大不列颠的海岸线有多长？统计自相似性和分数维数（曼德尔布罗特，《科学》1967，156:636-638）

汉密士20240101

于 2024-05-23 10:43:25 发布

阅读量45

点赞数 1

分类专栏：转载的文章文章标签：数学分形几何曼德尔布罗特海岸线

原文链接：http://li.mit.edu/Stuff/CNSE/Paper/Mandelbrot67Science.pdf

版权

转载的文章专栏收录该内容

24 篇文章 1 订阅

订阅专栏

文章目录

文章标题

B.B. Mandelbrot. How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension. （译：《大不列颠的海岸线有多长？统计自相似性和分数维数》），Science, 1967, 156:636-638.

摘要

Geographical curves are so involved in their detail that their lengths are often infinite or, rather, undefinable. However, many are statistically “self-similar,” meaning that each portion can be considered a reduced-scale image of the whole. In that case, the degree of complication can be described by a quantity D that has many properties of a “dimension,” though it is fractional; that is, it exceeds the value unity associated with the ordinary, rectifiable, curves.

地理曲线的细节非常复杂，以至于它们的长度通常是无限的，或者更确切地说，是不可定义的。然而，许多曲线在统计上表现出“自相似”的特性，这意味着每个部分都可以被认为是整个曲线的缩小版本。在这种情况下，复杂程度可以用一个名为 D 的量来描述，它具有许多“维度”的特性，尽管它是分数值；也就是说，它超过了与普通、可测量曲线相关联的单位值。

正文

Seacoast shapes are examples of high-ly involved curves such that each of their portion can in a statistical sense be considered a reduced-scale image of the whole. This property will be referred to as “statistical self-similarity.” To speak of a length for such figures is usually meaningless. Similarly (1) , “the left bank of the Vistula, when measured with increased precision, would furnish lengths ten, hundred or even thousand times as great as the length read off the school map.” More generally, geographical curves can be considered as superpositions of features of widely scattered characteristic size; as ever finer features are taken account of, the measured total length increases, and thier is usually no clearcut gap between the realm of geography and details with which geography need not be concerned.

海岸形状是高度参与曲线的示例，因此从统计学意义上讲，它们的每个部分都可以被视为整体的缩小比例图像。此属性将称为“统计自相似性”。谈论这些数字的长度通常是没有意义的。同样地（1），“维斯瓦河左岸，如果以更高的精度测量，其长度将是学校地图上读到的长度的十倍、一百倍甚至一千倍。更一般地说，地理曲线可以被认为是具有广泛分散特征大小的特征的叠加;随着越来越精细的特征被考虑在内，测量的总长度也会增加，并且它们在地理领域和地理不需要关注的细节之间通常没有明显的差距。

Quantities other than length are thus needed to discriminate between various degrees of complication for a geographical curve. When a curve is self similar, it is characterized by an exponent of similarity, D, which possesses many properties of a dimension, though it is usually a fraction greater than the dimension 1 commonly attributed to curve. We shall reexamine in this light some empirical observations by Richardson(2). I propose to interpret them as implying, for example, that the dimension of the west coast of Great Britain is D=1.25. Thus, the so for esoteric concept of “random figure of fractional dimension” is shown to have simple and concrete applications and great usefulness.

因此，为了区分一条地理曲线的各种不同复杂程度，就需要除长度之外的其他量。当一条曲线是自相似的时候，它的特征由一个相似性指数 D 来表示，D 具有许多维度的特性，尽管它通常是一个大于通常赋予曲线的维度 1 的分数。我们将从这个角度重新审视理查森（2）的一些经验性观察。我提议将其解释为，例如，大不列颠西海岸的维度是 D=1.25。这样，“具有分数维度的随机图形”这个如此神秘的概念就被证明具有简单而具体的应用以及极大的实用性。

Self-similarity methods are a potent tool in the study of chance phenomena, including geostatistics, as well a economics (3) and physics (4). In fact, many noises have dimension D contained between 0 and 1, so that the scientist ought to consider dimension as a continuous quantity ranging from 0 to infinity.

