Talk:Complex system

From Wikipedia, the free encyclopedia


This is the talk page for discussing improvements to the Complex system article.
Article policies
Archives: 1, 2
Archive
Archives
1: 2001 - 2004
1: 2007 -


Contents

[edit] The original intention of this article

When I originally wrote this, I framed the whole thing with an emphasis on non-linear dynamics. This seems to have been removed, but at a later state some kind of half educated twaddle about "chaos theory" ( which is a subset of non-linear dynamics) has been chucked in by someone who has only half an idea what they are going on about. The section even contradicts itself.

All very frustrating. I'd advise people to use scholarpedia instead. I know I will.

Duracell 10:44, 21 August 2007 (UTC)

What I don't understand is when you wanted to write an article about non-linear dynamics, why didn't you call it that way?
Allready on 6 March 2004 the user:Waltpohl made a REDIRECT from 'Nonlinear dynamics' to 'Chaos theory'. It's a bit late to complain about that three years later?
Since this article is called complex systems people have been writing about this. I think everybody, even the experts have only half an idea what this is about. So I refrased the title.
The wikipedia article non-linear dynamics is now a redirect. It's seems like a small step to collect you original ideas and make it to an own article non-linear dynamics. Good luck - Mdd 11:13, 21 August 2007 (UTC)
The point is that complex systems are described by predictive models that use the maths of non-linear dynamics. Non-linear dynamics itself is not synonymous with complex systems, but it is crucial to their understanding. And not just in describing the components themselves and their interactions; equations can be derived that explain changes in "emergent" properties without reference to the microstates that cause them. That is what makes the study of complex systems a definable discipline.
There is already a not-bad-looking article on non-linear dynamics (though I just skimmed over it) dynamical system. It is *that* body of work that needs to be refactored into this one, so that there is a proper link between theory and application.
also, this: http://www.scholarpedia.org/article/Complex_Systems is what the article here should look like.
Duracell 19:14, 21 August 2007 (UTC)
I kind of agree with you about non-linear dynamics, but I would only suggest that when it comes to complex systems, there's a good case that can be made that these systems cannot be modelled very well, let alone predicted... (perhaps that's not what you implied) --Childhood's End 19:36, 21 August 2007 (UTC)
I'm glad you "kind of agree". The extent to which they can be modeled is the extent to which they can be *understood*. The rest is hand waving. It's not science. A model is just a formal explanation. And complex systems *can* be predicted, at least to some extent. I'm not sure why you think a prediction is one step up from a model (".. let alone predicted"). Some models cannot be predicted in practice, although they are in principle predictable, because they are deterministic (e.g. "chaos" in dynamical systems of >=3 dimensions). Still that might be clouding the issue; models are usually used for prediction. But we can talk about levels of prediction. Can I predict the weather in Accra next year? Well, I cannot predict where all the individual air or water molecules will be (micro-level). I can't predict where individual weather systems will be (meso-level) but I *can* predict that it will be very hot (macro-level). This macro-level description is a bit trivial. The point in complex systems is that we are trying to make meso-level predictions, but not micro. and those predictions might involve, e.g. phase transitions which are modelled (amazingly well) by bifurcations in non-linear systems.
In respect of all this, I think anyone who wants to chip in on this should read the scholarpedia article (which is *peer* *reviewed*!!!!) I referenced above and look at the bibliography there. Actually, if scholarpedia is GPLed we should probably just copy that article in here (when it is out of review of course)
Duracell 21:00, 21 August 2007 (UTC)
turns out scholarpedia is copyright, so that a print version can be published. But a good place to start nonetheless. Duracell 21:03, 21 August 2007 (UTC)
I made some inquiries here and there... and first some findings:
  • You said here When I originally wrote this..., I guess you refer to the rewritting you did between 13 and 18 september 2005. After this makeover [1] the article kind of got his current form.
  • In the comment you gave 13 september 2005 on this talk page [2], you also stated. "Complexity" is a buzzword at the moment, and unfortunately many are confused over what is entailed.
  • Now the scholarpedia article you mentioned introduces complex systems as a "paradigm" (a set of common beliefs & fundaments shared in the group of scientists), further jumps from complexity research to aspects of complex systems.
  • ... and in your comment here you focuss very closely to those "complex systems (that can be) are described by predictive models that use the maths of non-linear dynamics"
  • Now I myselve have mentioned in this last talk item here, that a better organization is needed between the articles Complexity, Complexity theory, Complexity theory and organizations, Complex adaptive system and this Complex system.
Now in these five findings I see a common underlaying problem, that the buzzword "complex system" is not unraveled. The term "complex systems" has at least three meanings:
  • A specific kind of system
  • A field of science studying these systems
  • A paragdigm, that complex systems have to be studied with non-linear dynamics
The biggest problem (in my opinion) with the current article is, that I speaks of all these three meanings. I think it would be much better, if those three meanings would be explained in three different articles... Or at least in different sections, with a clear overview. - Mdd 21:04, 21 August 2007 (UTC)
These are good points, well met. But I think it should remain a single article and here's why: What is happening in the field right now is that complex systems has turned into an important branch of physics. The paradigm is best exemplified by Haken and Prigogine. Let's say you wanted to do an article on "massive things". Would you argue that it should be split into three articles? 1) things that are massive. 2) the field that studies them (physics) and 3) a paradigm (e.g. newton's, or aristotle's)? That would seem odd. Why? Well, the theories of Newton and Einstein kind of *define* massive things. I'd argue that, at this point, we're getting there in terms of defining complex systems. It's not *only* about non-linear math - but in large part it is the link between that math and the phenomena. The math can describe the behavior of emergent phenomena! that is an amazing thing. And it can derive those dynamics from the micro-level structures below. It's very cool. But it's not just the math - the math is there to describe physical phenomena - specifically, self organization (== emergance). What I'm saying is I'm not sure how you could write a *good* article that talked about these system without bringing in the physical/mathematically based theories of what they are. It would be hand waving - and it's not even necessary. Moreover, I'm not sure that the "field" and the "paradigm" are separate. I think they are the same thing at this point.
Duracell 21:31, 21 August 2007 (UTC)
Thanks for this feed back. I agree with your last remark. But I do believe that the current article has to be split in two... for the same reason there is a separate system and systems theory article. You have said it yourselve: This article has become a mess. To avoid this a more solid ground has to be shapen.
I've been working rather intens in this field lately. And much of the work I do is creating the (article) space in such a way that others can further contribute. I wikify a lot and bring articles up to one standard lay out. But in between I'm trying to thinking a lot about the obvious. I ask myselve the question: What is the first thing people need to know to get a little understanding about the subject? That should be a basic element of any article. The next question is: In what sequence, does those people should be further introduced into this subject matter?. That is one of my mayor concerns.
Now I have experiences more then a dozend times, that splitting the object of study from the study in seperate articles clearifies and creates space. I believe you also want to do something about it. Why not give it a try? - Mdd 21:56, 21 August 2007 (UTC)

