You are on page 1of 10

International Conference on Computational Sciences and Its Applications ICCSA 2008

Matlab Toolbox and GUI for Analyzing One-Dimensional Chaotic Maps

Akemi Gálvez Tomida


Department of Applied Mathematics and Computational Sciences, University of Cantabria, Spain
galveza@unican.es

Abstract cuits and lasers, meteorology, ecology, etc. Nonlinearity


appears in both discrete and continuous systems, which are
Although the notion of chaos is usually associated described by iterated functions (or maps) and (either or-
with continuous systems, discrete systems can also exhibit dinary or partial) differential equations (or flows), respec-
chaotic behavior. In fact, discrete systems have less require- tively. That means that the accurate analysis of such chaotic
ments than their continuous counterparts to show chaotic systems requires a bulk of specialized mathematical tools
regime, as chaos can be found in simple one-dimensional and techniques, designed to account for the kind of system
non-invertible maps. This paper deals specifically with this involved. In terms of computer software, this means that we
last kind of systems. To this aim, the author introduces a need very powerful computer programs and packages which
new Matlab toolbox and GUI (Graphical User Interface) are, at their own, extraordinarily resource-demanding.
providing the user with a suitable toolkit for a complete The list of numerical, symbolic and graphical require-
analysis of those systems by computer. The paper describes ments is so large that developers are now looking at the
the software developed to reach these goals, its main fea- computer algebra systems (CAS) as the key to tackle this
tures and components and illustrates its performance by ap- issue. As a consequence, there is an increasing interest on
plying it to the analysis of important issues regarding one- CAS-based programs for dynamical systems. Some exam-
dimensional maps, such as the calculation of fixed points, ples include [1, 2, 5] for Matlab, [4, 7, 10, 11, 12, 17] for
the study of their stability and the representation of their bi- Mathematica and [18] for Maple. In this context, the author
furcation diagram and Lyapunov exponents, cobweb plots presented last year a numerical-symbolic Matlab program
and phase space graphs for various dimensions. In our for the analysis of three-dimensional chaotic systems [3].
opinion, the tool described here is very useful in order to The program provides the users with a GUI (Graphical User
carry out such calculations in an efficient and simple way. Interface) that allows them to analyze (either symbolically
or numerically) any 3D flow with a minimal input, namely,
the symbolic ordinary differential equations describing the
1. Introduction system and (sometimes) some relevant parameters.
Although the notion of chaos is very often associated
Although the nonlinear systems and their wealth of with nonlinear continuous systems, discrete systems can
chaotic behaviors are known for a long time, only with also exhibit chaotic behavior. In fact, low-dimensional
the impressive developments in hardware and software hap- maps contain most of the phenomena commonly encoun-
pened during the last few decades we have been able to an- tered in higher-dimensional continuous systems. While, ac-
alyze them at full extent. It is well-known that the behavior cording to the Poincaré-Bendixson theorem, a strange at-
of nonlinear systems is much more complicated than that tractor can only arise in a continuous dynamical system
of linear systems. In fact, even the simplest chaotic sys- if it has three or more dimensions, no such restriction ap-
tems exhibit a number of different behaviors that can only plies to discrete systems, which can exhibit strange attrac-
be fully analyzed with the help of powerful hardware and tors in two or even one dimensional systems. In fact, one-
software resources. Because of that, the analysis of chaotic dimensional non-invertible maps are the simplest systems
systems has been a classical subject of research in the realm capable of chaotic motion. Consequently, one-dimensional
of Computational Science. maps are used as a very convenient starting point for the
The range of different phenomena associated with the study of chaos.
nonlinear systems is extremely varied. Chaos is ubiqui- In this paper, a new Matlab toolbox and GUI for analyz-
tous in nature: chaotic behaviors can be found in almost ing one-dimensional chaotic maps is introduced. The struc-
any field, ranging from chemical reactions to electronic cir- ture of this paper is as follows: in Section 2.1 the archi-

978-0-7695-3243-1/08 $25.00 © 2008 IEEE 321


DOI 10.1109/ICCSA.2008.7
tecture of our program and its main components are briefly
discussed. Some implementation issues are briefly reported
in Section 2.2. Then, Section 3 describes the main features
of the program through its application to the analysis of
some issues regarding one-dimensional maps: fixed points
and their stability in Section 3.1, bifurcation diagrams and
Lyapunov exponents in Sections 3.2 and 3.3 respectively,
cobweb plots in Section 3.4 and phase space graphs in Sec-
tion 3.5. For studying all these items, different 1D-maps are
used. The conclusions and some further remarks close the
paper.

2. Program Architecture and Implementation

2.1. Program Architecture

The program introduced in this paper for the analysis of


one-dimensional chaotic maps consists basically of two ma-
jor components:

1. a numerical-symbolic toolbox: a collection of com-


mands, functions and routines implemented to perform
the numerical, symbolic and graphical tasks.