自相似方法在包括地质统计学以及经济学（3）和物理学（4）等对随机现象的研究中是一种强有力的工具。事实上，许多噪声的维度 D 介于 0 和 1 之间，所以科学家应该将维度视为一个从 0 到无穷大的连续量。

Returing to the claim made in the first paragraph, let us review the methods used when attempting to measure the length of a seacoast. Since a geographer is unconcerned with minute details, he may choose a positive scale G as a lower limit to the length of geographically meaningful features. Then, to evaluate the length of a coast between two of its points A and B, he amy draw the shortest inland curve joining A and B while staying within a distance G of the sea. Alternatively, he may draw the shortest line are of straight segments of length at most G, whose vertices are points of the coast which include A and B. There are many other possible definitions. In practice, of course, one must be content with approximations to shortest paths. We shall suppose that measurements are made by walking a pair od dividers along a map so as to a count the number of equal sides length G of an open polygon whose corners lie on the curve. If G is small enough, it does not matter whether one starts from A or B. Thus, one obtains an estimate of the length to be called L(G).

回到第一段中提出的主张，让我们回顾在试图测量海岸线长度时所使用的方法。由于地理学家不关心微小的细节，他可能会选择一个正的比例尺 G 作为具有地理意义特征的长度下限。然后，为了评估其两点 A 和 B 之间的海岸线长度，他可能会绘制连接 A 和 B 的在距离海 G 以内的最短内陆曲线。或者，他可能会绘制由长度最多为 G 的直线段组成的最短直线弧，其顶点是包括 A 和 B 在内的海岸上的点。还有许多其他可能的定义。当然，在实践中，人们必须满足于对最短路径的近似。我们假设通过沿着地图移动一对分规来进行测量，以便计算一个其角位于曲线上的开放多边形的等边长 G 的数量。如果 G 足够小，从 A 点还是 B 点开始并不重要。这样，就得到了一个将被称为 L(G)的长度估计值。

分形几何
Fig. 1, Richardson’s data concerning measurements of geographical curves by way of polygons whicln have equal sides and have their corners on the curve. For the circle, the total length tends to a limit as the side goes to zero. In all other cases, it increases as the side becomes shorter, the slope of the doubly logarithmic graph being in absolute value equal to D-1. (Reproduced from 2, Fig. 17, by pern~ission.)——(图 1，理查森有关通过其边相等且顶点在曲线上的多边形来测量地理曲线的数据。对于圆来说，总长度随着边趋近于零而趋向于一个极限。在所有其他情况下，当边变短时它会增加，双对数图的斜率的绝对值等于 D-1。（经许可从 2，图 17 复制。）)

Unfortunately, geographers will dis-agree about the value of G, while L(G) depends greatly upon G. Consequently, it is necessary to know L(G) for several values of G. Better still, it would be nice to have an analytic formula link- ing L(G) with G. Such a formula. of an entirely empirical character, was pro- posed by Lewis F. Richardson ( 2 ) but unfortunately it attracted no attention. The formula is $L(G) = M G^{I-a}$ , where M is a positive constant and D is a constant at least equal to unity. This D, a “characteristic of a frontier, may be expected to have some positive cor- relation with one’s immediate visual perception of the irregularity of the frontier. At one extreme, D = 1.00 for a frontier that looks straight on the map. For the other extreme, the west coast of Britain was selected because it looks like one of the most irregular in the world; it was found to give D = 1.25. Three other frontiers which, judging by their appearance on the map were more like the average of the world in irregularity, gave D = 1.15 for the land frontier of Germany in about A.D. 1899; D = 1.14 for the land frontier between Spain and Portu- gal and D = 1.13 for the Australian coast. A coast selected as looking one of the snloothest in the atlas, was that of South Africa and for it, D = 1.02.”

