Talk:Nonlinear system

From Wikipedia, the free encyclopedia
Jump to: navigation, search
edit·history·watch·refresh Stock post message.svg To-do list for Nonlinear system:

Nonlinear equation and nonlinear function[edit]

--A nonlinear function is one which does not satisfy the superposition criteria, as this page says. But I'd like to see a warning about a common misconception:

With reference to, a linear equation is: "an equation involving only the sum of constants or products of constants and the first power of a variable." That is, "y + 3x +1 = 0" is a linear equation. However, the equivalent function y(x) = -3x -1 does not satisfy the additive criteria, as one can easily check for oneself. A linear equation correspond to a linear function or operator iff the line passes through the origin.

The nonlinearity pages have been greatly improved lately and I leave it to the main contributors to incorporate this information in the way they see fit. Lauritz

-- Given that this is the very basis of chaos and numerical Relativity, I shall be slowly attempting to clean it up. Help is appreciated. --Son Goku 4 July 2005 23:37 (UTC)

nonlinear or non-linear[edit]

What is the preferred spelling? Judging by searching for "nonlinear non-linear" and examples such as non-Hermitian, I would guess with the dash, but some conformation would be nice. --MarSch 12:51, 12 September 2005 (UTC)

I think specialities which use this term a lot, prefer it without the dash. If you really care about the dash, ask at the WikiMath talk page. (And be ready to disambiguate all the links if you think of doing any move, as well as replacing the instances in the main text and a bunch of instances in linking articles.) Oleg Alexandrov 16:01, 12 September 2005 (UTC)

Hey guys, sorry to be picky, but heck, you're all mathematicians or physicists here, so y'all understand precision: you're using a HYPHEN, not, repeat NOT a dash, when you type "non-linear"! As astonishingly few techies--or even published writers, sometimes!--seem to realize, a hyphen is nearly as different--in function, if not so much in appearance in most fonts--from a dash as, say, an exponent is from an operation sign! [signed] FLORIDA BRYAN — Preceding unsigned comment added by (talk) 17:35, 21 April 2012 (UTC)

Math markup[edit]

Why do some equations render as full-size, and others are inline? I checked the stuff inside the (angle-bracketed) math tags, but there don't seem to be any major differences. I noticed this in the 'Linear systems' susbection - here's some stuff copied straight out of it:

       * F(x) = x2, where x a real number
       * F(u)=-d_x^2 u + g(u), where u is a function u(x) and x is a real number and g is a function;
       * F(u,v) = (u + v,u2), where u, v are functions or numbers.

The second equation rendered properly, the first and third didn't. Anyone (who knows more about the math tags than me) know what's up? I was trying to change the inline equations to full-size, because they're too small to read properly the way they are at the moment.

Bird of paradox 15:46, 12 March 2006 (UTC)

In your settings is an option which controls whether math-tags are rendered as latex or as html. You can force latex rendering by adding a thin space like \' .--MarSch 08:52, 24 April 2006 (UTC)

The "Specific nonlinear equations" section is a little unreadable where the equations break up the explanations. Would there be any objections to me making section more readable? --Kenneth M Burke 23:39, 10 June 2007 (UTC)

It seemed to me that the page wasn't uniformly a textbook way of explaining the math with the markup. I checked some textbooks, it's fine. Having never edited a textbook, I never noticed that they really read in such a way. Kinda awkward honestly, but that is how it is done. --Kenneth M Burke 02:24, 13 June 2007 (UTC)


The notation etc is confusing. What's it supposed to mean? ? Partial ? Something else? What about ? Is this partial or ? Or even a metric d?

This is supposed to be an introductory article and should use standard notation wherever possible.

-- 17:50, 2 June 2006 (UTC)

(PS: I have a PhD in nonlinear dynamics / electronics and I can't understand it!)

Yes this probably should be changed. I would have jumped in and changed it already but I want to make sure what I put is accurate. I think it's saying
du/dx = u^2
and then
du^2/dx^2 + g(u) = 0
If there are no objections I (or somebody) could write this in math markup and update the page....

