風雲變幻誠多端──大自然之混沌與分形

風雲變幻誠多端──大自然之混沌與分形

Quote:
問題不大在於上帝「擲不擲」骰子,而是上帝「怎樣」擲骰子。
- Ian Stewart [1]

於一九八六年你不會在生理學書籍中找到分形一詞,我想於一九九六年你不會找到一本生理學書籍並不提及它。
- Ary L. Goldberger,哈佛醫學院[2]

  馮夢龍寫《喻世明言》有言:「榮枯貴賤如轉丸,風雲變幻誠多端[3]。」影響天氣的變化的因素很多,自古難測。三國時期諸葛亮借東風讓人歎為觀止,假如真有其事。近代西方自然科學取得成功,可是天氣預測的精確程度暫不如其他科學理論。試想,每天在各媒體發佈的天氣預告,可靠程度有多高?為何只有未來一週的預測?明年二月十四日澳洲聖誕島飛魚灣的天氣如何?能否預測?順便想想,足球球季尚未展開,博彩公司已為下屆賽事提供賠率,以英超為例,為何博彩公司視曼阿利車四強為熱門?為何沒有五十年後的賽事賠率?五十年後的賽事不能預測嗎?這些問題關乎一個近代科學的概念:混沌。筆者在本文試以簡單的文字,管窺此新知之一二,並列數書以供有興趣的讀者們錐指探察。

  近代科學有多可靠?艾耶爾[4]主張,只允許邏輯和純數學的命題為確鑿無疑的。我們不能憑體驗駁倒這類命題,因為它們非關經驗世界。而有關經驗世界事物的命題,只是假說,很可能成為事實但永不確鑿無疑。邏輯學分析語句,有助解釋邏輯和純數學的命題,跟有關經驗世界的語句,兩者之分別[5]。氣象學家羅倫茲亦認為,多數真實現象的理論研究,皆為近似值的研究[6]。費曼甚至說,任何物理理論皆是一種猜測[7]。科學理論是關乎經驗世界,不會像邏輯和純數學那樣百分百可靠。然而純數學建基於邏輯,自然科學建基於數學,科學理論的可靠程度是僅次於邏輯和純數學。而且,自然科學放諸四海皆準。只要實驗條件一致,你在北京做實驗所獲的結果,跟在東京做的,在可接受實驗誤差範圍內。偽科學、算命、風水命理或迷信的觀念,並非建基於理性思維,不可與科學相比。大體而言──我重申是大體並非個別例子──非科學的推斷不夠精確、預測力弱和推測不佳。

  我們看到自然科學跟其他學科或民間觀念相比,是相當可靠的。然自然科學各門學科之間呢?這可有些微差別。例如你打開一本大學程度物理書,當中的理論都相當可靠。而一些有關生命起源、宇宙濫觴的理論,由於事隔太久,很多數據已失佚,人類可能仍未找到滿意的解釋。天氣預測亦是一門很難操縱的學問,一來涉及的數據繁多,二來能夠處理大量數據的電腦是近代產物。現時的天氣預測已較遠古時期的有大改進。我亦強調,縱使天氣預測在各門科學之間不及某些很精準的理論,但跟其他非科學觀念比較,還是很可靠的。

  走筆及此,前戲也弄得差不多了,我們轉入正題。何為混沌?容許我先舉例子,稍後定義。較易為一般人理解的比喻是「蝴蝶效應」:在巴西,一隻蝴蝶輕拍翅膀,導致美國德州產生一個龍捲風。又,以電影為例,奇斯洛夫斯基的《盲打誤撞》[8]和提克威爾的《疾走羅拉》[9]都用了一個非主流的敘事方式。影片可分為幾個故事,每段故事的開局是一樣的,因為一些很細微的因素,引致結局不一。例如趕上一班列車和趕不上,就有不同的後果。羅倫茲[10]則以彈珠台為喻。他憶述大學時期,學生熱中於這種遊戲。有些學生偶爾獲勝,多數人輸掉。他發覺即使是有練習過的學生也不會成為長勝軍。羅倫茲解釋彈珠台機器對初始速度很敏感。留意,這些比喻較易理解,但不是混沌的完美例子。羅倫茲對混沌現象的解說為,看似隨機和不可預料的行為,然而是根據精確的通常易於表述的法則進行[11]。他又重申,不論必要的初始條件的差別如何細,附近的狀態最終會擴散,這是混沌行為的一個必要特性[12]。這可能不易理解[13],正如他說,精確的定義並不經常是方便的一種[14]。


「蝴蝶」──羅倫茲吸引子[15]

