible to travel through time, so in order to predict events over a large time span,

such as climate change or population growth, computer models are used as they are the

best way of making predictions. Another use of computer model is to accurately predict

when storms will occur and the severity of them, thus enabling early warnings to the

public about how to protect themselves and their property.

Time-efficiency
Modelling also saves time as well as money. Car manufacturers can perform more tests

in a shorter space of time with a computer model than by arranging tests of prototypes

with crash-test dummies inside them over and over again. Computer models make it

possible to run many more tests.

Planning
Simulations are a specific type of computer model and are very useful in planning for rare

or unexpected events. For example, they can be used to help predict and plan for the

after-effects of natural disasters, such as volcanic eruptions, earthquakes, and tsunamis.

Safety
Some situations are dangerous, such as experimenting with a nuclear reactor. A computer

model is a much safer way to do this.

Advantages of Spreadsheet models

I. One of the reasons for this is that formulae are recalculated automatically. It is also

possible to automatically import data into a spreadsheet from a database.

2. Spreadsheet models are very useful for observing patterns which take place over a

long period of time. They are particularly useful when studying animal population

growth, to observe population growth in real time would take many years. If an

accurate computer model is created, the results will be known quite quickly.
3. Another useful property of spreadsheet model is the automatic updating of graphs.

In a business spreadsheet, for example, graphs might be produced which show the

profits, output, costs, and income, among other variables. It is usual to show the

trends over a period of time. These are useful when making predictions about the

future prospects of a company.

4. Spreadsheet model can remove human fallibility. This means that the likelihood of

decisions being made based on personal feelings is no longer a problem. The

obvious benefit of this is that the decisions made by spreadsheet models are

consistent for the same inputs, which is not always the case with different humans.

5. Spreadsheet model templates exist for regularly used spreadsheets. Users can save

time setting up the spreadsheet by using a template which might just need to be slightly amended to
suit the particular situation.

6. Validation rules can also be built into spreadsheet models. This means that data

input errors are reduced as the validation will ensure only data which is reasonable

is allowed to be entered.

Disadvantages of spreadsheets Models

I. It is very difficult to include every possible variable in a spreadsheet model, In

population growth models,

2. It is very difficult to accurately predict patterns. There are numerous variables that

affect the study of a particular species; factors such as the number of predators,

birth rate, illness and so on, all need to be considered.

3. In financial spreadsheet models, there may be a number of variables which are

fairly straightforward to identify. However, there may be problems in predicting

future patterns of spending, borrowing and investing, as it is difficult to accurately

model human behaviour; it is just not possible to quantify or tum it into a number.

4. Another aspect of modelling which makes it difficult to predict trends is the effect

that sudden emergencies will have on real-life behaviours. The banking crisis of 2008

forced governments to make decisions. Some governments reacted by putting

more money in to their economy.

Other governments cut spending, there was no way of predicting how individual governments would
react. As a result, economic

spreadsheet modelling became problematic.

5. Some situations will need software to be specially designed and/or computer

experts to create a model, and neither of these is inexpensive to purchase or invest

in.

6. Spreadsheet users sometimes fail to take into account is the fact that it may take a

very long time to produce a complex model.

7. The spreadsheet model has to be created and then it needs to be tested.

Documentation for the user has to be produced. Finally, the model has to be

improved and updated as necessary. All this contributes towards the effectiveness

of the model, but it can take time.

8. When an expert has been through all these stages, the staff at the organization

which is going to use the spreadsheet need to be trained to use this model. This, of

course, creates an extra expense to the organization on top of paying the creator(s)

Of the spreadsheet model,

9. One Other factor Which affects the effectiveness of a spreadsheet model is human

errors introduced to the model by users. This can cause even the most

sophisticated model to give out inaccurate results when incorrect data is input,

Uses of Computer Models

I. Financial Forecasting
Companies need to be able to make a profit. It is important to them that they can model

possible future events. Details Of income and outgoing Costs can be entered into the

spreadsheet and a formula linking the two can be used to calculate the profit. This is

making use of the what-if feature of spreadsheets. It can also use the automatic

recalculation feature of spreadsheets which means you do not have to ask the computer

to run the calculation every time you change individual prices.

Conditional formatting is another spreadsheet feature that is beneficial in this context. It

allows you to highlight particular values that match particular criteria. Graphs are also

used to illustrate forecasts, such as which goods are likely to make profits over a period

of time, and allow you to compare the profit levels and sales of different goods; you can

see the profit made by each product very clearly. The model can then be used to predict

profits based on the rate of selling.