Richard Giuly 10:33, 10 March 2007 (UTC)

"To do"?[edit]

What's the "to do" bit under Examples about? Does it need tidying? David 21:03, 13 July 2006 (UTC)

"Not equal to the sum of their parts"[edit]

Who says? Has any system ever been proven not to be equal to the sum of its parts? And by whom? Sounds like pseudoscience to me. At any rate please provide a citation for this from a respected peer-reviewed physicist. -- fourdee ᛇᚹᛟ 09:13, 20 August 2007 (UTC)

It just means that solutions to a nonlinear system can't be written as a linear combination (i.e. scaled sum) of solutions from some basis set. I agree that it's a problematic wording, because it's not clear what the "parts" are supposed to be, and usually "sum" in the context means something like a disjoint union. It should probably be changed. -- BenRG 00:43, 21 August 2007 (UTC)

Lead paragraph[edit]

A recent re-write of the lead paragraph of this article says:

"a nonlinear system is one whose behavior can't be expressed in terms of a linear function of the dependent variables"

I think this is wrong in several ways:

  1. "Behavior" is too vague.
  2. The term "dependent variables" should be "independent variables". A closed form solution of a system of equations expresses the dependent variables as functions of the independent variables.
  3. Most importantly, it is based on a misunderstanding of what a linear system is. The solutions of the differential equation are the functions y=kex. These are not linear functions of x, but the differential equation is a linear differential equation.

I think any explanation of non-linearity must mention the superposition principle early in the lead section. This is the defining property of linear systems, and so its absense is the defining property of nonlinear systems. If no-one has any objections, I will correct these errors in the next few days. Gandalf61 09:32, 11 September 2007 (UTC)

I mostly agree. I was worried about that sentence when writing it. Please fix it by all means, but I have a few things to mention:

I think the main problem is vagueness. The sentence doesn't even imply if it's talking about the solution to the problem or the system defining the problem. The solution is by all means generally nonlinear, as you pointed out, and it ought to be mentioned that the equations defining the problem are linear, not the solution.

On that topic, the problem (again, not the solution) must be linear in the dependent variables, not the independent ones. So, for example, Bessel's equation is wholly linear, though it has nonlinear variable coefficients.

I'm not sure if it's ok to mention superposition before the definition section, ...but maybe it is?

The misunderstanding isn't concerning what the system is, the failure is in the extreme vagueness. And FYI, I rewrote a lot more then the opening paragraph. Please fix it if you can. -Ben pcc 16:27, 13 September 2007 (UTC)

wrong solution/wrong equation?![edit]

Re the example nonlinear diff. eq. [du/dx=-u^2], the given general solution [u=(3x+c)^(-1/3)] is wrong, so either the solution or the equation was presumably entered wrong. Note that for u=(3x+c)^(-1/3), du/dx=-(3x+c)^(-4/3)=-u^4, NOT -u^2. The correct solution to the given eq. is u=(x+c)^(-1)...So it seems there are 2 options to correct this: keep the given "solution" as listed but change the given eq. to du/dx=-u^4, or keep the given eq. and change to solution to the one listed here, i.e u=(x+c)^(-1). Can someone please do this. Thanks...

Yes, you are right. I have fixed this by changing the equation (replacing u^2 with u^4) to fit the solution. Gandalf61 11:17, 21 September 2007 (UTC)

I'm sorry about that (I cam up with it). However, it's still wrong, the factor on t is wrong. I'll fix it now. — Ben pcc 16:51, 23 September 2007 (UTC)

Hmm ? Where do you mean by "the factor on t" ? There is no t in the equation. Your edit is fine, but it was also correct before; u=(3x+c)^(-1/3) is the general solution to du/dx = -u^4. You can differentiate u to check. Gandalf61 17:28, 23 September 2007 (UTC)


Why was nonlinearity redirected to nonlinear system? There was a lot of information lost in doing this, nobody said a word or discussed it? A merger should have been discussed before a redirect. Has the information thus been lost? Disappointing. --Kenneth M Burke 21:12, 23 September 2007 (UTC)

Nothing was lost. I revamped the article and moved it, see its history. Old article did contain some interesting info, but it was badly written and it seemed easier for me to start over. A merger didn't happen since Nonlinear system didn't exist; it seems like a better name and is what's used in most other languages here on Wikipedia.
Just before my first edit:
Ben pcc (talk) 21:43, 25 November 2007 (UTC)

Less technically[edit]

The less technically bit, is equally technical, to the technical bit
ThisMunkey (talk) 13:49, 5 March 2008 (UTC)

Sources of nonlinearity[edit]

I would like to see some discussion of common sources of nonlinearity. For example, in solid mechanics, you can have nonlinear material properties but you can also have linear material properties but geometric nonlinearities due to finite strain. I think this sort of distinction would help readers wrap their minds around what nonlinearity is in a practical sense. I'm sure there are other good examples from other disciplines. (talk) 13:33, 11 March 2008 (UTC)

Questionable Relevance[edit]

The second paragraph seems fairly situational and subject. I advocate its removal. Flux (talk) 09:20, 17 August 2008 (UTC)

The superposition principle isn't sufficient for linearity[edit]

I've added "and/or whose output is not proportional to its input" and it was removed as "Redundant subset of superposition principle."