2. a graphical user interface (GUI): this component is re-


sponsible for input/output windowing and smooth in- Figure 1. Screenshot of a typical window of
teraction with the user. the GUI for our toolbox
At its turn, the toolbox is comprised of three different
modules, according to the processes (numerical, symbolic
and graphical) to be carried out: symbolic object-string conversion or symbol replace-
ment and assignment) have also been used for sym-
a) a set of numerical libraries containing the implemen- bolic tasks, as it will be shown later on.
tation of the commands, functions and routines for the
numerical tasks. They have been implemented by us- c) some graphical commands. The powerful Matlab
ing the native Matlab programming language and tak- graphical capabilities exceed those commonly avail-
ing advantage of the large collection of numerical rou- able in other CAS such as Mathematica and Maple.
tines available in this system. Usually, these Matlab Although our current needs do not require to apply
routines provide extensive control on a number of dif- them at full extent, they avoid the users the tedious
ferent options and are fully optimized to offer the high- and time-consuming task to implement many routines
est level of performance. In fact, this is one of the for graphical output by themselves. Some nice view-
major strengths of the program and one of the main ing features such as 3D rotation, zooming in and out,
reasons to choose Matlab as its optimal programming labeling, scaling, coloring and others are also automat-
environment. ically inherited from the Matlab windows system.

b) a set of symbolic routines and functions. They have Although this toolbox is enough to fulfill almost all our
been implemented by using the Symbolic Math Tool- computation needs, potential end-users might be challenged
box that provides access to several Maple routines for for using it properly unless they are really proficient on
symbolic tasks. This symbolic module is more impor- Matlab’s syntax and functionalities and the toolbox rou-
tant than it might seem at first sight; for instance, sys- tines. This kind of limitations can be overcome by creating
tem equations are inputted symbolically so that some a GUI; a well-designed GUI uses readily recognizable vi-
functional operators (such as derivatives) can be ef- sual cues to help the user navigate efficiently through infor-
fectively applied. Further, some additional operators mation. Fortunately, Matlab provides a mechanism to gen-
(like those for string manipulation, forward/backward erate GUIs by using the so-called guide (GUI development

322
environment). This feature is not commonly available in
many other CAS so far. Although its implementation re-
quires - for complex interfaces - a high level of expertise,
it allows end-users to deal with the toolbox with a minimal
knowledge and input, thus facilitating its use and dissemi-
nation.
Based on the previous discussion, a GUI for the already-
generated toolbox has been implemented. Figure 1 shows
an example of a typical window. It allows an effective use
of powerful interface tools designed according to the type of
values being displayed (e.g., drop-down menu for a choice
list, radio buttons for boolean options, text boxes for dis-
playing messages, dialog boxes for input/output user inter-
action, etc.) In general, functions associated with each topic
under analysis have been grouped and arranged in a rectan-
gular area with an indicative title on the top, such as: fixed
point analysis, bifurcation diagram, Lyapunov exponents,
cobweb plot and phase space graph (see Figure 1 and sub-
sequent description in Section 3). Additional functionalities
have been provided in hidden menus or in separate windows
which can be invoked at will if needed so as to keep the
main window streamlined and uncluttered. For instance, all
graphical output is displayed in separate windows so that
the information is better organized and “flows” in a natural
and intuitive way. As a result, this GUI presents an interface
which is both aesthetic and very functional to the user.
Figure 2. Fixed points and their stability for
the logistic map
2.2. Implementation Issues

Regarding the implementation, this program has been


developed by the author in Matlab [14] v2007b on Win- of one-dimensional chaotic maps.
dows XP operating system by using a PC with Intel Core
2 Duo processor at 2.4 GHz. and 2 GB of RAM. However, 3.1. Fixed Points and Stability
the program supports many different platforms, such as PCs
(with Windows 9x, 2000, NT, Me, XP and Vista) and UNIX Given a map defined by a function f (x), a fixed point
workstations. A version for Apple Macintosh with Mac OS of f (x) is a point that is mapped to itself by the function.
X system is also available provided that X11 (the imple- In other words, x∗ is a fixed point of the function f if and
mentation of the X Window System that makes it possible only if f (x∗ ) = x∗ . For instance, consider the logistic map
to run X11-based applications in Mac OS X) is properly in- defined by the expression:
stalled and configured. Figures in this paper correspond to
the PC platform version. xn+1 = λxn (1 − xn ) (1)
The graphical tasks are performed by using the Matlab
where n ∈ IN and λ is a system parameter. The logistic map
GUI for the higher-level functions (windowing, menus, or
became popular following a seminal paper by the biologist
input) while the built-in graphics Matlab commands are ap-
Robert May in 1976 [15], where he introduced the discrete
plied for rendering purposes. The numerical kernel has been
version of a demographic model due to Verhulst. Roughly,
implemented in the native Matlab programming language,
xn ∈ [0, 1] represents the population at year n, evolving
and the symbolic kernel has been created by using the com-
from x0 , the initial population value. Depending on the λ
mands of the Symbolic Math Toolbox.
value, the system evolves among a number of different sit-
uations, from the population eventually dying to stabilizing
3. Some Illustrative Examples or changing chaotically (see Section 3.2 for more details).
The logistic map can be analyzed by using the toolbox
In this section the main features of the program are de- and GUI introduced in this paper. To do so, we must enter
scribed through its application to some illustrative examples the system equation as shown in Figure 2. Note that the