不幸的是，地理学家们会对 G 的价值有不同意见，而 L(G)极大地取决于 G。因此，有必要知道 G 的几个值对应的 L(G)。更好的是，要是有一个将 L(G)与 G 联系起来的解析公式就好了。这样一个完全经验性的公式，是由刘易斯·F·理查森（2）提出的，但不幸的是它没有引起任何关注。这个公式是 $L(G)=MG^{1-a}$ ，其中 M 是一个正的常数，D 是一个至少等于 1 的常数。这个 D，作为一个边界的特征，可以预期与人们对边界不规则性的直接视觉感知有某种正相关。在一个极端情况下，对于在地图上看起来是直的边界，D=1.00。对于另一个极端，选择了英国西海岸，因为它看起来是世界上最不规则的之一；发现它给出的 D=1.25。另外三个边界，根据它们在地图上的外观判断，在不规则性方面更接近世界平均水平，大约在 1899 年德国的陆地边界给出 D=1.15；西班牙和葡萄牙之间的陆地边界 D=1.14，澳大利亚海岸 D=1.13。在地图集中被选为看起来最平滑的海岸之一是南非海岸，对于它，D=1.02。

Richardson’s empirical finding is in marked contrast with the ordinary be- havior of smooth curves, which are endowed with a well-defined length and are said to be “rectifiable.” Thus, to quote Steinhaus ( 1 ) again, “a state- ment nearly adequate to reality would be to call most arcs encountered in nature not rectifiable. This statement is contrary to the belief that not recti- fiable arcs are an invention of mathe- maticians and that natural arcs are rectifiable: it is the opposite that is true.”

理查森的经验发现与光滑曲线的常见行为形成了鲜明对比，光滑曲线具有明确界定的长度，并且被称为是“可求长的”。因此，再次引用施泰因豪斯（1）的话，“一个几乎足以反映现实的陈述将是称自然界中遇到的大多数弧是不可求长的。这个陈述与认为不可求长的弧是数学家的发明而自然的弧是可求长的这种信念相反：事实恰恰相反。”

I interpret Richardson’s relation as contrary to the belief that curves of dimension greater than one are an invention of mathematicians. For that, it is necessary to review an elementary feature of the concept of dimension and to show how it naturally leads to the consideration of fractional dimen- sions.

我将理查森的关系解释为与那种认为维度大于 1 的曲线是数学家的发明的信念相反。为此，有必要回顾一下维度概念的一个基本特征，并展示它如何自然地导致对分数维度的考虑。

分形几何2
Fig. 2. Nonrectifiable self-similar curves to can be obtained as follows. Step 1: Choose any of the above drawings. Step 2: Replace each of its N legs by a curve deduced from the whole drawing through similarity of ratio 114. One is left with a curve made of $N^2$ legs of length $1/4)^2$ . Step 3: Replace each leg by a curve ob-tained from the whole drawing through similarity of ratio $1/4)^2$ . The desired self-similar curve is approached by an infinite sequence of these steps.（图 2. 不可求长的自相似曲线可以如下获得。步骤 1：选择上述任何一个图形。步骤 2：用通过比例为 1/4 的相似性从整个图形推导出来的曲线替换其每一条 N 条边。会留下一条由 $N^2$ 条长度为 $1/4)^2$ 的边组成的曲线。步骤 3：用通过比例为 $1/4)^2$ 的相似性从整个图形获得的曲线替换每一条边。通过这样的无限序列步骤可以接近所期望的自相似曲线。）

To begin, a straight line has dimension one. Hence, for every positive integer N, the segment ( $\leq x \lt X$ ) can be exactly deconlposed into N nonoverlapping segments of the form $[(n - 1) X / N x < n X / N]$ , where n runs from 1 to N. Each of these parts is deducible from the whole by a similarity of ratio $r (N) = 1/ N$ . Similarly, a plane has dinlension two. Hence, for every perfect square N, the rectangle ( $\leq x \lt X$ ; $\leq y \lt Y$ ) can be decomposed exactly into N nonoverlapping rectangles of the form ${\left[ (k-1)X]\right]}/{ \sqrt{N} } \leq x \lt { kX}/{ \sqrt{N}}$ ; ${(h-1)Y}/{\sqrt{N}}\leq y \lt {hY}/{\sqrt{N}}$ , where k and h run from 1 to $sqrt{N}$ . Each of these part is deducible from the whole by a similarity of ratio $r(N)=1/\sqrt{N}$ . More generally, whenever $N^{{1}/{D}}$ is a positive integer, a D-dinlensional rectangular parallelepiped can be deconlposed into N parallelepipeds deducible from the whole by a similarity of ratio $r(N)=1/{N^{{1}/{D}}}$ . Thus, the dimension D is characterized by the relation $D = - l o g N / l o g r (N)$ .