  清朝大學士紀昀﹝字曉嵐﹞曾以「螳螂捕蟬,黃雀在後[16]」喻眼光短淺。有沒有荒誕地想過,那黃雀之後又有蟬,蟬之後又有螳螂,螳螂之後又有黃雀,如此類推,不斷循環?這可能是分形的其中一個精髓,儘管不夠準確。曼德勃羅[17]介紹分形時斷言,多數大自然的形態都是無規律和無條理的,遠比歐幾里得幾何的圖形龐雜。歐氏幾何不足以描繪現實世界。曼德勃羅[18]曾研究英國的海岸線有多長。他著眼於海岸線,我在這兒談談面積,希望有助理解。簡單的圖形如圓形、方形和三角形,我們都有簡單的方程計算它們的面積。但英國島的面積呢?英國島不是簡單的圖形,我們沒有簡單的方程計算它們的面積。曼德勃羅談及海岸線是自相似曲線﹝self-similar curves﹞,這些曲線後來被稱為分形。這自相似是什麼東西?閒言「兩條平行線終會有交彙的一天」。試想,我們建築由兩條平行線組成的火車軌道,圍繞地球一周。走到軌道中間,平行線看來有交彙的一點。但當我們向這點走近,不斷不斷的走,不斷循環也抓不住這交彙點,平行線並不交彙。我們不斷走近這點的影像,與 Koch 曲線的自相似近似[19]。再幻想你有一輛甲蟲車,望著它,越望越仔細,甚至用顯微鏡放大,車的表面看似越來越滑,但當你越放越大,它的表面越崎嶇不平,哥倫比亞大學的 Christopher Scholz 發現分形幾何是描述地球表面的強大工具[20]。


曼德勃羅集[21]

  前兩般分別勾畫了混沌和分形的概念,接著談談一道科學門外漢很喜歡問的問題,這些東西有什麼用?假如我說,混沌的應用很廣,包括數學、物理學、生物學、電腦科學、工程學、心理學、經濟學、金融學等等,他們也不會真正了解混沌有什麼用。假如我深入點說,在生物學中生態學的動力系統中,他們大概已經頭昏腦脹。所以我不打算在這裡解答。現代科學發展到今天,分工極細。前沿的理論不易為普通人理解,因為學習科學是要循序漸進,一步一步慢慢積聚知識。在現代社會,要精通每一個學科近乎不可能,但對眾多學科有一些基本的認知不是不可能。介紹混沌,這篇短文所能做到的十分有限。有興趣的讀者,可先從 James Gleick 的著作入手,這本書按混沌的歷史發展,由蝴蝶效應到自然之幾何到生物模型到將來展望,對各種概念和應用都著墨不少。Ian Stewart 那本也可視為通俗入門。羅倫茲和曼德勃羅的著作比較專門,適合有科學底子的讀者。Anton 和 Rorres 的《基礎線性代數應用版》[22]則是以最簡單的數學作較專業的入門。Peitgen、Jurgens 和 Saupe 的《混沌與分形──科學之新邊疆》[23]圖例豐富多彩,屬專門著述,是研究生的尚佳良伴。這些書多列有頗詳盡的書目以供深入研習。

[1] 「上帝擲骰子」這話該理解為「世界沒規律」,不應鑽牛角尖爭辯這話預設上帝存在而沒有實証。原文為:The question is not so much whether God plays dice, but how God plays dice. 見:Ian Stewart, Does God Play Dice? The New Mathematics of Chaos. (2nd ed.) London: Penguin Books, 1997, p.xii.

[2] 轉引自 James Gleick, Chaos: Making A New Science. London: Vintage, 1998, p.282.

[3] 馮夢龍:《喻世明言.卷十八.楊八老越國奇逢》。

[4] A.J. Ayer, Languages, Truth and Logic. London: Penguin Books, 2001, p.9.

[5] 簡言之,語句有分析和綜合。分析語句的真偽只取決於表達式的用法,是先驗、必然和無經驗內容的;綜合語句的真偽除了取決於表達式的用法,還要檢查現實世界,是後驗、蓋然和有經驗內容的。科學理論通常是綜合語句。請參看張海澎:《分析邏輯──理性思維的基石》﹝香港:青年書屋,2004 年 6 月初版﹞,頁 75-79。

[6] Edward Lorenz, The Essence of Chaos. Seattle: The University of Washington Press, 1993, p.5.

[7] Richard P. Feynman, Robert B. Leighton and Matthew Sands, The Feynman Lectures on Physics. (Definitive ed.) Reading, Massachusetts: Addison-Wesley Publishing Company, 2006 (7th printing, 2007), Volume 1, p.6-1.

[8] 奇斯洛夫斯基﹝Krzysztof Kieslowski﹞:電影《盲打誤撞》﹝Przypadek,1987 年﹞。

[9] 提克威爾﹝Tom Tykwer﹞:電影《疾走羅拉》﹝Lola Rennt,1998 年﹞。

[10] 同注 6,p.9-11.