[edit] A new design in two seperate articles

I have made a first draft for a separate article about complex system, see User:Mdd/Complex system. This gives a first impression on what I intend to do. Of cause a second article about complex systems theory is missing, but this will contain all the information from the current article which is not in the new complex systems article. - Mdd 23:25, 21 August 2007 (UTC)

I just can't see what an article on complex systems that ommitted all theory would have to offer - it might even stoke the fires of confusion in that it would just be a list of "complicated things", which is not the same as a complex system. Morevoer, the scholarpedia article is an excellent introduction. At the moment I'm busy researching the brain as a complex system, for my PhD in "complex systems", and if I took time to write something as I did in 2005, I reckon it would just end up deleted. Duracell 01:46, 22 August 2007 (UTC)

I have recreated the second article User:Mdd/Complex system (study). And I think I am going to implement this sone. I'm not so sure any more that the scholarpedia article is an excellent introduction... of complex systems. It's probably an excellent introduction of complexity. But since when are complexity and complex systems one and the same thing? I think that if you name an article complex system and then write about complexity... you don't make a contribution to society. It mistifies instead of clearifies things. - Mdd 13:32, 22 August 2007 (UTC)

Both these terms have multiple meanings in the English language. But as far as technical, scientific meanings go, "complexity" and "complex system" are so close in meaning that they are effectively the same thing. All complex systems exhibit complexity. All things that exhibit complexity are complex systems. Duracell 16:14, 22 August 2007 (UTC)

Thanks for this explaination. I have just recreated the complexity article, and I have come to the point that I have three designs:

Now I'm trying to compare these with scholarpedia article!? That article on first hand is far more impressive then the three (what you already called) listings. I have faith however that the three related article can and will also grow.

It remains clear to me that something has to been done with the current situation... If you have no really big objections I like to give it a try to split this article in two. - Mdd 17:30, 22 August 2007 (UTC)

I'll just implement it. Then people can really see what happens and respond anyway. - Mdd 18:55, 22 August 2007 (UTC)

I made one more step. I changed the articles name complex systems (study) into complex systems. - Mdd 19:21, 22 August 2007 (UTC)

[edit] The Concept

The main reason for so many misunderstandings about complex systems and complexity theory are ambiguous and overcomplicated explanations. People seem to think that words “complexity” and “complex” justify complicated explanations in which even “explanators” (coined word) cannot find their way.