首先，一条直线的维度为 1。因此，对于每个正整数 $N$ ，线段 $\leq x \lt X )$ 可以被精确地分解成 N 个不重叠的线段，形式为 $[(n-1)X/N\leq x\lt nX/N]$ ，其中 n 从 1 到 N。这些部分中的每一个都可以通过相似比 $r (N) = 1/ N$ 从整体中推导出来。同样，一个平面的维度为 2。因此，对于每个完全平方数 N，矩形 ( $\leq x \lt X$ ; $\leq y \lt Y$ ) 可以被精确地分解成 N 个不重叠的矩形，形式为 ${\left[ (k-1)X]\right]}/{ \sqrt{N} } \leq x \lt {kX}/{ \sqrt{N}}$ ； ${(h-1)Y}/{\sqrt{N}}\leq y \lt {hY}/{\sqrt{N}}$ ，其中 $k$ 和 $h$ 从 1 到。这些部分中的每一个都可以通过相似比 r(N)=1/从整体中推导出来。更一般地，每当是一个正整数时，一个 D 维矩形平行六面体可以被分解成 N 个平行六面体，这些平行六面体可以通过相似比 $r(N)=1/{N^\frac{1}{N}}$ 从整体中推导出来。因此，维度 D 的特征在于关系 $r(N)=1/{N^\frac{1}{N}}$ 。

This last property of the quantity D means that it can also be evaluated for more general figures that can be exactly decomposed into N parts such that each of the parts is deducible from the whole by a similarity of ratio r(N), or perhaps by a similarity followed by rotation and even symmetry. If such figures exist, they may be said to have $D = - l o g N / l o g r (N)$ for dimension (5). To show that such figures exist, it suffices to exhibit a few obvious variants of von Koch’s continuous non- differentiable curve. Each of these curves is constructed as a limit. Step 0 is to draw the segment $(0, 1)$ . Step 1 is to draw either of the kinked curves of Fig.2, each made up of N intervals superposable upon the segment $(0, 1/4)$ . Step 2 is to replace each of the N seg- ments used in step 1 by a kinked curve obtained by reducing the curve of step 1 in the ratio $r (N) = 1/4$ . One obtains altogether N2 segments of length $1/16$ . Each repetition of the same process adds further detail; as the number of steps grows to infinity, our kinky curves tend toward continuous limits and it is obvious by inspection that these limits are selfsimilar, since they are exactly decomposable into N parts deducible from the whole by a similarity of ratio $r (N) = 1/4$ followed by translation. Thus, given N, the limit curve can be said to have dimension $D = - l o g N / l o g r (N) = l o g N / l o g 4$ . Since N is greater than 4 in our examples, the corresponding dimensions all exceed unity. Let us now consider length: at step number s, our approximation is made of NS segments of length $G=(1/4)^8$ , so that $L=(N/4)^8=G^{1-D}$ ". Thus, the length of the limit curve is infinite, even though it is a “line.” (Note that it is not excluded for a plane curve to have a dimension equal to 2. An example is Peano’s curve, which fills up a square.)