Another feature of a spreadsheet is the use of absolute and relative cell referencing. This

ensures you only increment the parts of a formula you need to, such as the prices or costs

of individual items to see their effect. One of the most useful features of spreadsheets is

the use of cell protection. This prevents these inexperienced users changing formulae or

data by accident.

Macros are another useful feature of spreadsheets; if a company wanted to compare the

effect of different costs at the same time, a macro can be used which makes it simpler to

create more complex formulae or functions.

More sophisticated financial models are often used by governments to predict the future

Of a nation's economy. Governments can use computer models to see how much extra

money would be raised if they increased taxes. They can do this for a variety of scenarios

2. Population growth
A population-growth computer model tries to predict the population of an organism. It

assumes that the organism reproduces according to a set of rules. There are a number of

variables, such as how often the organism reproduces, how many new organisms it

produces each time, limiting factors (such as limited resources like food and water),

natural death rates and types of predators.

Models provide a way of understanding how populations of organisms change over time

or when compared to each other. Many patterns can be identified by using computer

modelling. Other studies can be made such as the age distribution within a population

and how this change. Computer models are used to calculate the maximum harvest for

farmers and also to help with environmental conservation, Computer models are also
used to understand the spread of parasites, viruses, and disease. They are also useful

when modelling the future of particular endangered species.

3. Weather systems
Weather forecasting used to involve several weather forecasters and days of calculations.

The arrival of satellites and computer models has really helped forecasters to make more

reliable predictions, Weather presenters on television will sometimes refer to the

uncertainty of their predictions because the computer models are not in agreement. The

technical term given to producing weather forecasts using a computer is 'numerical

weather prediction' (NWP).

The computer needs to know the current state of the atmosphere, which is provided to it

through observations from weather stations and weather balloons as well as satellites.

These readings are fed into models that analyse the data and combine it with the most

recent forecasts to produce a 3-D model of the Earth's atmosphere. It uses mathematical

equations that represent horizontal and vertical air motions, temperature changes and

moisture processes among other variables to calculate what the atmosphere might look

like. Computer models have to include hundreds of mathematical equations to produce

a weather forecast.

A lot of computing power is needed to run a computer model. It is usual for

supercomputers to be used in such activities. As we have seen, these supercomputers are able to
perform quadrillions of calculations every second!

4. Climate Change
In order to create a computer model, climatologists have to make several assumptions

about how the climate works. Problems arise because climate is affected by a large

number of variables and relating all these variables to each other is a very difficult task.

When you consider that even the largest supercomputer has its limitations, there are a

number of people who are very sceptical about the accuracy of computer models in this

scenario.
5. Queue Management
Let us consider a supermarket which has an automatic queue measurement system using

computers. A number of the large supermarket chains use such systems. Automatic

queue measurement systems can use people-counting sensors at entrances to the

supermarket and above checkouts. They accurately detect the number and behaviour of

people in the queue.

Built-in predictive algorithms can provide information on how many checkouts will be

needed to move customers through quickly. This is done as the queues are forming and

the information can be displayed on computer monitors or tablet devices. A range Of

information is provided, such as the length of the queue at any given time and how long

the waiting time for customers is. Management teams can be automatically alerted

before the waiting times get too long. Outputs produced by such systems are:

» The number of people entering the supermarket

» Queue length

» Average wait time

» Amount Of time each Checkout operator is idle

» Total wait time.

Other software is available for more complex queuing systems.

6. Traffic Management
Traffic flow depends on two variables: the speed of the traffic and the density of vehicles

on the road (the number of vehicles per given length of road). Traffic flow can be measured using
sensors, (induction) loop detectors, video cameras, and by manually counting the number of
vehicles.

Traffic-flow computer models use a series of complex algorithms. Some models treat each

mode of transport (such as vehicle, bus, train, tram, cyclist or pedestrian) as a unique

object, since each mode has its own behaviour and targets. Other models concentrate on

traffic-flow characteristics such as density, flow, and average speed of the traffic.
7. Construction
Models are frequently used in the construction industry. The most common type of

software used is computer-aided design (CAD). It is used by architects and engineers to

create precise drawings or technical illustrations. CAD software can be used to create two-

dimensional (2-D) drawings or 3-D models. CAD software is used to design Structu such

as bridges and buildings, and can also be used to generate animations and other

presentational material. The software often allows for the inclusion of additional

information, such as dimensions and descriptions of components. CAD packages also

incorporate libraries. A library, in this sense of the word, is a collection of pre-drawn parts

and features that you can insert into drawings.

