Computer simulations. How do they support theory?

[Robh (Rob)]

We seem to be relying increasingly on computer simulations to model complex problems. Simulations are being used to model the weather, climate change, earthquake activity, the evolution of the Universe, the evolution of galaxies and the distribution of dark matter. Consider this article about a simulation that describes how dark matter is distributed in a galaxy. The simulation assumes that dark matter exists.
http://www.sciencedaily.com/releases...0609073156.htm

In any computer simulation, we have to set and prioritise what we think are the important initial conditions. A model of the theory is reduced to mathematical expressions that operate on the data. More often, computer simulations use an iterative process whereby data produced in one run becomes the input for the next run. Many simulations require a supercomputer due to an enormous amount of data being modified over a huge number of iterations.

How can a simulation support a particular theory? I suspect that the verification of a theory can only occur if a simulation modeling it makes an observable prediction which can be confirmed. My worry is that many simulations are scaffolding an imaginary Universe without any real-world confirmation. For example, the simulation in the link above.
Anyone else with any thoughts on this?

Regards, Rob.

[xelasnave]

My thoughts, without even having any experience in this field and therefore not corruptted by any real facts, is the predictive process although with its good points could have folk adjusting things to move things in the direction they seek.

AS I have said often science can not be corruptted but humans can and that does not mean anyhting other than not all folk will play the game without cheating to win.

I am very suspicious of the models produced to support climate change simply because of the various folk who seek profit from the fear mongering.

Carbon trading is there to do one thing in spite of how it is presented and that is to raise the price of energy to open the door for currently expensive neuclear power...in my view.

It is cleaer the NP folk set out to side line coal by geting the greenies focused upon it as being bad offering in its place a lovely clean efficient fuel... and they are winning the game folk who would nevcer consider NP now think well maybe we should think about NP otherwise we will lose the Great Barrier Reef and the Polar bears.

To think this country with a contribution to the carbon footprint of a mere 1.5% is to made to feel guilty by restating the situation on a per capita basis is simple trickery the mob buys ...I suspect the computer models developed by the NP lobby could be slanted to support their addgendah...

But on the positive humans have the ability to think things thru and hopefully temper results with a consideration of the benefits that may accrue to those commissioning construction of the model.

alex

[Coen]

Simulations are a fantastic aid in:
- Engineering, saving billions of dollars in design and prototyping costs (e.g. aircraft design)
- Visualisation, allowing the implications of complex concepts to be distilled in a form readily digestible
- Plus many others...

However the simulations must have a level of confidence attached to them in order to then be able to understand what their outputs can tell us and how far we can extend them. The resulting formal verification, validation and accreditation (VV&A) process is very difficult/expensive for many simulations but necessary especially if lives are involved.

Many science simulations serve a different purpose and thus usually do not need a formal VV&A with an equivalent of a wet-finger in the air enough. The basis of their accreditation, besides peer review, is how well does the simulation support/replicate what is observed i.e. can it be tied back to known experimental/observed results and other established phenomena. If it can not or it does not make predictions that can be then explored then it is perhaps a nice conceptual art piece serving a purpose different to that of scientific exploration.

When a simulation is established in a particular field (name a field and you'll find one that has been around since, well, a long time) then it often becomes a basis for other simulations i.e. can the new simulation X produce results consistent with simulation Y that has been around since the beginning of the field. This domino approach can be good but does need to be checked periodically lest various assumptions become hidden and later jump out and bite the user(s). Extrapolation beyond assumption basis can lead to dangers. Another factor is that computers resort to numbers and simulations use real numbers (non-integer), for irrational numbers it is impossible to store them accurately in a computer thus any errors have the potential to compound as time goes on (see chaos theory: an early met simulation led to the "butterfly flapping its wings" comment based on the very different results obtained from a very small change in initial conditions). Plenty of books on the topic.

Simulations are good but as with any tool, the right tool for the job at hand is important and not all tools will do all jobs.

Simulations work on the principle of a deterministic universe, the universe is not necessarily deterministic thus simulations are ultimately always an approximation.

[Robh (Rob)]

Originally Posted by xelasnave

My thoughts, without even having any experience in this field and therefore not corruptted by any real facts, is the predictive process although with its good points could have folk adjusting things to move things in the direction they seek.

alex

As a simulation depends on its base data and the mathematical expressions selected to model it, there is certainly the danger that data and expressions are used selectively, whether consciously or unconsciously, to produce a desired outcome. One would hope the scientific community would monitor this but how many people would have a nuts and bolts understanding of the particular simulation in question?

Regards, Rob.

