1.

Introduction

1.1. Definition
Fractals are any of various extremely irregular curves or shapes for which any suitably
chosen part is similar in shape to a given larger or smaller part when magnified or
reduced to the same size [1]. This is also called self-similarity. Computer-generated
fractals are produced mathematically, and these can create detailed pictures of
mountains, plants, waves, and planets. Some mathematicians, such as Benoit
Mandelbrot, Niels Fabian Helge von Koch and Gaston Julia, study nature first, looking
for fractal behavior.

1.2. History
The word fractal originated in French around the year 1970 from Latin fract meaning
broken. Corresponding to Latin verb frangere means to break or to create irregular
fragments. In wider term fractal is a geometrical shape or pattern made up of identical
parts, which are in turn identical to the overall pattern.
The mathematics behind fractals began to take shape in the 17th century when
mathematician and philosopher Leibniz considered recursive self-similarity. It took until
1872 before a function appeared whose graph would today be considered fractal, when
Karl Weierstrass gave an example of a function with the non-intuitive property of being
everywhere continuous but nowhere differentiable.
In 1904, Helge von Koch, dissatisfied with Weierstrass's very abstract and analytic
definition, gave a more geometric definition of a similar function, which is now called
the Koch snowflake. In his 1904 paper entitled Sur une courbe continue sans tangente,
obtenue par une construction gomtrique lmentaire he used a mathematical
techniques to show that it is possible to have figures that are continuous everywhere but
differentiable nowhere [2]. The equilatorial triangle can be built up into a much larger
but still similar pattern. Any part of the snowflake is equally crinkly, whatever scale it is
viewed at.
In 1918, Bertrand Russell had recognized a "supreme beauty" within the mathematics of
fractals that was then emerging. The idea of self-similar curves was taken further by Paul
Pierre Levy, who, in his 1938 paper Plane or Space Curves and Surfaces Consisting of
Parts Similar to the Whole described a new fractal curve, the Levy C curve. Georg
Cantor also gave examples of subsets of the real line with unusual properties - these
Cantor sets are also now recognized as fractals.
Iterated functions in the complex plane were investigated in the late 19th and early 20th
centuries by Henri Poincare, Felix Klein, Pierre Fatou and Gaston Julia. However,
without the aid of modern computer graphics, they lacked the means to visualize the
beauty of many of the objects that they had discovered.
Some of the most remarkable fractals are the Julia sets, devised by the French
mathematician Gaston Julia (18931978). He was a French fractal mathematician who
devised the formula for the Julia Set, fractal shapes defined on the complex number
plane. Although Gaston Maurice Julia was famous in the 1920s, his work was essentially
forgotten until Benot B. Mandelbrot brought it back to prominence in the 1970s through
his fundamental computer experiments in what soon became to be known as fractal
geometry. Through computer technologies Benot B. Mandelbrot was able to visually
calculate it onto a screen and produced what is now known as Julia sets.
Benoit Mandelbrot was a Polish-born French mathematician who coined the term fractal:
a mathematical model for natural shapes that are irregular and do not form perfect
squares, pyramids, spheres, and so on. He invented a basis for the theoretical
investigation of complex dynamics [3].
As one of the worlds most influential mathematicians, Mandelbrot is recognized as the
mathematician principally responsible for originating fractal geometry and applying it to
science and engineering research. Examples of fractals range from natural objects with
self-similar patterns, such as clouds and plants, to unique computer-generated graphical
art forms created using mathematical formulas. The best known of these is called the
Mandelbrot Set. Researchers use fractals to model and measure irregular patterns and
structures, such as the rough coastline that cannot be represented by classical geometry.
Benoit Mandelbrot was largely responsible for the present interest in Fractal Geometry.
He showed how Fractals can occur in many different places in both Mathematics and
elsewhere in Nature.
1.3. Scope of Application
Fractals are not just complex shapes and pretty pictures generated by computers.
Anything that appears random and irregular can be a fractal. Fractals permeate our lives,
appearing in places as tiny as the membrane of a cell and as majestic as the solar system.
Fractals are the unique, irregular patterns left behind by the unpredictable movements of
the chaotic world at work. In theory, one can argue that everything existent on this world
is a fractal, the branching of tracheal tubes, the leaves in trees, the veins in a hand, water
swirling and twisting out of a tap, a puffy cumulus cloud, tiny oxygen molecule, or the
DNA molecule and even the stock market. The applications of fractals are wide
including in turbulence in fluids, complex number bases in computer science, polymer
chemistry, earthquake prediction, weather prediction, video compression, capacitor
design, filters, surfaces for heat transfer, crystallization, pavement distress, computer
graphics, nerve regeneration/ tissue cells, oil recovery, image generation, soot
agglomerates, bacterial growth, economic markets, satellite images, grammar school
mathematics and many more.






2. Mathematical Elements

A fractal is a geometric figure or natural object that combines the following characteristics:

a) Its parts have the same form or structure as the whole, except that they are at a different
scale and may be slightly deformed;
b) Its form is extremely irregular or fragmented, and remains so, whatever the scale of
examination;
c) It contains "distinct elements" whose are very varied and cover a large range;
d) Formation by iteration;
e) Fractional dimension.
One of example of fractals is the Sierpinski Triangle.

2.1. Sierpinski Triangle
The Sierpinski's Triangle is named after the Polish mathematician Waclaw Sierpinski
who described some of its interesting properties in 1916. It is one of the simplest fractal
shapes in existence. It can be generated by infinitely repeating a procedure of connecting
the midpoints of each side of the triangle to form four separate triangles, and cuting out
the triangle in the center [4].

2.1.1. Construction of The Sierpinski Triangle
1. Start with the equilateral triangle.

2. Connect the midpoints of each side of the triangle to form four separate
triangles.


3. Cut out the triangle in the center.



4. Repeat the steps 1, 2 and 3 on the three black triangle left behind. The center
triangle of each black triangle at the corner was cut out as well.




5. Further repetition with adequate screen resolution will give the following
pattern.





2.1.2. Area of The Sierpinski Triangle
As we know, at each level, one quarter of the triangle is removed. That is, three
quaters of the area of the original triangle is left after the first iteration. Thus, it is
not hard to infer that after n iterations, the area of the Sierpinski's Triangle would
be (0.75)
n
times the area of the original triangle. So after an infinite number of
iterations, we would find there was no area at all.




2.1.3. Number of Triangles Grow
After observing the number of triangles pointing down for several iterations, we
have the following table:

Iteration No. of Triangles pointing down
1 1
2 4
3 13
4 40
5 121

From the table, we can come up with a general formula to predict the number of
triangles being removed for any iteration:

i=n-1
At nth iteration, the number of triangles being removed, N =

i=0

2.2. Iterated Funtion System
Fractals can be constructed through limits of iterative schemes involving generators of
iterative functions on metric spaces. Iterated Function System (IFS) is the most common,
general and powerful mathematical tool that can be used to generate fractals. Moreover,
IFS provides a connection between fractals and natural images. It is also an important
tool for investigating fractal sets. In the following, an introduction to some basic
geometry of fractal sets will be approached from an IFS perspective. In a simple case,
IFS acts on a segment to generate contracted copies of the segment which can be
arranged in a plane based on certain rules. The iteration procedure must converge to get
the fractal set. IFS replaces polygons by other polygons as described by a generator
(Nikiel 2007). On every iteration, each polygon is replaced by a suitably scaled, rotated,
and translated version of the polygons in the generator. Figure 1 shows two such
generators made of rectangles and the result after one and six iterations. From this
geometric description it is also possible to derive a hopalong description which gives the
image that would be created after iterating the geometric model to infinity. The
description of this is a set of contractive transformations on a plane of the form

Hopalong or "The Chaos Game"
A technique exists by which the resulting
form after an infinite number of iterations
can be derived. This is a function of the
form

This gives a series of (x,y) points all
which lie on the result of an infinite IFS.
Although it still takes an infinite number
of terms in this series to form the result
the appearance can be readily appreciated
after a modest number of terms.

For example consider the generator on the right.
It consists of three rectangles, each with its own center,
dimensions and rotation angle. The initial conditions
usually consist of a single square, the first iteration then
consists of replacing this square by a suitably positioned,
scaled and rotated version of the generator.


The next iteration involves replacing each of the rectangles
in the current system by suitabled positioned, scaled, and
rotated versions of the generator resulting in the following

The next iteration replaces each rectangle above again by
the initial generator as shown

With this method, it is possible to create the image at any scale. In many but not all cases
zoomed in examples will be exhibit self-similarity at all scales. Applications generally involve
data reduction for model files. If a generator can be found for a complex image then storing the
generator and the rules of production results in a great deal of data reduction.
Another examples:












2.2.1. IFS Application
Fractal geometry is at the conceptual core of understanding nature's complexity
and IFS provides an important concept for understanding the core design of the
natural objects as well as approximating the natural design. The Banach's fixed-
point theorem forms the basis of the IFS applications. However, applying the
theorem in practiced raises two central questions. One is to find the attractor for a
given IFS. The other is to find IFS for a given attractor, an inverse problem of the
first.
For the first problem, the attractor can be obtained by successive approximations
from any starting point theoretically. From a computational point of view, two
techniques, deterministic and stochastic, can be applied. The deterministic
algorithm starts with an arbitrary initial set to reach the attractor. The stochastic
algorithm is often more complex but more efficient. A stochastic algorithm
attributes to the IFS system with a set of probabilities by assigning a probability to
each mapping, which is used to generate a random walk. If we start with any point
and apply transformations iteratively, chosen according the probabilities attached,
we will come arbitrarily close to the attractor. The associated probabilities
determine the density of spatially contracted copies of the attractor. Therefore, the
probabilities have no effect on the attractor but influence significantly the
rendering of its approximations. The second problem, the inverse problem, can be
solved by Barnsley's Collage Theorem, a simple consequence of Banachs fixed
point theorem. Such procedure was illustrated nicely through the 'Barnsley fern' in
and using four-transformation IFS with associated probabilities. Figure below
shows our calculation examples of fractals using four-transformation IFS with
variations and their associated probabilities produced by Matlab, where 20000
iterations were set. These fractals actually have more than one attractor. In Figure
below, the four-transformation matrices are


D has the probability 0.75 and others 0.083.




3. Fractal Features
A fractal often has the following features:
It has a fine structure at arbitrarily small scales.
It is too irregular to be easily described in traditional Euclidean geometric language.
It is self-similar (at least approximately or stochastically).
It has a simple and recursive definition.
It has a Hausdorff dimension which is greater than its topological dimension.
3.1. Architecture
Architectural forms are handmade and thus very much based in Euclidean geometry, but
we can find some fractals components in architecture, too. We can divide the fractal
analysis in architecture in two stages [14]:
3.1.1. Little scale analysis (analysis of a single building)
The Little Scale Analysis comprises of:
3.1.1.1. The building's self-similarity
The buildings self-similarity is a building's component which repeats
itself in different scales. In the research of the buildings' self-similarity
by the 1104, the cathedral of Anagni (Italy) has floor which is adorned
with dozens of mosaics, each in the form of a Sierpinski gasket fractal.
In the Figure (a) there is a part of one mosaic, showing the fractal at its
fourth stage of iteration. The cathedral and its floor were built in the year
1104 and is possibly the oldest handmade fractal object [15]. In Saint
Paul church in Strasbourg (France), there are presences of selfsimilar
shapes on each two towers. In Venice there are many palaces that have a
rising fractal structure (e.g., Ca' Foscari, Ca' d'Oro, Duke Palace,
Giustinian Palace), and for this reason we can talk about "fractal Venice"
(figure c).



Figure (a) Floor of the cathedral of Anagni

Figure (b) Sierpinski Gasket

Figure (c) Ca' d'Oro, Venice (1421-1440).

3.1.1.2 The box-counting Dimension
The box-counting dimension is to determine the fractal dimension of a
building) [8]. To understand fractal concepts, we have to familiar with
two dimensions [8, 14]:
self- similarity dimension (Ds);
box-counting dimension (Db).

All these dimensions are directly related to Mandelbrot's fractal
dimension (D). In all self-similar constructions there is a relationship
between the scaling factor and the number of smaller pieces that the
original construction is divided into [7]. This is true for fractal and non-
fractal structures. The relationship is the law:

Where a is the number of piece and s is the reduction factor. For non-
fractal structure the exponent D is an integer. Solving for D:

The self-similarity dimension (DS) is equivalent to the Mandelbrot's
fractal dimension (D). In the Koch curve the scaling factor is 1/3 and the
number of pieces is 4; the fractal dimension is:
Ds=log(4) /log(3) =1.26 .
The box-counting dimension is presented with the problem of determining
the fractal dimension of a complex two-dimensional image.
It is produced using this iterative procedure:
superimpose a grid of square boxes over the image (the grid size as
given as s1);
count the number of boxes that contain some of the image (N(s1));
repeat this procedure, changing (s1), to smaller grid size (s2);
count the resulting number of boxes that contain the image (N(s2));
These procedures are repeated changing s to smaller and smaller
grid sizes.
Box-counting dimension is:

Where 1/s is the number of boxes across the bottom of the grid. We can
apply in architecture the box-counting dimension. It is calculated by
counting the number of boxes that contain lines from the drawing inside
them.
Application
Figure (d) illustrates the box count of a Frank Llyod Wright's building
(Robie house) [7]. Table 1 contains the number of boxes counted, the
number of boxes across the bottom of the grid, and the grid size.
Box Count Grid Size Grid Dimension (feet)
16 8 24
50 16 12
140 32 6
380 64 3
Table 1

Figure (d) Robin Houses Box-counting grids places over elevations
Three fractal dimensions can be calculated. The first is for the increase in
number of boxes with lines in them from the grid with 8 boxes across the
bottom (24 feet) to the grid with 16 boxes across the bottom (12 feet):

The next scanning range compares boxes that are 12 feet across with
boxes that are 6 feet across:

The final scanning range compares boxes that are 6 feet across with
boxes that are 3 feet across:

The last two calculations are in closer agreement than the first
calculation.

3.1.2. Large scale analysis (the urban growth)
The exposition of fractal geometry by Mandelbrot provided a critical tool for the
study of urban growth, and since 1990, the discipline of urban geography has
developed a large literature with an emphasis on the use of fractals to study these
complex irregularities. Mathematical modeling of fractal growth has been used to
simulate and understand urban growth patterns (White and Engelen, 1993;
Andersson, et al, 2002a; Andersson, et al 2002b; Onural, 1991; Paget, 1999;
Batty, 2005). Makse, et al (1995; 1998) used correlated percolation simulation to
produce a pattern similar to the growth of Berlin from 1850 through 1945 (Figure
e,f,g).

Figure (e) Berlin 1875

Figure (f) Berlin 1920


Figure (g) Berlin 1945
3.1.3 Conclusion
The research of the building's self-similarity and the calculus of the building's
fractal dimension are only two different approaches in the fractal little scale
analysis. Other fractal components are present in architecture:
Fractal geometrical description of real built-up area (e.g. using the self
similarity or some particular fractal shape);
Simulation of urban growth using fractal algorithm (e.g. using the
Diffusion Limited Aggregation (DLA) model 1) [5, 6].
In the first approach there is present of an interesting syllogism [11, 14]:
The house like a part of the town, and it is little image of the town, too;
The town is constituted from copies of itself; this is an example of self-
similarity.
Other fractal approach in architecture is to find the connections between the
structure of the towns and the fractal sets (e.g. Julia sets or Mandelbrot sets) [11].
The geographer Michael Batty suggested that the fractal geometry can describe the
urban growth: "The morphology of cities bears an uncanny resemblance to those
dendritic clusters of particles which have been recently simulated as fractal growth
processes" [6]. Fractal Cities is the title of Batty and Longley's (1994) book in which
they show how Mandelbrot's (1983) theory of fractal geometry can be applied to the
study of cities, their structure and evolution [12]. They show how complex geometries
of urban form, growth and evolution, can be generated by means of Mandelbrot's type of
fractal. This is an example of fractal approach in large scale [13]. Portugali introduces a
new idea in the urban growth: "Cities are self-organizing systems" 2 [13]. To
substantiate his revolutionary concept, he uses several interlinked methods. He employs
in his argument theoretical tools developed in the interdisciplinary field of synergetics.
He has performed detailed model calculations on cellular nets [17]. The aim of this work
is to present how the fractal geometry is helping to newly define a new architectural
models and an aesthetic that has always lain beneath the changing artistic ideas of
different periods, schools and cultures [9, 10, 12, 14].











References
[1] http://www.merriam-webster.com/dictionary/fractal
[2] http://www.fractalus.ch/biograf/bio_e.html
[3] http://www-history.mcs.st-and.ac.uk/Biographies/Mandelbrot.html
[4] http://www.math.nus.edu.sg/aslaksen/gem-projects/maa/World_of_Fractal.pdf
[5] M. Batty, P. Longley, S. Fotheringham, Urban growth and form: scaling, fractal geometry,
and diffusion-limited aggregation. In Environment and Planning, 1989, vol. 21, pp. 1447 -
1472
[6] M. Batty, Cities as Fractals: Simulating Growth and Form. In Crilly A. J., Earnshaw R. A.,
Jones H., Fractals and Chaos, Springer - Verlag, 1991, New York, pp. 43 69
[7] M Batty, P. Longley, Fractal Cities, Academic Press, 1994, London

[8] C. Bovill, Fractal Geometry in Architecture and Design, Birkhuser, 1995, Boston

[9] J. Briggs, Fractals The Patterns of Chaos, Thames and Hudson, 1992, London

[10] J. Briggs, Estetica del caos, Red Edizioni, 1993, Como

[11] F. Donato, L. Lucchi Basili, L'ordine nascosto dell'organizzazione urbana, Franco Angeli
Editore, 1996, Milano

[12] B. Mandelbrot, The Fractal Geometry of Nature, 1983, W.H. Freeman, New York

[13] J. Portugali, Self-Orgnizing and the City, Springer, 1999, Berlin

[14] N. Sala, Chaos and Fractals in Arts and Architecture. International Journal of Chaos and
Applications, 2000 (will appear)

[15] N. Sala, Some News About The Fractals, 1997, resource online available:
http:// www. arch.unisi.ch/fractals/fract1e.htm