数量 D 的最后一个性质意味着它也可以针对更一般的图形进行评估，这些图形可以被精确地分解成 N 个部分，使得每个部分都可以通过相似比 r(N)从整体中推导出来，或者可能通过相似性，然后是旋转甚至对称。如果存在这样的图形，则可以说它们的维度为 $D = - l o g N / l o g r (N)$ （5）。为了证明这样的图形存在，只需要展示一些 von Koch 连续不可微曲线的明显变体。这些曲线中的每一条都是作为极限构造的。第 0 步是绘制线段（0，1）。第 1 步是绘制图 2 中的任何一条折线曲线，每条曲线都由 N 个间隔组成，可叠加在线段 $(0, 1/4)$ 上。第 2 步是将第 1 步中使用的 N 个线段中的每一个都替换为通过将第 1 步中的曲线按比例 $r (N) = 1/4$ 缩小而获得的折线曲线。总共得到 N2 个长度为 $1/16$ 的线段。每次重复相同的过程都会添加更多细节；随着步骤数量增加到无穷大，我们的曲折曲线趋向于连续极限，通过观察很明显这些极限是自相似的，因为它们可以精确地分解成 N 个部分，通过相似比 $r (N) = 1/4$ 后跟平移从整体中推导出来。因此，给定 N，极限曲线可以说具有维度 $D = - l o g N / l o g r (N) = l o g N / l o g 4$ 。由于我们的例子中 N 大于 4，相应的维度都超过了 1。现在让我们考虑长度：在第 s 步，我们的近似值由 NS 个长度为 $G=(1/4)^8$ 的线段组成，因此 $L=(N/4)^8=G^{1-D}$ 。因此，极限曲线的长度是无限的，即使它是一条“线”。（请注意，平面曲线的维度等于 2 并不排除。一个例子是 Peano 的曲线，它填满了一个正方形。）

Practical application of this notion of dimension requires further consid- eration, because self-similar figures are seldom encountered in nature (crystals are one exception). However, a sta- tistical form of self-similarity is often encountered, and the concept of di- mension may be further generalized. To say that a (closed) plane figure is chosen at random implies several definitions. First, one must select a family of possible figures, usually designated by $\Omega$ . When this family contains a finite number of members, the rule of ran- dom choice is specified by attributing to each possible figure a well-defined probability of being chosen. However, $\Omega$ is in general infinite and each figure has a zero probability of being cho- sen. But positive probabilities can be attached t o appropriately defined “events” (such as the event that the chosen figure differs little-in some specified sense-from some specified figure).

维度这一概念的实际应用需要进一步考虑，因为自相似图形在自然界中很少遇到（晶体是一个例外）。然而，经常会遇到一种统计形式的自相似性，并且维度的概念可以进一步推广。说一个（封闭的）平面图形是随机选择的，需要进一步的定义。首先，必须选择一个可能的图形家族，通常用表示。当这个家族包含有限个成员时，通过给每个可能的图形指定一个明确的被选择概率来指定随机选择的规则。然而，通常是无限的，每个图形被选择的概率为零。但是可以为适当定义的“事件”（例如所选图形在某种指定意义上与某个指定图形相差不大的事件）分配正概率。

For the family $\Omega$ , together with the definition of events and their proba- bilities, to be self-similar, two condi- tions are needed. First, each of the possible figures must be constructible by somehow stringing together N fig- ures, each of which is deduced from a possible figure by a similarity of ratio r; second, the probabilities must be so specified that the same value is ob- tained whether one selects the overall figure at one swoop or as a string. (The value of N may either be arbitrary, or chosen from some specific sequence, such as the perfect squares relative to nonrandom rectangles, or the integral power of 4, 5, 6, or 7 encountered in the curve built as in fig.2.) In case that the value of r is specified be choosing N, one can consider $- l o g N / l o g r$ a similarity dimension. More usually, however, given r, N will take different values for different figures of $\Omega$ . As one considers points “sufficiently far” from each other, the details on a “sufficiently fine” scale may become asymptotically independent, in such a way that $- l o g N / l o g r$ almost surely tends to some limit as r tends to zero. In that case, this limit may be considered a similarity dimension. Under wide conditions, the length of approximating polygons will asymptotically behave like $L(G)~G^{1-n}$ .