[11] 原文為:seemingly random and unpredictable behavior that nevertheless proceeds according to precise and often easily expressed rules. 同注 6,p.ix.

[12] 原文為:an essential property of chaotic behavior is that nearby states will eventually diverge no matter how small the initial differences may be. 同注 6,p.32.

[13] 考慮到讀者的程度不一,我不打算作太專門的解釋﹝即涉及隨機、偽隨機、決定性混沌等等,其實 James A. Yorke 甚至說過:「If you could write down the solution to a differential equation, then necessarily it's not chaotic」轉引自同注 2,p.76-68.﹞。這只是一篇可能還算不上普及科學的文章,旨在介紹混沌與分形這些上世紀後半期才發展的科學,並喚起讀者興趣而作伸延閱讀。寫科普讀物最難之處是不用專門名詞而讓門外漢了解科學又不乏味。

[14] 原文為:Precise definitions are not always convenient ones. 同注 6,p.16.

[15] 以 MATLAB R2006a 繪。

[16] 紀昀:《閱微草堂筆記.卷四.槐西雜志十四》。語本見《莊子.山木》:「睹一蟬,方得美蔭而忘其身,螳螂執翳而搏之,見得而忘其形;異鵲從而利之,見利而忘其真。」

[17] Benoit B. Mandelbrot, The Fractal Geometry of Nature. New York: W.H. Freeman and Company, 2000, p.1.

[18] 同注 17,第五章。或見:Benoit B. Mandelbrot, How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension. Science. 156, 1967, p.636-638.

[19] 參看:http://en.wikipedia.org/wiki/Image:Kochsim.gif

[20] 同注 2,p.105-106.

[21] 以 MATLAB R2006a 繪,程式由 Alberto Strumia 編寫。參看:http://www.ciram.unibo.it/~strumia/

[22] Howard Anton and Chris Rorres, Elementary Linear Algebra Application Version. (8th ed.) John Wiley & Sons, Inc., 2000, ch.11.14-11.15.

[23] Heinz-Otto Peitgen , Hartmut Jurgens and Dietmar Saupe, Chaos and Fractals: New Frontiers of Science. (2nd ed.) New York: Springer-Verlag, 2004.


回應瀏覽選項

選擇你喜歡的顯示回應的模式,並點選「儲存設定」,以啟用你所做的改變。
._.

恭候 wslee 或其他隱名或非隱名高人指教。

不高人 windy

//問題不大在於上帝「擲不擲」骰子,而是上帝「怎樣」擲骰子。//

如果不是詩不當預設 ^^

ps. 謝 nick 生的藝術品.

混沌?

 //走筆及此,前戲也弄得差不多了,我們轉入正題。何為混沌?容許我先舉例子,稍後定義。較易為一般人理解的比喻是「蝴蝶效應」//

好似係 wslee 先生 phD 研究項目 bor ~

.^^

Broad and interesting collection. I remember that Nick mentioned that he is very interested in Fractal Geometry. Perhaps, it would be nice for him to share with us further the many beautiful pictures so related. Further, a question which you guys might be interested in: while the Butterfly Effect illustrates the occurence of chaos in a vivid way, but in reality you know that things are not that bad, so what's gone "wrong" in the chaos picture?

------------------------------------------------------------------------------------

<<好似係 wslee 先生 phD 研究項目 bor ~>>

Ha.. related, but not exactly.

._.

Yup, I mentioned it in your Chaos post. Sadly I didn't master to plot complex fractal pictures yet. More beautiful pictures can be seen from http://images.google.com/images?q=fractal .

I guess in real life application, we keep improving our initial values and re-input so it's kind of recurrence relation. While in the other situation we only input the initial values once.

More beautiful pictures can be seen from http://images.google.co

very beautiful ^^

believe it or not

When I was heading Sydney yesterday, I found two movies on the flight, one titled Chaos Theory and the other Butterfly Effect. As I have a habit of goolging things out, I was not dreaming, here we go:

http://www.imdb.com/title/tt0460745/ Chaos Theory (2007)
http://www.imdb.com/title/tt0289879/ The Butterfly Effect (2004)

And it turned out I watched neither of them... due to the fact that it was a night flight...

Anyway... a few video about this topic:







精彩

讓我慢慢消化消化。

請問 D 圖係咪用 CHAOS THEORY GEN 出黎?

請問 D 圖係咪用 CHAOS THEORY GEN 出黎?

._.

遊俠@思想者 寫到 :
請問 D 圖係咪用 CHAOS THEORY GEN 出黎?