As any other entry in any other encyclopaedia, the explanation needs to be simple, easy to follow and with examples to support the explanation. It needs to satisfy readers of all possible profiles: experts and non-experts.

It always helps to start with a bit of history (George Henry Lewes and Jules-Henri Poincaré) and example like emergent property saltines of salt (NaCl). Here we see that the first notions of complexity were based upon observations of natural phenomena. The math (theory) that followed is still having some issues to match itself with observations, mainly because of linear thinking. (See complexity theory bible: Thinking in Complexity by Klaus Mainzer.)

The concept is, however, clear and supported by numerous examples. (Market Equilibrium, planetary systems, molecules, bees, ants, us, culture etc. each describing interactions of agents (oscillators) and result of these interactions called emergent property.) I think that I have managed to explain the concept clearly enough on my web pages: Imagination is Greater than Knowledge. However, I can surmise it here:

  • A limited number of agents (closer to each other than to other agents - horizontal relationships) impact each other with each of them modifying its impact on others (its behaviour) depending on how it was impacted by others.
  • The whole group (system) is initially unstable (while individual agents are modifying their behaviour - see above), but with few possible outcomes that will stabilise the system as a whole. (In case of some carbohydrates we can have two such outcomes or emergent properties: “left” or “right” orientation of the molecule.)
  • Once the system is stabilised (symmetry splitting), it, as agent (oscillator), enters into interplays with other, stable systems on the same or different “level”. And the same story repeats itself.

It is interesting to note that unstable system is in a higher energy level than stable system and we need to introduce energy into stable system if we do not like the outcome of the previous symmetry splitting.

As for this whole entry: It is much better and more competent than previous one. (Scrapped, I guess.) However, when I consider that previous one listed attributes and this one is listing seemingly different meanings of word complexity - I’m not so sure. There is one, and only one, meaning of complexity in all listed (and some non-listed) disciplines and this meaning needs to be articulated before we get into different aspects of it.

Damir Ibrisimovic 08:27, 25 August 2007 (UTC)

You are speaking of "many misunderstandings about complex systems and complexity theory" and those "being ambiguous and overcomplicated explanations". I just created a new structure here with a simplification, with:
a complex system is a specific form of system
and complex systems is a field of study
and complexity is a (new) paradigm in science.
I'm wondering what you think of this new situation? I'll really like some feed back on this subject, because I'm having some discussion about this at Wikitionary as well. - Mdd 16:10, 25 August 2007 (UTC)
The trouble is that "complex systems", "the complexity theory" and "complexity" got their names from people originally working (and thinking) within "ordinary systems" (input -> processing -> output) areas. And using "ordinary systems" methodology made descriptions of natural phenomena ("complex systems") extremely complex. Although a new, more appropriate methodology is far from being articulated, there is growing realisation that natural systems are far from being complex. So we could say that (according to complexity theory):
  • All (natural) systems (or observations of them) are "complex", i.e. there are no "complex systems" as a subset of "systems".
  • A description of a system could be complex or simple depending on methodology used.
It could be said that components of "ordinary", man made systems (input -> processing -> output) are constructed so that a "noise" from neighbouring components does not interfere. In essence, the complexity of a natural (complex) system is eliminated to "keep the system under control i.e. in the framework of theoretical (input -> processing -> output) concept.
Since we are stuck with "complexity", as a concept, we are compelled to use it, but we need to be aware of the fact that "complexity" refers to "methods of description" rather than to nature of natural phenomena. You are right about "complexity" being a (potential) new paradigm in science. But for this potential to be realised we need a new methodology that will SIMPLIFY descriptions of natural phenomena. Just like Kepler's (description of) universe replaced the Ptolemaic one.
I would suggest that such approach will yield much better results in understanding and structuring. Good luck.
The following is my attempt to bridge a physics/chemistry gap in complexity terms:
  • Intraction between an electron and nucleus is stabilised when the electron "drops" to the lowest orbit or energy state. (Let's say: symmetry splitting - level 1.)
  • A molecule formation through interactions between atoms (level 2) seeks a lowest energy level for the whole of the molecule.
  • The impact of level 2 interplays on level 1 interplays results in some electrons jumping from the lowest orbit to a higher energy level to satisfy one of few possible stable states of the molecule as a whole.
This view is quite close to what is known in physics and chemistry. Note that this also explains observed phenomena in a bit different way. It also contributes to a better understanding of why formations of some molecules release energy while other need energy from the environment.
Damir Ibrisimovic 21:46, 25 August 2007 (UTC)

[edit] What is required here?