对于家族，以及事件的定义和它们的概率，要使其自相似，需要两个条件。首先，每个可能的图形都必须通过某种方式将 N 个图形串接在一起而构造出来，每个图形都是通过相似比为 r 的相似性从一个可能的图形推导出来的；其次，概率必须被指定为使得无论一次选择整个图形还是作为一个字符串选择，都能得到相同的值。（N 的值可以是任意的，也可以从某个特定的序列中选择，例如相对于非随机矩形的完全平方数，或者在图 2 中构建的曲线中遇到的 4、5、6 或 7 的整数次幂。）如果通过选择 N 来指定 r 的值，则可以将视为相似维度。然而，更通常的情况是，给定 r，N 对于中的不同图形将取不同的值。当考虑“足够远”的点时，“足够精细”的尺度上的细节可能会渐近地独立，使得几乎肯定会随着 r 趋于零而趋于某个极限。在这种情况下，可以将这个极限视为相似维度。在广泛的条件下，逼近多边形的长度将渐近地表现为。

To specify the mathematical conditions for the existence of a similarity dimension is not fully solved problem. In fact, even the idea that a geographical curve is random raises a number of conceptual problems family in other applications of randomness. Therefore, to return to Richardson’s empirical law, the most that can be said with perfect safety is that it is compatible with the idea that geographic curve are random self-similar figures of fractional dimension D. Empirical scientists having to be content with less than perfect interpretation stated at the beginning of this report.

指定相似维度存在的数学条件并不是一个完全解决的问题。事实上，即使地理曲线是随机的这一想法，在随机性的其他应用中也会引发一些概念性问题。因此，回到理查森的经验定律，可以说最安全的是，它与地理曲线是具有分数维度 D 的随机自相似图形的想法是一致的。在本报告开头不得不满足于不完美解释的经验科学家。

BENOIT MANDELBROT

参看文献与注释

1.H.Steinhaus, Colloquium Math. 3.1(1954), where earlier reference are listed.
2.L.F.Richardson, in General Systems Yearbook 6, 139(1961).
3.B.Mandelbrot, J. Business 36, 394(1963), or in The Random Character of Stock Market Prices, P.H.Cootner, ED. (M.I.T Press, Cambridge, Mass, 1964), p.297.
4.B.Mandelbrot, IEEE Inst. Elect. ELectron. Eng. Trans. Commum. Technol. 13, 71(1965) and IEEE Inst. Elect. Electron. Eng. Trans. Inform. Theory 13(1967). Very similar considerations apply in turbulence, where the characteristic sizes of the “feature” (that is, the eddies) are also very widely scattered, as was first pointed out by Richardson himself in the 1920’s.
5. The concept of “dimension” is elusive and very complex and is far from exhausted by the simple considerations of the kind used in this paper. Different definitions frequently yield different results, and the field abounds in paradoxes. However, the Hausdorff-Besicovitch dimension and the capacity dimension, when computed for random self-similar figures, have so far yielded the same value as the similarity dimension.
14. November 1966:27 March 1967.

1.H.施泰因豪斯，《数学座谈会》3.1（1954 年），其中列出了更早的参考文献。
2.L.F.理查森，在《一般系统年鉴》6，139（1961 年）。
3.B.曼德尔布罗特，《商业杂志》36，394（1963 年），或在《股票市场价格的随机特征》，P.H.库特纳编（麻省理工学院出版社，马萨诸塞州剑桥，1964 年），第 297 页。
4.B.曼德尔布罗特，《电气电子工程师学会通信技术学报》13，71（1965 年）和《电气电子工程师学会信息理论学报》13（1967 年）。在湍流中也有非常相似的考虑，其中“特征”（即漩涡）的特征尺寸也非常分散，这最早是由理查森本人在 20 世纪 20 年代指出的。
5.“维度”的概念是难以捉摸且非常复杂的，远非本文中使用的这类简单考虑所能涵盖。不同的定义经常得出不同的结果，该领域充满了悖论。然而，对于随机自相似图形计算时，豪斯多夫-贝西科维奇维度和容量维度迄今为止得出了与相似维度相同的值。
14.1966 年 11 月：1967 年 3 月 27 日。

P.S.2024年5月23日上午10点36分，全篇文字校准和数学公式latex语法转换初步完成，该篇文章全文机器翻译，博主本人仅仅是搬运和数学公式、文字校准。分形几何（自相似），这门数学科目我大概一生都要仰望。