If a building is being designed, a library could be used which would have parts such as

doors, windows, staircases and roofs available. All that is needed is to select the required

parts, changing the dimensions to suit.

CAD packages produce 3-D views and different viewing angles. Most packages can create 3-D
walkthroughs so that the viewer can

actually, see inside buildings or other structures. Printers are used to produce hard copies

of designs; however, it is more usual for plotters to be used to produce blueprints. More

recently 3-D models can be produced using 3-D printers.

Advantages & Disadvantages Of CAD

1. CAD can allow a user to design a part much faster than traditional drawing.

2. There is a library of shapes and designs so only minor modifications need to be

made. However, work can be lost because Of the sudden breakdown of computers.

3. Using computer-aided design software, it will be much easier to make any changes

because you can fix the errors and modify the drawings more easily.

4. It is also more accurate, the number of errors that used to occur with manual
designs are significantly reduced.

5. However, it takes time for users to learn how to use the software and with every

new release of the CAD software the user has to update their skills which can take

time. It is expensive to regularly update CAD software or operating systems.

6. The different parts of the design can be reused over and over again and CAD tools

make it easier to save files of the drawings and store them in a way that they can

be used time and again.

7. Open-source CAD software is available for free, though it tends to be limited in

scope, but the better forms of CAD software can be expensive to buy, and it costs

a lot of money to train workers to use the software,

Simulation
A simulation is the use of a computerized model to predict how a real-life system might

behave, as with modelling, simulations can be used for training, forecasting and

construction.

Advantages of Simulation
I. It is not necessary to create expensive prototypes or the real thing in order to

experiment with different variations and answer what-if questions.

2. Changes to the model can be made very quickly and the effect can be seen just as

quickly.

3. Alternative models and designs can be used to see how they react differently.

4. Unusual events (e.g. earthquakes) can be tested without the need to wait for them

to happen in real life.

5. Equipment does not suffer from wear and tear or damage when being tested or

experimented with.

6. Dangerous situations such as aeroplane equipment failure can be simulated without putting
people in danger.

7. Simulations can be 'speed up' so that effects can be analysed over a long period of

time Without having to wait for that period of time to elapse.

Uses of Simulation
1. Natural Disaster Planning
When planning for natural disasters, it is necessary to know what the effects of a natural

disaster might be. It is impossible to do this in real life without the actual disaster

happening and, as they are natural disasters, it is impossible to force them to happen. If

the planners wait for the natural disaster to happen, then it is too late to plan.

Simulations based on models of natural disasters such as earthquakes, volcanic eruptions,

hurricanes, bush fires and tsunamis can be used to see what the effects might be. Planners

can experiment with different variables such as wind speed and direction to see how

quickly a fire might spread, Which Will help them to plan evacuations and firefighting.

2. Car Driving
Simulators can be used to learn to drive a car. In most countries, learner drivers start

immediately on the road having never used any of a car's controls before. They

immediately have to deal with hazards such as other cars, pedestrians, cyclists, potholes

and dangerous junctions. In addition to this, other drivers on the road are held up because

of the learner driver being slow and hesitant. Simulations can enable a learner driver to

become familiar with the controls of a car and dealing with hazards before taking a car

onto the road for the first time. This is a safe environment and could reduce insurance

premiums for driving instructors.

Simulations can also be used for racing drivers to experiment with different car setups to

find the optimum configuration fora racing circuit, rather than risking crashing. Racing

drivers would be able to take the car to the limit knowing that the virtual crash will not result in an
injury.

However, a simulation will never behave exactly the same as the real

car as the rules of the model will never be perfect. Therefore, it is still necessary to

practice using a real racing car.

3. Nuclear Science Research
Nuclear science is very dangerous. Even a small accident could expose a person to

radiation that could cause serious disfigurement or death. It is impossible to experiment

with nuclear reactions in the real world in the hope that something might work.

Simulations can be used to try out different nuclear reactions by adjusting the coolant

temperature, changing the way the control rods are used and the rate of reaction. What-

if questions can be asked such as 'What happens if I increase the temperature?' and the

outcomes can be seen. It's also possible to speed up the simulation so that rather than

waiting hours, days or even years to see what the effects might be, results can be seen

much more quickly.