I've repeatedly stated how I think this article should be done. I greatly improved it in 2005, and would be willing to work on it further. All that is required here is not 3 articles, or impressionistic wanderings, but a simple article after the manner of the article on scholarpedia, as I have said above. - Duracell 00:36, 10 September 2007 (UTC)

I have made this a separate new talk item.
Now I also have explained why I made the changes I made. There are now three article to explain three related subjects:
  1. the concept of a complex system.
  2. the existing of a scientific discipline.
  3. the existing of a phenomenon called complexity.
I still think that most of the things said in the scholarpedia article could fit here (in the wikipedia) in the complexity article. - Mdd 00:51, 10 September 2007 (UTC)
@Duracell. I agree with your comments about chaos theory being over-hyped. I think right now, the biggest problem with this article is that it's a collection of lists and random facts and sections, and does not have a coherent narrative. This is why I created the history section! But I'd like to see the other sections become more unified and read more like an actual article too...not just a collection of random statements. Cazort (talk) 19:30, 17 November 2007 (UTC)

[edit] History Section

I thought there really needed to be a history section on this page. I created one and put a few tidbits there but it's only the beginning. I would be very grateful if anyone could contribute enough material to weave together a coherent narrative. I would be glad to come back and add material on ecology, but I don't know as much about other subjects. Cazort (talk) 19:30, 17 November 2007 (UTC)

I do agree that a historical section can add some coherence to this article. I will give it some thoughts. - Mdd (talk) 20:30, 18 November 2007 (UTC)
There are some remarks in the historical section in the system article which could also fit in here. And you also start looking in the historical sections of other related Wikipedia articles. Now I will leave that you to decide. Good luck. - Mdd (talk) 20:34, 18 November 2007 (UTC)

There used to be relevant info about the origins and early developments of complex systems theory, which was, it seems, deleted during this edit [3]. I would bring back some of this info as the article is pretty "dry" right now. --Childhood's End (talk) 19:17, 20 December 2007 (UTC)

Hi, I guess that section moved towards the complex systems article. -- Mdd (talk) 22:28, 20 December 2007 (UTC)
Oh, so there's two articles now. I see, thanks for the note ! --Childhood's End (talk) 01:34, 21 December 2007 (UTC)
Maybe the first sentence here should be more precise. Maybe the two articles should be merged after all. There are other voices telling, those two articles should be merged with complexity. And an other voice telling ther three articles should be removed and the citizendium article should take their place. These discussions are small in compare to the argumentations you seemed to be in. But if you have other idea's here, please let me know. -- Mdd (talk) 01:41, 21 December 2007 (UTC)
I keep an eye on the complex system(s) articles. Besides, nice work there Mdd, and hats off for all the effort. --Childhood's End (talk) 15:18, 21 December 2007 (UTC)

[edit] Non-linear = dependent on initial conditions?

Regarding the following entry:

Relationships are non-linear In practical terms, this means a small perturbation may cause a large effect (see butterfly effect), a proportional effect, or even no effect at all. In linear systems, effect is always directly proportional to cause. See nonlinearity.

I could be wrong, but I don't think the argument follows. I can envisage a linear system that is ill-conditioned, although I don't know if a non-linear system CAN be well posed. Does Hadamard's definition of a well posed problem apply to non-linear systems too? Or am I completely misguided here? Also I couldn't find any reference to a system being sensitive to initial conditions from the page on nonlinearty that is referenced. Maybe someone with more experience in the field could follow this up?

Maszanchi (talk) 16:55, 21 December 2007 (UTC)


Broadly, non-linear does mean dependent on different initial conditions. in simple (one dimensional) non-linear systems, there are only fixed points in the dynamics, to which trajectories will converge. But there can be more than one fixed point, and which one you end up at is determined by initial conditions. In two dimensions, it gets more complicated, and in 3+ dimensions, you (can) get chaos. In chaotic dynamics, it is impossible to predict where the system will go (as far as is known), even though it is deterministic (i.e. is not random in any way).

A one dimensional linear system could only have one fixed point, at the intersection of the line describing the rate of change, and the line describing the position. Two dimensional linear system can have orbits as well as fixed points, but still just one fixed point. So, in that sense they are not dependent on initial conditions. Strogatz's book is good for all thisDuracell (talk) 23:47, 22 February 2008 (UTC)

Actually, now that I think about it, two dimensional linear systems have orbits, and the amplitude of the orbit depends on the initial conditions. So that is an example of a linear system that depends on inital conditions. but *qualitatively* the dynamics don;t depend in IC. Duracell (talk) 23:50, 22 February 2008 (UTC)