[Coen]

Originally Posted by Robh

As a simulation depends on its base data and the mathematical expressions selected to model it, there is certainly the danger that data and expressions are used selectively, whether consciously or unconsciously, to produce a desired outcome. One would hope the scientific community would monitor this but how many people would have a nuts and bolts understanding of the particular simulation in question?

Regards, Rob.

How many understand the mechanical internals of an aircraft? a car? a lawnmower? or some other piece of machinery?

You can if you want to but many do not, content to use until it no longer works and have a mechanic they trust to look after it. If it does not work to some standard (tangible and/or intangible) then that manufacturer will soon be out of business.

Often things are built from known pieces put together in various ways likewise with simulations. Many are built on various libraries and the libraries are generally well documented, tested etc. This enables further confidence to be obtained. Nevertheless short of every person using the simulation reading every line of code making it up, understanding it and so forth then a level of black-box abstraction is essential. Still if the simulation in question does not give expected results in simple test cases then warning flags will be raised.

It also comes back to levels of documentation available especially with regards to the assumptions made, nature of the inputs required and the nature of the underlying model(s)/mathematics.

Ultimately the "market" is "self-regulated" (peer review etc) so the adage "buyer beware" applies.

[Robh (Rob)]

Originally Posted by Coen

Another factor is that computers resort to numbers and simulations use real numbers (non-integer), for irrational numbers it is impossible to store them accurately in a computer thus any errors have the potential to compound as time goes on (see chaos theory: an early met simulation led to the "butterfly flapping its wings" comment based on the very different results obtained from a very small change in initial conditions). Plenty of books on the topic.

You make a good point here.
Rounding off numbers can produce large errors over many iterations. For those not familiar with the term, an iteration is where data produced from one run is then re-used and modified again on the next run. I would also think that any small change in initial conditions could produce a markedly different final outcome in a simulation that has a large number of iterations, the so called Butterfly Effect in Chaos Theory. A danger which can arise from scientists being rather selective of initial data that may be in question.

Regards, Rob.

[Robh (Rob)]

Originally Posted by Coen

.... Nevertheless short of every person using the simulation reading every line of code making it up, understanding it and so forth then a level of black-box abstraction is essential. Still if the simulation in question does not give expected results in simple test cases then warning flags will be raised.

It also comes back to levels of documentation available especially with regards to the assumptions made, nature of the inputs required and the nature of the underlying model(s)/mathematics.

Ultimately the "market" is "self-regulated" (peer review etc) so the adage "buyer beware" applies.

Sorry, missed your last post as I was doing mine at that time.
Thanks for your view. I'd agree entirely.

Rob

[SMR (Steve)]

Hi Rob

Originally Posted by Robh

Consider this article about a simulation that describes how dark matter is distributed in a galaxy. The simulation assumes that dark matter exists.
http://www.sciencedaily.com/releases...0609073156.htm

This article, which appeared in an on-line popular science magazine and is based on other popular science sources, clearly states that the results are not about a simulation, but about the development of a "much more reliable model" of the density distribution and other properties of dark matter in galaxies that is NOT based on a simulation. So in this case at least, your question is unfounded.

As to your general question about the validity of simulations in verifying other scientific theories, here are two responses:

- There are many problems in physics that cannot be solved by closed-form equations. One excellent example is the three-body problem. The only way to work out an accurate ephemeris of the objects in our solar system, for example, is to run a "simulation". This process is called numerical integration, and groups like JPL have perfected this process. There's nothing magical or "theoretical" about it -- it's based on Newton's simple law of gravity. (OK, with a little relativity thrown in, mostly for Mercury.)

Guess what? The same problem applies to modelling the movement of stars in galaxies. The only way to do this is to use the same kind of "simulations". Again, these models are simply applying well-established (at least in our local galactic neighbourhood) physics to observed stellar motions in galaxies.

As the article points out, the problem of stellar motions in galaxies has been known since Zwicky in 1933. The only ways that the models can match reality is to introduce more mass -- dark matter -- or to invent new theories of gravity. The vast majority of astrophysicists have chosen the first option. And, yes, they might yet be proven wrong. Anyone who can come up with a new theory of gravity that explains what we see can expect a call from the Nobel committee. That's the way science works. 76 years later, they haven't called any of the advocates of new gravity theories because their ideas have not stood up to scrutiny.

What the researchers here seem to have done -- and I remind you that this is article is a rewrite of another popular science article -- is create a mathematical model of dark-matter distribution, density, temperature, etc, without relying on numerical integration which requires guessing these parameters then comparing them against observations. As such, it's very interesting work. I just wish that they'd cited the original source so that we could read it ourselves.

(My guess, BTW, is that the author of the article preferred the word "simulation" to "numerical integration" because it would be more understandable to the general audience. Would you have reacted the same way if the world simulation had not appeared?)