可以這樣說。

For the Mandelbrot set i plotted, the equation behind is indeed quite simple: f(z)=z^2+c; while the code to program, which was written by Alberto Strumia, is slightly more involved:

Quote:

col=20;
m=400;
cx=-.6;
cy=0;
l=1.5;
x=linspace(cx-l,cx+l,m);
y=linspace(cy-l,cy+l,m);
[X,Y]=meshgrid(x,y);
Z=zeros(m);
C=X+i*Y;
for k=1:col;
Z=Z.^2+C;
W=exp(-abs(Z));
end
colormap copper(256);
pcolor(W);
shading flat;
axis('square','equal','off');

._.

張海澎 寫到 :
讓我慢慢消化消化。

會不會寫得太深?

蝴蝶效應與餬碟效應

呀李天命話:

學混需要餬口,
餬口需要餬碟效應。

http://rthk27.rthk.org.hk/php/leetm2/message.php?forum_id=2&topic_id=4&p...

^^

._.

有人以為「蝴蝶效應的論說沒有什麼出奇、特別之處」。其實蝴蝶效應只是 metaphor,一種令不懂科學的人都看懂的表述例子。當中的理論並不只是「闡明了細微的偏差經過連鎖的效應可能會導致重大的改變」。「這道理很可能連一位幼稚園學生也會明白」,科普的目的是要以簡單的文字解釋滿紙方程那些理論,令普通人讀明白。區區幾句話便自以為是,實乃自暴其短。

.^^

Yes. Very often, when we speak about a scienfitic concept casually, we just want to pinpoint what its essential features are only. So, it naturally follows that there are exceptions. I guess it would be fun to mention two common wrong ideas about chaos which I once committed as follows:

1. Exponential divergence, or in short sensitive dependence to initial conditions, means Chaos.

2. A necessary condition for a system to exhibit Chaos is that it is nonlinear.

Can you think of two trivial examples to disprove the above?

._.

1. Sensitivity to initial conditions is not the only factor for being Chaotic, isn't it? In theory, if something is sensitive dependent to initial conditions, but not aperiodic long term behavior in deterministic system, then it's not chaotic. Example... err... an equilibrium but unstable system, a ball standing on the of sharp object?

2. It's not necessary, but sufficient, is it? Chaotic behavior has to be nonlinear, but nonlinearity may not imply chaos. In fact, most nonlinear systems aren't chaotic, are they?

.^^

Nick 寫到 :
1. Sensitivity to initial conditions is not the only factor for being Chaotic, isn't it? In theory, if something is sensitive dependent to initial conditions, but not aperiodic long term behavior in deterministic system, then it's not chaotic. Example... err... an equilibrium but unstable system, a ball standing on the of sharp object?

Yea, more or less. The ball falls and stops, ha.. so nothing interesting happen afterwards. One more example, the simple differential equation

dx/dt = a x

for a>0 and x being a vector in R^n exhibits sensitive dependence too. But it is obviously not too interesting since everything just escapes to infinity. For chaotic behavior to show up, you need something like squeezing and elongating the phase trajectories, etc. And this is why Smale's horseshoe is regarded as one of the earliest insights into this sort of problems.

So far, the only necessary condition, that I am aware of, in order for Chaos to exhibit is that the dimension of the system N need to be three or more. This is a topological constraint and is a "corollary" from even earlier studies due to Poincare at the turn of the 20th century.

Nick 寫到 :
2. It's not necessary, but sufficient, is it? Chaotic behavior has to be nonlinear, but nonlinearity may not imply chaos. In fact, most nonlinear systems aren't chaotic, are they?

Nonlinearity is neither sufficient nor necessary sometimes. Chaos can also be due to something quite different. A simple example is to think of a billard ball inside a irregular shaped enclosure say, and then consider how two initially close-by trajectories get separated as time progresses. This is a kind of chaotic behavior due to boundary condition. This consideration is a bit different from what I mentioned by chaos in the "N<3 implies no chaos" thread; and it relates to deeper problems on how chaos are to be related between classical and quantum systems. But at least, one should be able to appreciate chaos in the sense of sensitive dependence. With further imagination, one can even taste a further sense of chaos in this case being that the billard is going to visit every point inside the enclosure.

Addendum: In the wave equation context, the equation in use is actually linear for this class of chaos.

nick 兄及 wslee 兄好野!

//科普的目的是要以簡單的文字解釋滿紙方程那些理論,令普通人讀明白。//

nick 兄及 wslee 兄好野! 令我這個科盲數盲也看的津津有味

拜謝. OZ

^____^

._.

Very detailed explanation, thanks wslee.

.^^

You guys are very much welcome.^^

發表新回應

此欄位內容將保密,不會公開顯示出來。
驗証碼
這個問題驗證您是否人類訪客,以防止這個網站被自動化程式貼入大量的垃圾資訊
圖片的 CAPTCHA