Lesen Sie diese Seite auf Deutsch
Advantage and Benefits of Computer Simulation in the 21st Century
The benefits of computer simulation is becoming more and more evident. It has an enormous potential, and the prerequisites for fruitful implementation are excellent. Simulation expert Dr. Schuler tells from trends and examples.
Today, computer simulation should not be tied to a single precise situation. It has become much more interdisciplinary. In a multitude of procedures and models, and in nearly all scientific branches, computer simulations have taken roots: From aerodynamics to zymosis, a long list could be enumerated.
For simplicity, let us mention three well-established simulation areas together with some illustrative examples.
At the beginning, the calculations were static. This can be explained by the compute power available at this time. To determine the load capacity of a new bridge, a force grid calculation without any illustration is sufficient. Twenty years ago, an architect was able to impress his audience with a few model pictures of a building. Today, photo-realistic montage or even video animations are a must.
Or the models were strongly simplified or measurable data was missing. Nowadays many measures can be taken, thanks to laser, satellites, tomographs, etc. that the beginning of last century one couldn't even have dreamed of. This numerous devices lead to an excellent data basis to start with a model to simulate.
Where does this lead to? The following chain depicts a progression, that will be repeated in many industry lines.
Development of Computer SimulationAt a very early time - but only by means of immense compute power and adequate precision - this principle was realized for weather forecasting.
That is why today's wheather forecasts are more reliable than twenty years ago - especially the long-term forecasts.
The benefit is evident: «A reliable prediction enables better decisions and saves consequential costs of false estimation!»
But how can one be sure that computer simulation leads to promising - i. e. correct - predictions? Every simulation is based on a model. The model shall be able to reflect and predict reality. Would anyone find a model to correctly predict stock quotes, it would be easy to be rich. But unfortunately it is not that simple.
«Complex problems cannot be solved with a simple model.»
A hard, but promising way is to provide the model with the necessary profoundness. If more compute power and dynamic computer simulation can reach timely or other measurable dimensions, an adjustment to the real world is possible and the model may be furthermore and interactively improved. The simulated model may realise what simulation predicts wrong and correct its model for future predictions.
Sometimes to evaluate single but real events is feasible by simulation only. Not even a single time it would be desirable to make a real but fatal experiment. In such a case simulation can provide much information about risks and security measures.
The manyfold and even increasing potential we estimate therefore very promising: Computer simulations advance further in prognostic dimensions and help saving costs!
What if ..?
Future simulations and models will allow a "what-if" question. They will be able to tell the consequences of heavy rain, power black-outs, or stress hormons. And they will be able to tell where to effectively excert leverage before disaster happens.
A flooding like in August 2005 in Switzerland with 2.6 billon consequential costs could anytime repeat, as nobody is able to banish torrential rain. But it should be possible to simulate the waters in soil and streams in order to be ahead of the real water fronts. In such a way, alerting and evacuation of residents, protection of infrastructure, and counter-measures should be more precise, and well ahead in time. Thus, figuratively, the sand bags would be present before the water.
Anyone involved, and insurance companies, could benefit greatly.
Dr. Lukas Schuler
Dr. Lukas Schuler is scientist and owner of a didactic certificate of ETH Zürich. He simulated various lipid aggregates on a molecular level and significantly improved the available molecular model.
Replication allowed only with written agreement with xirrus