Naked Science Forum
General Science => General Science => Topic started by: random_soldier1337 on 18/02/2023 20:41:22
-
Hi, I understand there is computational research as well as completely theoretical research compared to experimental research. I was curious about what they are like. Could anyone point out some good papers?
Thanks.
-
That question is so broad that it is hard to give a meaningful response.
For a starter, the title mentions engineering/physics research
- While the body of the question only mentions computational research
- The title only mentions theoretical research, while the body adds experimental research
- Most theoretical research in any field involves computation
- Most experimental research (including in computation) involves building things
Some extreme examples:
- Manipulating real stars and galaxies is far beyond our capabilities, so cosmology is one area of research which is dominated by theoretical & computational methods.
See: https://en.wikipedia.org/wiki/Illustris_project
- Manipulating real protons and lead ions has become more effective over the past 50 years. So the Large Hadron Collider at CERN does experiments in nuclear physics. Processing and storing and analysing the results involves lots of computation, which is why CERN originated the concepts between today's World Wide Web, and Cloud Computing.
See: https://en.wikipedia.org/wiki/CERN
-
I should clarify when I say computational, I mean computer simulation being the major/only component.
Regarding the title, I was reading some papers on radiation detection as well as application of computational fluid dynamics for part of the system design within some of those papers. I wasn't really sure whether they count as engineering or physics.
Also, If I was solely to base my understanding of what research is from those examples, it would seem that there is both computational as well as experimental in every paper. But from what little I understand, there are papers that are purely theoretical as well as those that are purely computational. If those are something that exist, I was hoping someone could point out papers which serve as good examples of the method of doing such research.
Cosmology is probably an obvious example. I was hoping for research where one could do experimental but has chosen to submit simulation results only (again only if such a thing exists). If you are aware, you could probably give an example in the fields I mentioned (radiation detection and CFD). I just didn't want to restrict you to them.
-
both computational as well as experimental in every paper
Computation has become such a common way of studying phenomena that there is even a name for it: in silico
Traditionally, biology experiments have been done:
- in vivo: in living organisms
- in vitro: in a glass dish (eg Petri dish)
- in situ: as found in nature
We now have:
- in silico: Done in a computer (silicon chips being the basis of today's computer hardware)
https://en.wikipedia.org/wiki/In_silico
-
Yes, that's correct! In the modern age, computer simulations have become a very common way to study and understand various phenomena, including in the field of biology. These simulations are often referred to as "in silico" experiments, which is a play on words using the Latin term "in vivo" (meaning "in the living") and the fact that modern computers are based on silicon chips.
In contrast to traditional biological experiments done in vivo (in living organisms), in vitro (in a glass dish or other controlled environment), or in situ (in the natural environment), in silico experiments allow researchers to simulate complex systems and study their behavior under different conditions. This can help scientists to gain insights that might not be possible to obtain through traditional experimental methods alone.
-
There is a danger that significant decisions can be taken on the basis of a model that follows and extrapolates from the recent smooth behavior of an inherently chaotic system. You might for instance attempt to forecast energy prices on the apparently rational presumptions that
the Russian authorities would prefer to sell gas rather than declare war,
that they would obviously win that war quickly,
that energy retailers would not used the war as an excuse for naked profiteering,
that there is a competitive market for gas and electricity
that the US embargo on Venezuelan exports would remain
and that nobody else would increase production once the price was attractive.
All of which are obviously false. But you can't produce a stable model unless you assume them to be true!
My thoughts on climate modelling are well known. Mao Tse-Tung famously said "sound analysis impossible without intimate knowledge of life and without real understanding of the pertinent contradictions".
-
a model that follows and extrapolates from the recent smooth behavior of an inherently chaotic system
There are a lot of chaotics systems around us - and a lot of models assume equilibrium conditions
- The whole field of economics was based on the assumption that supply and demand converge to an equilibrium
- But it really is a chaotic system
I am listening to an interesting series of podcasts called "Simplifying Complexity" which is about fractal/chaotic systems, addressing effects in economics, biology etc - a series of interviews with staff at the Santa Fe institute.
Web-page index: https://podcasts.apple.com/au/podcast/simplifying-complexity/id1651582236
(Other podcast players are available...)
-
I am familiar with chemistry, and there is a LOT of computational chemistry (some is based on quantum mechanical models, some on classical "force field" models, and some using statistical models). Many publications include both experimental and computational components because it is good practice to test the validity of the simulation models against the real world. There's an adage something along the lines of, "experiment can show results (what happened), but rarely mechanism (how it happened), while computation shows possible mechanisms, but rarely identifies results (how something could happen, but not if it does)"
Since it sounds like the OP is requesting purely computational papers, I would draw their attention to this open access and purely theoretical journal:
https://www.mdpi.com/journal/computation
Note: the subject matter spans a very broad range, and is highly technical, so don't expect to understand any of it on a first read :-)