- I think you underestimate the power of peer review in the scientific world. If a group publishes results based on a poorly constructed simulation, there will be dozens of other groups around the world who will leap at the opportunity to point out every little flaw. The science community is much, much more aggressive and territorial than the general community thinks. The majority of academics and researchers around the world will fight over every bit of status and prestige that they can get because that's their "currency" -- they're not in it for the money.

Final questions: why did you feel the need to misrepresent the article, and question the validity of a whole bunch of other areas of research, when the article clearly said the opposite of what you claimed? Why the red herring from you?

Steve.

[bojan]

Originally Posted by SMR

- I think you underestimate the power of peer review in the scientific world. If a group publishes results based on a poorly constructed simulation, there will be dozens of other groups around the world who will leap at the opportunity to point out every little flaw. The science community is much, much more aggressive and territorial than the general community thinks. The majority of academics and researchers around the world will fight over every bit of status and prestige that they can get because that's their "currency" -- they're not in it for the money.
Steve.

I also think this is the case here.
it is very dangerous to be opinionated about things if they are not understood properly... or if our understanding is based on "popular science".
I am not saying we should not discuss those things, but.. we can not judge them (the individuals who are dealing with those things on a daily basis are excepted of course)

[Enchilada]

Originally Posted by Robh

How can a simulation support a particular theory? I suspect that the verification of a theory can only occur if a simulation modeling it makes an observable prediction which can be confirmed. My worry is that many simulations are scaffolding an imaginary Universe without any real-world confirmation. For example, the simulation in the link above.
Anyone else with any thoughts on this?

Regards, Rob.

Here is an interesting web simulator, which might answer your questions on simulators.

This is the N-body simulator, which I've used regarding some papers on multiple stars. Fairly easy to use and add parameters too. It runs in Java,
showing effects like interactions, perturbations, relativistic effects, etc.

It might seem rater complex, but it is a useful example of simulations in astrophysics.

Some Examples
http://nbodylab.interconnect.com/

Some N-body samples are accessed on the ;

http://nbodylab.interconnect.com/nbl...ml#triple-pyth
http://nbodylab.interconnect.com/nbl...ns_sa.html#Run

The binary star example is;
http://nblg6.webhop.info/cgi-bin/ast...art_mode=kz5_4

An explanation of the parameters is at;
http://nbodylab.interconnect.com/nbl...es.html#Inputs

[Enchilada]

Originally Posted by bojan

I also think this is the case here.
it is very dangerous to be opinionated about things if they are not understood properly... or if our understanding is based on "popular science".
I am not saying we should not discuss those things, but.. we can not judge them (the individuals who are dealing with those things on a daily basis are excepted of course)

I totally agree. Sadly, this is happening everywhere these days. Just because someone doesn't understand some underlying theory doesn't mean it is made up or inconsequential. The sudden rise of pseudo-science is becoming a serious issue, as neither democracy or popular opinion applies to the scientific method nor in how science operates. Unfortunately, it gets in the way by the influence of "the mob", which directs diverts the funding of science to less productive outcomes or new advances.

The danger is of course groups with other agendas, like intelligent design, non-scientific biology (I.e. against cloning or stem cells), or with the electronic universe proponents.

Furthermore, the education system or the media doesn't help either, as it tends to dumb down science for the masses without understanding the observations, methods, or conclusions.

[renormalised (Carl)]

Be very careful about using computers to support anything, theory or otherwise. They're only as good as the data that they're given and their output is commensurate with that data. It's the old adage "GIGO". You can make a computer simulation/program prove/support anything...and nothing at all.

What they're great for is visualisation, constructing representations of theoretical models.

Like all tools, they're only as good as those that use them.

[DJDD]

good points raised earlier.

Originally Posted by Enchilada

The danger is of course groups with other agendas, like intelligent design, non-scientific biology (I.e. cloning or stem cells), or with the electronic universe proponents.

would you expand upon your comments that cloning and stem cell (research, I presume) is "non-scientific biology"?

[bojan]

Originally Posted by DJDD

would you expand upon your comments that cloning and stem cell (research, I presume) is "non-scientific biology"?

I would not subscribe to this either...
Those disciplines are just "applied biology".
However, they are still 100% science - they are subject to peer review etc.

[Robh (Rob)]

Originally Posted by SMR

Hi Rob
Final questions: why did you feel the need to misrepresent the article, and question the validity of a whole bunch of other areas of research, when the article clearly said the opposite of what you claimed? Why the red herring from you?

Steve.

