Join 3,438 readers in helping fund MetaFilter (Hide)


Help me better describe dynamic scientific processes in general terms
February 25, 2014 4:33 PM   Subscribe

I'm looking for scientific or mathematical examples, ideas, which could rightly refer to the imagined class of dynamic systems I'm trying to describe.

(Preventing me from naming this imagined class is I only have an impression of difference in dynamic systems and a few generally grouped examples).

A recent post here, Failure, reminded me of a book by Dietrich Dorner titled The Logic of Failure. (Basic Books, 1997). One of the takeaways from this book was the difficulty for human cognition to comprehend dynamic processes.

From this and some other research on dynamic systems, I've been working to better describe these systems and divide them into sub-classes by investigating how other scientific domains refer to dynamic processes which have a particularly difficult quality--they don't "stop" or are generally not subjected to human intervention.

In this imagined class of dynamic systems I thought to put astrophysics, and that part of the physics of the very small that says "we can't know both the speed and position of electrons", and global economics, and environmental biology, and global weather to name a few examples. For better understanding I've tried to contrast this imagined class of dynamic processes to human-scale sciences and their Earth-based events. For example, while the travel of a bullet is a dynamic event, it has a beginning and end--when its fired and when it stops. On the other hand the astronomical observations of planetary orbits are not discrete (in my relative argument)--there is no event in the orbit marking its end or beginning. The global economy, business, commerce is constantly "happening" and doesn't "stop" so that an economist might wish to ask some fundamental question, like how much money is there in circulation right now? (and certainly I'm aware that at some level even if you could stop all economy, the count would be prone to error as much as any election is prone to counting error--but these are different problems).

Such distinctions are arbitrarily convenient for me too. (Its a useful example if its useful. If your first reaction is, "That's just not right! They're not bullets. They're projectiles." That's certainly true, but not relevant. ) I'm the first to admit my use of terms may be unorthodox (maybe even heretical) and cause some confusion (aggravation), thus it is more useful to interpret my question in relative terms. Don't think of the canon, just consider how you might describe dynamic processes better.

I like to follow the "multi-model" approach, so I'm not married to any particular perspective. Since I've been thinking about this, its not a topic I hear referred to commonly. I suspect it rarely comes up (I'm sure for good reasons). Nevertheless, I'm interested in what posters might suggest I investigate. And, MeFier's have such broad experience and knowledge and interests, I think in some way I'm asking people who might also be interested in the answer.

// PS. I was admonished for 'thread sitting'--I'm new to askmetafilter, so I'll avoid this practice.
posted by xtian to Science & Nature (8 answers total) 1 user marked this as a favorite
 
In manufacturing they often refer to "batch" vs "continuous" process.

And in software there's "continuous deployment" and "devops" vs "releases".
posted by straw at 5:00 PM on February 25 [1 favorite]


Maybe you should look at some books about control theory.
posted by zscore at 5:06 PM on February 25


What about cellular processes like photosynthesis, conversion of sugar into energy, etc.?
posted by bleep at 5:07 PM on February 25


On second thought, what's probably more helpful is the literature on "complex systems."
posted by zscore at 5:14 PM on February 25


Pretty much all of Ecological science is like that. When the field first got going, it was an article of faith among biologists that ecologies were stable; that over decades or centuries the system would find an equilibrium.

It took a while for them to finally admit that this is rare. Most ecologies ring, and keep ringing. They never settle down. (Even disregarding the effects of varying weather from year to year.)

One of the earliest described examples is an ecology in Finland and NW Russia where the critical players are voles, owls, and foxes. The voles are extremely prolific breeders; females often raise several batches of young per year, and they mature so fast that daughters from the first batch often raise their own first batch before winter sets in.

The owls and foxes don't breed as fast, so the voles outrace them and eventually run up against a limit of food supply. Then their population stabilizes as excess voles starve.

With huge food supply (lots of voles), the fox and owl numbers continue to rise over a period of years, and as a result more and more excess voles become fox/owl food and fewer starve. Eventually the foxes and owls overbreed.

Then they overpredate the voles, whose numbers crash. With little food available, owl and fox numbers crash. With the predators mostly gone, the cycle resets with the voles breeding like mad.

The cycle is about 12 years long and it's been going on as long as anyone can tell.

These days they use chaos theory to analyze this kind of ever-dynamic ecology.
posted by Chocolate Pickle at 5:17 PM on February 25 [2 favorites]


From this and some other research on dynamic systems, I've been working to better describe these systems and divide them into sub-classes by investigating how other scientific domains refer to dynamic processes which have a particularly difficult quality--they don't "stop" or are generally not subjected to human intervention.


Some systems, like a weight on a spring subject to air friction, will oscillate around their final position for a while before stopping. Other systems, like a weight on a spring with no friction, will just oscillate forever. How this looks to an observer depends on how often you measure the system. It's not clear from your examples what property of a system you're trying to use for categorization.

If you want to seriously study dynamical systems, you need to use math. Look up "dynamical systems". In those terms, the categorization you describe could be expressed in a lot of ways, e.g. as "systems with long time constants vs systems with short time constants" or "systems with limit cycles vs systems with fixed points", etc.
posted by serif at 1:06 PM on February 26 [1 favorite]


@Serif, Its not so much a single property of a system (and your answer helped my response) I'm thinking about. Its the addition of a property--the difference which makes a difference--(here "no friction") when the system jumps from one category to another.

Its this point where the language, or, in absence of a sufficient vocabulary, class categories are useful (as evident by your answers). And this helps me appreciate @Chocolate Pickle's answer. Any group or individual could possess some portion of the narrative CP describes and they might miss the realization this plays out in a "12 year" "cycle".

Its the recourse to figurative language which is partially driving this inquiry to the community. Other answers such as
"continuous deployment" & "devops" vs "releases" & "batch"
"cycles"

All in all very good thinking. Thank you all.

PS. I'll drop one more of my inspirations, Wolfram's Classifications for Cellular Automata. Thank goodness the rest of the universe is more complex.
posted by xtian at 3:16 PM on February 27


For your other categories of dynamic systems you might consider consulting "Systems Theory" of course. Watch the movie Mindwalk for inspiration there.
Also, for more-or-less closed systems, like a factory floor, look into "Theory of Constraints". This has more to do with how slowly kinking a water hose on one side of the system can boost or slow processes on the other. (In a Rube Goldberg kind of way).
posted by AskAnswerAct at 6:51 PM on February 27


« Older I drive a Mobilio Spike (2005)...   |  Do we have to leave our Charte... Newer »

You are not logged in, either login or create an account to post comments