This is not true. For example consider the complex field C as a one-dimensional vector space over itself. Then the function f(z) = z* (complex conjugation) satisfies the superposition principle:

but it's not linear, as that would require f(λz) = λf(z), but

unless λ is real or z is zero. (In technical terms, complex coniugation is antilinear.) Therefore, the second condition doesn't follow from the first, as there is a counter-example in which the first applies and the second doesn't. -- Army1987 – Deeds, not words. 21:11, 6 December 2008 (UTC)

There are two problems with your edit and explanation. First, what does "and/or" mean? If either condition is not sufficient, neither is the or of them. Second, "proportional" is generally taken to mean with a real lambda in your expression. I realize that complex conjugation is not an analytic function, but I did think it was linear by the usual definition. Do you have a source that backs up your definition? Dicklyon (talk) 21:42, 6 December 2008 (UTC)
Any book of vector algebra will do that. λ can be any element of the base field of the vector space over which f is defined. (Of course, complex coniugation is linear if you take C as a 2D vector space over R, but in this example I took it as a 1D vector space over itself.) -- Army1987 – Deeds, not words. 21:53, 6 December 2008 (UTC)
So find such a book and use its definition. Dicklyon (talk) 22:58, 6 December 2008 (UTC)
Modern Quantum Dynamics by J. J. Sakurai says "With the single exception of the time-reversal operator to be considered in Chapter 4, the operators that appear in this book are all linear, that is, " (And the time-reversal operator is additive, only that it is antilinear rather than linear, i.e. ) But I think that using symbols in the lead section (see WP:LEAD), which are supposed to be generally accessible (see WP:MTAA) is not a very good idea, also "use its definition" might border on copyvio (and it is understood that using another definition but which is equivalent is not OR). -- Army1987 – Deeds, not words. 02:52, 7 December 2008 (UTC)
I've seen that "anti-linear" concept in some QM books, but I haven't seen anything like a corresponding definition in any books on linear systems or vector algebra, which you said there are plenty of. Can you find one and point it out? Dicklyon (talk) 08:46, 7 December 2008 (UTC)
I didn't mean that there are plenty of definitions of antilinear on books of linear algebra. I said that books of linear algebra define linear as f(u + v) = f(u) + f(v) and f(λv) = λf(v) (or, more concisely, f(λv + μw) = λf(v) + μf(w)) but none of them requires λ to be real, they just require it to be in the base field of the vector space over which f is defined. Can you back your claim that '"proportional" is generally taken to mean with a real lambda in your expression'? -- Army1987 – Deeds, not words. 10:23, 7 December 2008 (UTC)
I didn't mean a definition for anti-linear, I mean a definition for linear that corresponds to what you're telling us. Just point one out. Dicklyon (talk) 16:58, 7 December 2008 (UTC)
The one which I shown above. -- Army1987 – Deeds, not words. 17:44, 7 December 2008 (UTC)
See Linear map#Definition and first consequences, which I never edited and makes the very same example. -- Army1987 – Deeds, not words. 17:48, 7 December 2008 (UTC)
Yes, it says that, but I'm asking for a reliable source. I'll tag it for one. Dicklyon (talk) 19:54, 7 December 2008 (UTC)

← IIRC, the linear algebra book I used made that same example, I'll add a citation when I find it (it is buried somewhere in my garage right now, I'm going to look for it tomorrow). Unfortunately, it is in Italian, but I think it can be kept as a placeholder until a source in English is found. -- Army1987 – Deeds, not words. 21:35, 7 December 2008 (UTC)

You had said "Any book of vector algebra will do that." So look at some of these and find a definition... Here's one that appears to specifically restrict the multiplicative factor to be real. Dicklyon (talk) 21:58, 7 December 2008 (UTC)
Because that is an example, in which the vectors are in the real plane. This one says that λ must be in F, which is the field over which the vector spaces V and W are. So does this one. So does this one, except that it calls k the scalar and K the field. -- Army1987 – Deeds, not words. 02:15, 8 December 2008 (UTC)
See also this. -- Army1987 – Deeds, not words. 02:18, 8 December 2008 (UTC)
Thanks; I'm satisfied now. Dicklyon (talk) 05:36, 8 December 2008 (UTC)

metaphorical use?[edit]

As an engineer, I have never heard anyone be called "non-linear" be it in the academic or professional spheres. Perhaps this is a phenomenon limited to countries other than the United States or languages other than English. If so, that should be stated under that section. Otherwise, the "Metaphorical Use" section reads more like something from UrbanDictionary and less like an informative article section. (talk) 21:58, 17 December 2008 (UTC)

I added a source. Just don't go nonlinear on us when you see it. Dicklyon (talk) 03:33, 18 December 2008 (UTC)

Algebraic geometry[edit]

would algebraic geometry be considered as the study of non-linear algebraic systems? If so, why not state this in the algebraic equations section? Also, if this interpretation is right, than it should be pointed out that much is still unknown about systems of algebraic equations (for example, it is not known if there is a curve of genus 6 and degree 8 in the three dimensional projective space. Namely, it is unknown if there is a system of two polynomial equations on three variables with degree 8 whose space of solutions has 6 holes - see Hartshorne's book page 345). Perhaps one should mention that their study heavly uses topology, geometry and algebra, among other tools, rendering these disciplines interesting from the viewpoint of equation solving. —Preceding unsigned comment added by (talk) 17:08, 3 March 2009 (UTC)

Some practical sense[edit]

Problem-solving is universally a non-linear thing. The reason for this is that real-world solutions aren't ever linear. Mathematics can be tricky as it is important to realize that solving linear problems by linear solutions and non-linear problems by non-linear solutions do not exist. Hence, a solution is always about something other than itself. And there is no such thing as a type logic by or for itself. This is the essence of mathematical non-linearity. (talk) 18:58, 18 November 2009 (UTC)

More work probably needed[edit]

I think this article has a more general idea that can be served. It is that nothing exists by or for itself. This is why there are non-linearities in the first place. (talk) 19:04, 18 November 2009 (UTC)

One question...(kinda dumb)[edit]

After reading the following bit of the article I was wondering if quadratic functions are non-linear equations.

"...the nonlinearity is due to the x2."

x2 is what makes a function quadratic. My sincere apologies if this is in the wrong place or the question is ridiculous. —Preceding unsigned comment added by Fyrra (talkcontribs) 18:54, 15 May 2010 (UTC)

Non-linear is not linear[edit]

You've got to be kidding...this is a "B" class article, an article that starts off with "a nonlinear system is a system which is not linear"? This write-up explains everything and at the same time explains nothing. In the introductory paragraph numerous, undefined terms are introduced e.g. "superposition principle". He then goes on to use equations to explain a "system", never having defined "system" in the first place. Who is this guy trying to impress? Himself, must be. This article must be scuttled and a new writer brought in who knows how to write understandably. The only person more confused than this writer is the one who rated this article "B". Dangnad (talk) 04:18, 3 September 2011 (UTC)

Second law of thermodynamics[edit]

About the formulation Simple changes in one part of the system produce complex effects throughout.

This formulation describes a phenomenon which violates the Second Law of Thermodynamics. Where did those changes stem from, and who chose to look at exactly those changes rather than other changes? There seems to be a certain popular confounding between a nonlinear system and the mysterious essence called randomness.

Has any research been made about the link to actuality and potentiality?Narssarssuaq (talk) 22:28, 21 July 2012 (UTC)

Definition in lead sentence[edit]

Currently the lead sentence says

... a nonlinear system is a system in which the output is not directly proportional to the input.[1]

That isn't right, as it includes systems of non-homogeneous linear equations. I don't have access to the alleged source of this assertion, but I suspect it was intended there as a necessary condition rather than a definitional necessary and sufficient condition. Can someone come up with something better? Loraof (talk) 20:16, 7 March 2017 (UTC)

... a nonlinear system is a system in which the output is not a linear function of the input.
Note that the link is piped and refers to linear function (calculus). D.Lazard (talk) 20:43, 7 March 2017 (UTC)
Thanks. I'll put that in. Loraof (talk) 21:25, 7 March 2017 (UTC)
To Loraof: Thinking again, I find my suggestion not convenient: It is too technical, and not convenient for physical systems. In fact, using this definition requires not only to know what is a linear function, but also to have modelled the system with a function. Therefore, I would prefer
... a nonlinear system is a system in which the change of the output is not proportional to the change of the input.[1]
The main advantage is that one may verify this directly with physical measures, without constructing any mathematical model. Solitons have been recognised as non-linear a long time before any mathematical modélisation. Another advantage is that one can probably keep the reference. D.Lazard (talk) 22:23, 7 March 2017 (UTC)
OK. Loraof (talk) 23:15, 7 March 2017 (UTC)