Steve,
You make many good points. Ouch!
I would agree with you that the scientific community puts the results of any new research under the microscopic and are very critical of claims that cannot be established through evidence. I guess the same applies here on IIS!
It was not my intention to create a red herring, nor to devalue the research done by scientists. From my understanding of the referred article, the results were more reliant on new equations and functions but still with the assistance of a computer to produce the overall results. I assumed that if certain theoretical assumptions were made, values given to constants and equations applied within a computer program then it was still some form of simulation. Perhaps, it was a poor example.
However, in the research discussed in the article, there was an assumption of the existence of dark matter. There are other competing theories e.g. MOND that may yet prove to be correct.

I guess my general question was about the increasing reliance on computer simulations in describing the real world and about some of the initial assumptions that were being made. In my thread starter, I stated "My worry is that many simulations are scaffolding an imaginary Universe without any real-world confirmation ... Anyone else with any thoughts on this?". My intention was to simply to open a discussion on the increasing role of simulations and their validity in supporting theories.

Thanks for your input, Rob.

[renormalised (Carl)]

Originally Posted by Robh

Steve,
You make many good points. Ouch!
I would agree with you that the scientific community puts the results of any new research under the microscopic and are very critical of claims that cannot be established through evidence. I guess the same applies here on IIS!
It was not my intention to create a red herring, nor to devalue the research done by scientists. From my understanding of the referred article, the results were more reliant on new equations and functions but still with the assistance of a computer to produce the overall results. I assumed that if certain theoretical assumptions were made, values given to constants and equations applied within a computer program then it was still some form of simulation. Perhaps, it was a poor example.
However, in the research discussed in the article, there was an assumption of the existence of dark matter. There are other competing theories e.g. MOND that may yet prove to be correct.

I guess my general question was about the increasing reliance on computer simulations in describing the real world and about some of the initial assumptions that were being made. In my thread starter, I stated "My worry is that many simulations are scaffolding an imaginary Universe without any real-world confirmation ... Anyone else with any thoughts on this?". My intention was to simply to open a discussion on the increasing role of simulations and their validity in supporting theories.

Thanks for your input, Rob.

Go back to my post above...that sums up the use of computers.

[Robh (Rob)]

Originally Posted by renormalised

Be very careful about using computers to support anything, theory or otherwise. They're only as good as the data that they're given and their output is commensurate with that data. It's the old adage "GIGO". You can make a computer simulation/program prove/support anything...and nothing at all.

What they're great for is visualisation, constructing representations of theoretical models.

Like all tools, they're only as good as those that use them.

Carl,
As usual, straight and to the point!

Thanks, Rob

[SMR (Steve)]

Hi, Oh Spicy One

Originally Posted by Enchilada

The sudden rise of pseudo-science is becoming a serious issue, as neither democracy or popular opinion applies to the scientific method nor in how science operates.

It's worse than that, IMO. Democracy itself requires a well-informed populace. Many of the decisions facing the human race can only be answered by science, but the populace don't understand science.

Originally Posted by Enchilada

The danger is of course groups with other agendas, like intelligent design, non-scientific biology (I.e. cloning or stem cells), or with the electronic universe proponents.

Yep. The issue is not pseudo science. It's politics and religion that are at at play here, which are both way beyond the limits of this forum so I'll end my comments here.

Steve.

[bojan]

Originally Posted by SMR

Yep. The issue is not pseudo science. It's politics and religion that are at at play here, which are both way beyond the limits of this forum so I'll end my comments here.

Steve.

Exactly right

[SMR (Steve)]

Hi Rob

Originally Posted by Robh

From my understanding of the referred article, the results were more reliant on new equations and functions but still with the assistance of a computer to produce the overall results.

You're using a computer to communicate with me. Does that automatically make your output suspect?

Seriously ... the article clearly stated that the work was a collaboration with the Applied Mathematics Department. Last I checked, applied mathematicians have stepped beyond using slide rules, or rooms full of women adding up numbers. Some of them have really, really big super computers to play with. And guess what -- they know what GIGO means.

Originally Posted by Robh

I assumed that if certain theoretical assumptions were made, values given to constants and equations applied within a computer program then it was still some form of simulation.

It's pretty clear that you don't know what a computer program is. Written much code recently?

Originally Posted by Robh

However, in the research discussed in the article, there was an assumption of the existence of dark matter. There are other competing theories e.g. MOND that may yet prove to be correct.

You're right -- competing theories may be correct. The interesting thing about this research, assuming that it's been reported correctly, is that it shows that there is no need to use simulations to explain the properties of dark matter that we observe in studying galactic dynamics.

Originally Posted by Robh

I guess my general question was about the increasing reliance on computer simulations in describing the real world

Do you have any evidence to show that this statement is true? Or are you just really looking for a way to discredit ideas that you don't understand, and extend that to other topics that you feel political about? Honest answer, please.

Thanks!

Steve.