A computer model is a representation of a system or process created on a computer, to assist with calculations and predictions. A computer simulation is the execution of that computer model with data over a period of time to generate results data.
The complexity of computer models can vary from small, such as a few calculations in a spreadsheet for personal finances, to massive amounts of numbers and complicated calculations such as studying the potential effect of climate change over the next 1000 years.
Computer models rely on variables and processes to create predictions that can help people make decisions in the real world. Given that people may use computer models to make important decisions, it is imperative to know exactly how accurate a model is.
Applications for Computer Models
These models can analyse traffic levels at different times of the day and allow for predictions of change based upon potential changes to the infrastructure (for example, adding a new road). Other variables such as noise levels or pollution can also be deduced from such models. To ensure these models are as accurate as possible, they need to have up-to-date data on the types and speeds of vehicles on the road as well as where and when humans might be moving during the day.
This type of model can be used to accurately predict how strong buildings and bridges can be in the case of natural disasters such as earthquakes or heavy storms/hurricanes. Details about the construction of the buildings such as their materials and the land that they are built on all need to be entered into the model to make the predictions as accurate as possible.
The structure of black holes and supernovas in space can be modelled to help back-up or debunk theories about how they behave. Scientific understanding can be improved over time using this method in a scenario where it would be very difficult/impossible to observe manually.
Medical researchers can use models to help understand bacteria and viruses that can cause diseases. Rather than testing drugs on human subjects, they can first test the effects using computer models.
Car crash models can be used to see what effect an accident would have on a car, it's passengers and pedestrians. Computer model tests are currently used alongside real world physical tests to test many more possible outcomes based on different input variables of speed, time, object materials being crashed into etc... By using a computer model, the testers need not destroy a new car each time a test is completed. As well as that benefit, a computer model can analyse far more variables than a manual real world physical test can.
Computational Fluid Dynamics
CFD models the flow of liquids and gasses, and is commonly used to test aero dynamics. Examples of this in industry are race car testing, turbines and aircraft designers. All of these use CFD to fully understand the flow of air around a physical item. NASA also make use of CFD to understand how their space shuttles will manage when they reenter the Earth's atmosphere. Testing this without computer models would be prohibitively expensive.
Climate Models and their Development
Used for modelling the Earth's climate for purposes such as predicting the weather, or predicting climate change. Data such as radiation from the sun, greenhouse gasses, ocean currents, temperature changes and more allow scientists to inform governments to help make decisions that can save money and save the planet.
Decision Support Systems
In the world of Business, decision support systems are often used to answer questions about future performance by using historical data to identify patterns. These systems can answer 'what if' questions such as "what if we increased our production costs by 2%" or "what if we hired an additional 20 employees?". Other areas these systems can be used include analysing the potential effects of a marketing campaign.
Why Use Models?
Although they seem to be very complicated, computer models can often be too simple. This can become a problem when trying to get consistently accurate results for use in decision making in the real world. For example, a model using data about the sun might need to use how much radiation is emitted. This could be between 0.1 and 0.2%, which may not seem much in the scheme of things, but that 0.1 range in data could have massively different results for a model on climate change. Furthermore, a computer model used to predict climate change could have 1000s of variables entered to generate as accurate predictions as possible, but there is no accounting for the behaviour of humans and what their potential effect could be on the model.
Models are simplified in the case that there are just too many potential variables to handle, which would need a supercomputer to calculate any meaningful sample of data (what would be the point of accurately predicting tomorrow's weather if it took 24 hours to do it?). Also, the ability to represent every possible scenario for a model requires complex coding, particularly for things that we don't have a full understanding of.
All of the models discussed on this page so far have some form of graphical output known as a visualisation. The numeric data generated by a computer model can be turned into a graph or a physical representation of a scenario such as weather patterns or ice movement of a glacier. These visual representations played next to each other can help to show the change over a set period of time. Another good example of this is predicting a missile trajectory.
Combines the computer models of the world, with high quality input and output devices to help a user experience a situation.
These combine the computer models of aircraft structure, flight physics, aerodynamics with a 3D model of the world. The flight simulators try to accurately replicate the cockpit controls of an aircraft. They are used to train pilots before they get chance to fly the real thing!
In much the same way as a flight simulator does, a driving simulator combines computer models of a car, and the potential environment around it. It can be used to help train learner drivers or give refresher courses to those who have not driven for a long time. Hazards such as pedestrians and obstacles can be used to help test a driver's reactions and control of a vehicle.
Racing car teams also use simulators to modify a car's potential design before actually going ahead with producing/modifying the physical version.
Advantages and disadvantages of simulations
Spreadsheet applications such as Microsoft Excel and Apple Numbers can be used to generate computer models. By using basic functions such as MAX, MIN, SUM, -, *, /, AVERAGE and LOOKUP basic computer models can be created. One simple model to create is a student grade record for a teacher. An example of such is shown below: