Upload
skh
View
217
Download
0
Embed Size (px)
Citation preview
8/14/2019 Fractal 5
1/16
Fractal (Fractile)
Introduction
1. Derived from the Latin wordfractus, meaning fragmented, or broken. It was coined by the Polish-born mathematician Benoit B. Mandelbrot in 1975 when he asked a question about the length of the
coastline of Britain. If one takes aerial pictures of the coastline at different heights (or use ruler of
different divisions), one sees different amount of details. Since the coastline is irregular with bays, inlets,
etc., the estimates of the coastline become arbitrarily large. Does it approach infinity?
2. This idea is different from estimating the circumference of a circle (which is regular) using regular
polygons. The estimates converge.
3. A fractal curve is one whose length becomes arbitrarily large as the measuring scale becomes smaller and
smaller. A fractal may have one or more of the following characteristics:
(a) complex structure at all scales
(b) self-similarity
(c) fractional dimension
(d) infinite branching
(e) chaotic dynamics
4. There are numerous applications to diverse fields dealing with irregularly shaped objects or spatially non-
uniform and random phenomena: the rugged terrain of mountains, coastlines, Brownian movement,
vascular networks, population growth, the shapes of polymer molecules, distribution of galaxies, and the
intricate branch systems of trees and ferns in nature. Computer graphics are used to generate lifelike
images of these complicated, highly irregular natural objects and processes.
5. Introduction of fractals in school mathematics has the following benefits:
(a) as a recent topic in mathematics, it shows students that mathematics is a constantly developing
discipline (unlike most school mathematics which has been fossilised for thousands of years);(b) they provide visual illustrations of many basic mathematical results, such as sequences, indices,
limit, complex numbers, and symmetry (a multi-modal approach);(c) challenge students to find and explore visual patterns, linking these activities to computer graphics
(use ofExcel, Logo, and special fractal-generating programs); the planet in Star Trekwas computer-generated fractal landscape;
(d) it is fun to discover new fractals;
(e) develop more positive attitudes towards learning mathematics.
Self-Similar Objects
1. A self-similar object: its parts resemble the whole; each part, when magnified, will look basically like the
object as a whole; it remains invariant under changes of scale, i.e., it has scaling symmetry (self-symmetry
under magnification).
2. Koch curve (by Helga von Koch,1870 1924, a Swedish mathematician, introduced this curve in1904).
On a piece of A4 paper, draw a line segment of length 18 cm. Take this as a unit length. Trisect it and
replace the middle segment by the other two sides of an equilateral triangle of side31
the original length.
Repeat the process up to stage 3. Use pen of different colour for each stage.
3. (Logo) Load fractal.lgo. Type side1 50 0 and change the side (first input) and stage (second input).
1
8/14/2019 Fractal 5
2/16
Stage
nLength of each segment
sn
Number of segments
pn
Total length
Ln0 1 1 1
131 4
34
2
n ( )n31 4
n
( )n34
4. The Koch curve is self-similar. As n , the total length . Imagine this curve is used to estimate thelength of a coastline. As the scale becomes smaller, the total length becomes increasingly larger.
5. Repeat the above using the following division (quadric Koch curve). (Logo) Type side2 50 0 and so on.
6. These fractal curves are highly complex and contain a great deal of detail. The degree of complexity of
self-similar fractals is described by thesimilarity fractal dimension. These curves have fractional
dimensions, so FRACTAL also stands for FRACTionAL dimension.
7. LetNbe the ratio of the number of final pieces to the number of original pieces (replacement ratio) and rthe ratio of the length of the originalpiece to the length of each final piece (scaling ratio). Then the
dimensionD is defined by NrD = or rN
Dlog
log= . This is an extension of the dimensions of classical
geometric objects, and is based on the notion of Hausdorff dimension for topological spaces.
Figure r N Dimension,D
Line 2 2 = 21 1
Square 2 4 = 22 2
Cube 2 8 = 23 3
Koch curve 3 4 = 3D 1.26
Quadric Koch curve 3 5 1.46
8. These curves form the L-system fractals, where L stands for Lindenmayer, a mathematician who wrote
about the fractal nature of plant development.
2
8/14/2019 Fractal 5
3/16
8/14/2019 Fractal 5
4/16
8. Curves like the Sierpinski curve are closed paths that pass through every point in the unit square. Space-
filling curve will do the same for a cube.
9. Sierpinski triangle (or gasket). Construct an equilateral triangle of side 12 units on an isometric dot paper.
At step 1, bisect each side and shade the middle triangle. This leaves three unshadedtriangles. See the first
two stages below.
Stage
n
Number ofunshadedtriangles
pn
Total area ofunshadedtriangles
An0 1 1
1 34
3
n 3n ( )n4
3
Show that its dimension is2log
3log 1.58.
10. Sierpinski triangle and the chaos game. This activity demonstrates that a non-deterministic process may
lead to regularity.
Draw an equilateral triangle with top vertex T and left and right vertices L and R. Take any point inside
the triangle. Throw a die and mark a point accordingly:
For 1 or 2, mark the mid-point between the current point and L.For 3 or 4, mark the mid-point between the current point and T.
For 5 or 6, mark the mid-point between the current point and R.
Repeat this process for a number of points.To collect class result, give each member a transparency of the same triangle and ask them to mark using
different colours. Put them together later.
11. Let L = (-1, 0), R = (1, 0), and T = (0, 3). The chaos game is generated by three linear mappings:
L =
221 ,
yx, R =
+
221 ,
yx, T =
+
2
3
2,yx
.
12. Instead of taking the mid-point between a point and a vertex, choose a line that divides the segment from
the point to a vertex in the ratio k: (1 k). Ifk< 0.5, we have an overlaid Sierpinski triangle; ifk> 0.5,we have a strict Sierpinski triangle where the triangles have no common point. (Logo) Type sierst andenter the value ofkto use.
13. Repeat the above using a square (or any regular polygon!). (Logo) Type siersqst.
4
8/14/2019 Fractal 5
5/16
Chaos
Many physical phenomena like a swinging pendulum are deterministic and regular. However, some aredeterministic but irregular (atmospheric conditions); a minor change in one parameter may give rise to drastic
changes; this is a state of chaos.
A Logistic Example
1. Letpn + 1 be the population at year (n + 1). A simple model may be:
pn + 1 =pn + b pn= (1 + b) pn, where b is the birth rate (> 0).This model will show that the population will grow without bound!
2. Letpn + 1 = (1 + b)s pn, wheres is the survival rate. Here the overall growth rate is (1 + b)s.
If 0 < (1 + b)s < 1, the population will eventually die out.If (1 + b)s = 1, the population will remain constant.If (1 + b)s > 1, the population will grow without bound.
3. These linear models are not realistic. May and Feigenbaum examine a logistic function. Suppose there is
a maximum populationPthat the resources can support. Multiple the above equation by (P pn); notice
that aspn increases, this term becomes smaller. We have
pn + 1 = (1 + b)s(P pn)pn orP
p
P
p
P
p nnn
=+ 11 , or
xn+1 = (1 xn) xn,where 0
8/14/2019 Fractal 5
6/16
8/14/2019 Fractal 5
7/16
4. The boundary of the prisoner set (consisting of thez0 values) is called a Julia set for the specific value ofc.
5. (Logo) Type juliaset to plot all the points defined by -1.5 x 1.5 and -1.5 y 1.5. Examples to try:c = 0.3; c = -0.05 + 0.745i, c = -0.5; c = -0.71 + 0.36i; c = 0.27 + 0.49i. The cross shows the position of thepoint c; in the last two cases, the regions cover the point c. Each diagram takes about 15 minutes.
c = 0.3 c = -0.05 + 0.745i c = -0.6
Mandelbrot Set
1. These sets are named after the mathematician, Benoit Mandelbrot, who first produced them in 1979. The
Mandelbrot set is shown below. It consists of a main bay (the lake), three baylets, and fractal shorelines.
2. Each value of the Mandelbrot set corresponds to one Julia set c. Ifc is inside the lake, the correspondingJulia set is connected. As c moves towards the shore, the Julia set becomes convoluted and eventually
becomes disconnected.
3. (Logo) The mandelst procedure useszz2 + 1.5c,and points are plotted for 1.5 x 0.5 and 1 y 1.Using the default parameters, the following plot is obtained; time taken is about 12 minutes. In the
construction, take a value ofc and check whether the origin is a prisoner. If so, the value ofc is included(plotted).
4. In colouring Mandelbrot set, points outside the set are coloured according to the number of iterations to
test a point. For example, if a test point escapes in a few iterations, it might be coloured red; if it escapes
in many iterations, it might be coloured blue, and so on.
References
Bills, C., & Bills, L. (1997). A little bit of chaos. In M. Sewell (Ed.), Mathematics masterclass: Stretching the
imagination (pp. 1-15). Oxford: Oxford University Press.Field, M., & Golubitsky, M. (1992). Symmetry in chaos: A search for pattern in mathematics, art and nature.
Oxford: Oxford University Press.Muller, J. (1997). The great Logo adventure: Discovering Logo on and off the computer. Madison, AL: Doone
Publications.
Peitgen, H.O., Jrgens, H., Saupe, D., Maletsky, E., Perciante, T., & Yunker, L. (1991).Fractals for theclassroom: Strategic activities (Vol. 1). New York: Springer-Verlag & NCTM .
Fractals: A Fractal Unit for Elementary and Middle School Students (http://math.rice.edu/~lanius/frac/)
7
http://math.rice.edu/~lanius/frac/http://math.rice.edu/~lanius/frac/8/14/2019 Fractal 5
8/16
The Spanky Fractal Database of the Stone Soup Group (http://spanky.triumf.ca/www/fractint/fractint.html) for
the freeware,Fractint, used to generate various types of fractals under DOS. Instructions and examplesgiven in Wegner, T. & Tyler, B. (1993).Fractal creations (2nd ed.).
Fractal Pictures and Animations (http://www.cnam.fr/fractals.html)
Logo Procedures
TO SIDE1 :SIZE :LEVEL
IF :LEVEL = 0 [FD :SIZE STOP]SIDE1 :SIZE / 3 :LEVEL - 1 LT 60
SIDE1 :SIZE / 3 :LEVEL - 1 RT 120
SIDE1 :SIZE / 3 :LEVEL - 1 LT 60
SIDE1 :SIZE / 3 :LEVEL - 1
END
TO KOCH :SIZE :LEVEL
REPEAT 3 [RT 120 SIDE1 :SIZE :LEVEL]
END
TO KOCHST
CS WINDOW PU SETXY [-200 80] PD
SETPC 0 KOCH 200 1 WAIT 30SETPC 9 KOCH 200 2 WAIT 30
SETPC 5 KOCH 200 3 WAIT 30
SETPC 12 KOCH 200 4
END
TO SIDE2 :SIZE :LEVEL
IF :LEVEL = 0 [FD :SIZE STOP]
SIDE2 :SIZE / 3 :LEVEL - 1 LT 90
SIDE2 :SIZE / 3 :LEVEL - 1 RT 90
SIDE2 :SIZE / 3 :LEVEL - 1 RT 90
SIDE2 :SIZE / 3 :LEVEL - 1 LT 90
SIDE2 :SIZE / 3 :LEVEL - 1
END
TO QKOCH:SIZE :LEVEL
REPEAT 4 [RT 90 SIDE2 :SIZE :LEVEL]
END
TO QKOCHST
CS WINDOW PU SETXY [-200 80] PD
SETPC 0 QKOCH 100 1 WAIT 30
SETPC 9 QKOCH 100 2 WAIT 30
SETPC 5 QKOCH 100 3 WAIT 30
SETPC 12 QKOCH 100 4
END
TO SIDE3 :SIZE :LEVELIF :LEVEL = 0 [FD :SIZE STOP]
SIDE3 :SIZE / 3 :LEVEL - 1 LT 90
SIDE3 :SIZE / 3 :LEVEL - 1 RT 90
SIDE3 :SIZE / 3 :LEVEL - 1 RT 90
SIDE3 :SIZE / 3 :LEVEL - 1 RT 90
SIDE3 :SIZE / 3 :LEVEL - 1 LT 90
SIDE3 :SIZE / 3 :LEVEL - 1 LT 90
SIDE3 :SIZE / 3 :LEVEL - 1 LT 90
SIDE3 :SIZE / 3 :LEVEL - 1 RT 90
SIDE3 :SIZE / 3 :LEVEL - 1
END
8
http://spanky.triumf.ca/http://www.cnam.fr/fractals.htmlhttp://spanky.triumf.ca/http://www.cnam.fr/fractals.html8/14/2019 Fractal 5
9/16
8/14/2019 Fractal 5
10/16
TO SIER
PU
MAKE "V RANDOM 3
IF :V=1 [SETPC 1 SETXY LIST (1-:C)*XCOR (1-:C)*YCOR+119*:C]
IF :V=2 [SETPC 2 SETXY LIST (1-:C)*XCOR-139*:C (1-:C)*YCOR-119*:C]
IF :V=3 [SETPC 12 SETXY LIST(1-:C)*XCOR+139*:C (1-:C)*YCOR-119*:C]
PD FD 1 BK 1
SIER
END
TO SIERST
CS WINDOW PU
TYPE [ENTER THE DIVISION RATIO (SAY .5)] MAKE "C READ
SETXY LIST ( 140 - RANDOM 280) (120-RANDOM 240)
PD FD 1 BK 1
SIER
END
TO SIERSQ
PU MAKE "V RANDOM 4
IF :V=1 [SETPC 1 SETXY LIST (1-:c)*XCOR-139*:C (1-:C)*YCOR+119*:C]
IF :V=2 [SETPC 4 SETXY LIST (1-:C)*XCOR+139*:C (1-:C)*YCOR+119*:C]
IF :V=3 [SETPC 9 SETXY LIST (1-:C)*XCOR-139*:C (1-:C)*YCOR-119*:C]
IF :V=4 [SETPC 12 SETXY LIST (1-:C)*XCOR+139*:C (1-:C)*YCOR-119*:C]
PD FD 1 BK 1
SIERSQ
END
TO SIERSQST
CS WINDOW PU
SETXY LIST ( 140 - RANDOM 280) (120-RANDOM 240)
PD FD 1 BK 1
TYPE [ENTER THE DIVISION RATIO (SAY .5)] MAKE "C READ
SIERSQ
END
TO JULIAPT
CS HT WINDOW SETPC 0
TYPE [ENTER THE SCALING FACTOR FOR AXES (SAY 60) ==>] MAKE "SCALE READ
TYPE [ENTER THE REAL PART OF C ==>] MAKE "CR READ
TYPE [ENTER THE IMAGINARY PART OF C ==>] MAKE "CI READ
MAKE "LIMIT 0.5 + SQRT(.25+ SQRT (:CR*:CR + :CI*:CI))
SETPC 12 DOT LIST :CR*:SCALE :CI*:SCALE
TYPE [HOW MANY ITERATIONS TO TAKE (SAY 100) ==>] MAKE "TIMES READ
GETPOINTS1
END
TO AXES
CS SETPC 0 BK 2*:SCALE TICK
REPEAT 4 [FD :SCALE TICK] BK 2*:SCALE RT 90 BK 2*:SCALE TICK
REPEAT 4 [FD :SCALE TICK] HOME
END
TO TICK
LT 90 FD 3 BK 6 FD 3 RT 90
END
10
8/14/2019 Fractal 5
11/16
TO GETPOINTS1
TYPE [ENTER THE REAL PART OF A COMPLEX NUMBER ==>] MAKE "R0 READ
TYPE [ENTER THE IMAGINARY PART OF THE COMPLEX NUMBER ==>] MAKE "I0 READ
MAKE "RT :R0 MAKE "IT :I0
AXES
PU SETXY LIST :R0*:SCALE :I0*:SCALE PD SETPC 12
CHECKSTATUS :RT :IT
IF :J = 0 (PRINT :R0 :I0 "PRISONER)
IF :J = 1 (PRINT :R0 :I0 "ESCAPER)
GETPOINTS1END
TO CHECKSTATUS :XR :XI
MAKE "J 0
JULIA 1
END
TO JULIA :T
IF :T = :TIMES [STOP]
MAKE "R1 (:RT*:RT - :IT*:IT + :CR)
MAKE "I1 2*:RT*:IT+:CI
MAKE "DIST SQRT (:R1*:R1 + :I1*:I1)
SETXY LIST :R1*:SCALE :I1*:SCALE
(PR :R1 :I1 "DISTANCE :DIST) WAIT 10
IF :DIST > :LIMIT [MAKE "J 1 STOP]
MAKE "RT :R1 MAKE "IT :I1
JULIA :T+1
END
TO JULIASET
CS WINDOW HT
TYPE [ENTER THE SCALING FACTOR FOR AXES (SAY 50) ==>] MAKE "SCALE READ
TYPE [ENTER THE REAL PART OF C ==>] MAKE "CR READ
TYPE [ENTER THE IMAGINARY PART OF C ==>] MAKE "CI READ
MAKE "LIMIT 0.5 + SQRT(.25+ SQRT (:CR*:CR + :CI*:CI))
SETPC 12 DOT LIST :CR*:SCALE :CI*:SCALE MARK
MAKE "TIMES 20TYPE [ENTER THE NUMBER OF DIVISIONS TO TAKE (SAY 200)==>] MAKE "D READ
MAKE "LP 3/:D
FOR "N 0 :D [MAKE "R0 (-1.5)+:N*:LP FOR "M 0 :D [MAKE "I0 (-1.5)+:M*:LP
GETPOINTSJ]]
END
TO MARK
PU SETXY LIST :CR*:SCALE :CI*:SCALE SETH 45 PD
REPEAT 4 [FD 5 BK 5 RT 90]
PU
END
TO GETPOINTSJ
MAKE "RT :R0 MAKE "IT :I0
CHECKSTATUSJ :RT :IT
IF :J = 0 [(PRINT :R0 :I0 "PRISONER) SETPC 0 DOT LIST :R0*:SCALE :I0*:SCALE]
;IF :J = 1 [(PRINT :R0 :I0 "ESCAPER) SETPC 13 DOT LIST :R0*:SCALE :I0*:SCALE]
END
TO CHECKSTATUSJ :XR :XI
MAKE "J 0
JULIAJ 1
END
11
8/14/2019 Fractal 5
12/16
TO JULIAJ :T
IF :T = :TIMES [STOP]
MAKE "R1 (:RT*:RT - :IT*:IT + :CR)
MAKE "I1 2*:RT*:IT+:CI
;(PR :R1 :I1) WAIT 10 (INCLUDE THIS TO SHOW THE SEQUENCE)
IF (SQRT (:R1*:R1 + :I1*:I1)) > :LIMIT [MAKE "J 1 STOP]
MAKE "RT :R1 MAKE "IT :I1
JULIAJ :T+1
END
TO MANDELST
CS WINDOW HT
TYPE [ENTER THE SCALING FACTOR FOR AXES (SAY 50) ==>] MAKE "SCALE READ
TYPE [ENTER THE NUMBER OF C TO CHECK (SAY 500)==>] MAKE "C READ
TYPE [ENTER THE LIMIT OF DISTANCE TO TEST (SAY 10) ==>] MAKE "LIMIT READ
TYPE [ENTER THE MAXIMUM NO. OF ITERATION (SAY 20) ==>] MAKE "TIMES READ
MAKE "LP 2/:C
FOR "N 0 :C [MAKE "CR (-1.5)+:N*:LP FOR "M 0 :C [MAKE "CI (-1)+:M*:LP GETPOINTS3]]
END
TO GETPOINTS3
MAKE "R0 0 MAKE "I0 0
MAKE "RT :R0 MAKE "IT :I0
CHECKSTATUS3 :RT :IT
IF :J = 0 [(PRINT :CR :CI "PRISONER) SETPC 0 DOT LIST :CR*:SCALE :CI*:SCALE]
;IF :J = 1 [(PRINT :CR :CI "ESCAPER) SETPC 13 DOT LIST :CR*:SCALE :CI*:SCALE]
END
TO CHECKSTATUS3 :XR :XI
MAKE "J 0
MANDEL 1
END
TO MANDEL :T
IF :T = :TIMES [STOP]
MAKE "R1 (:RT*:RT - :IT*:IT + 1.5*:CR)
MAKE "I1 2*:RT*:IT+1.5*:CIIF (SQRT (:R1*:R1 + :I1*:I1)) > :LIMIT [MAKE "J 1 STOP]
MAKE "RT :R1 MAKE "IT :I1
MANDEL :T+1
END
12
8/14/2019 Fractal 5
13/16
Koch Curve (stage 3)
Koch Curve (stage 4)
Quadric Koch Curve (stage 3)
Koch Snowflake (stages 1, 2, 3)
13
8/14/2019 Fractal 5
14/16
Koch Anti-Snowflake (stages 1, 2, 3)
Peano Curve (stages 1, 2, 3)
Quadric Island Curve (stages 1, 2, 3)
Sierpinski Triangle
14
8/14/2019 Fractal 5
15/16
Julia Sets
c = 0.3 c = -0.05 + 0.745i
c = -0.6
Mandelbrot Set
15
8/14/2019 Fractal 5
16/16
Instructions for Fractals on PC Logo
In each case, change the input values and see what happen!
Koch Curve ? cs rt 90? side1 100 0
? side1 100 1? side1 100 2
Quadric Koch Curve ? cs rt 90? side2 100 0? side2 100 1? side2 100 2
Koch Snowflake ? cs koch 100 0? koch 100 1? koch 100 2? kochst
Koch Anti-Snowflake ? cs kochanti 100 0? cs kochanti 100 1? cs kochanti 100 2? kochantist
Quadric Koch Snowflake ? cs qkoch 100 0
? qkoch 100 1? qkoch 100 2? qkochst
Peano Curve ? cs rt 90? side3 100 0? side3 100 1? side3 100 2
Quadric Island ? cs rt 90? side4 100 0? side4 100 1? side4 100 2
Sierpinski Triangle as a chaosgame
sierst (enter a value between 0 and 1)
16