323
Figure 3. Bifurcation diagram of the logistic Figure 4. Close up of Fig. 3 on the interval
map on the interval (0, 4) (3.5, 4)

equation is given in a mathematical-looking way similar to


Eq. (1) so as the expression can be subsequently processed
for further symbolic-numerical calculations. Then, we can
compute the fixed points, which means to solve the equation
f (x) = x (or equivalently, f (x) − x = 0 so that the solve
command can be applied directly onto the expression).
We can also analyze the stability of those fixed points.
An attractive fixed point (or stable fixed point) of a function
f is a fixed point x∗A of f such that for any value of x in
the domain that is close enough to x∗A , the iterated function
sequence converges to x∗A . Conversely, iteration may give
the appearance of points diverging away from a fixed point;
this would be the case for repulsive fixed point (or unstable
fixed point). To perform such stability analysis, we consider
the eigenvalue of the fixed point, given by:
 
df (x)
φ= (2) Figure 5. Close up of Fig. 4 on the interval
dx x=x∗
(3.826, 3.86)
According to [13], the fixed point is stable (or attractive)
if |φ| < 1, neutral if |φ| = 1, unstable (or repulsive) if
|φ| > 1, and superstable (or superattractive) if φ = 0.
Figure 2 shows the stability analysis performed for the or periodic orbits) of such system as a function of a system
λ−1 parameter, let’s say p. It is computed as follows (see [16]
logistic map. The fixed points are: x = 0 and x = . for details):
λ
The stability of such points is established according to the
previous paragraph, as shown in the text box in Fig. 2. 1. Set p = p0 /initial value of the system parameter, p;
for instance, λ = 0 in Fig. 3 or λ = 3.5 in Fig. 4
3.2. Bifurcation Diagrams 2. Set x0 ∈ Im(f ) /initial value of x0 ; for the logistic
map, any value on the interval (0, 1) is admisible
Given a dynamical system, its bifurcation diagram
shows the possible long-term values (equilibria/fixed points 3. Iterate the map 500 times

324
Figure 6. Bifurcation diagram of the Mandel-
brot map on the interval (−2, 0)

4. Iterate the map another 1000 times (starting from x500 )


and plot the resulting 1000 values of x

5. Increase λ by a small amount and return to step 2. The


procedure is repeated until the final parameter value is
reached. /For instance, until λ = 4 in Figs. 3, 4 and
7, λ = 3.86 in Fig. 5, c = 0 in Fig. 6 or µ = 1.08 in
Fig.8.

Note that an initial transient is usually rejected to retain


the asymptotic behavior of the system. This is the reason to
discard the first 500 iterations in our plot (step 3). Note also
that only the stable solutions are shown here, but there are Figure 7. Bifurcation diagram of the cubic
many other unstable solutions which are not shown in this map on the interval (1, 4) for the initial con-
diagram. ditions x0 = 0.5 (top) and x0 = −0.5 (bottom)
Figure 3 shows the bifurcation diagram of the logistic
map for λ ∈ (0, 4). The initial input consists of the initial
point x0 and the initial and final values of the system param-
eter. Additional settings such as the number of iterations for λ−1
the transient and for the final plot are subsumed into specific . This is a fixed point of the system. However, for
λ √
menus in the upper toolbar. The corresponding graphical larger values of the parameter, λ ∈ (3, 1 + 6) two dif-
diagram is displayed in a separate window, where interac- ferent points appear meaning that, under iteration, the sys-
tive tools for graphical manipulations are automatically pro- tem will jump back and forth between these points. This is
vided. The bifurcation parameter λ is shown on the horizon- called a period-2 orbit. For λ ∈ (3.45, 3.54) approximately,
tal axis of the plot and the vertical axis shows the possible the population oscillate between four values (period-4 or-
long-term population values of the logistic function. From bit). Increasing λ beyond 3.54, the population will oscil-
Fig. 3, it can be seen that the behavior of the system changes late between 8 values, then 16, 32 and so on, following the
depending on the values of the system parameter. For val- so-called period doubling bifurcation route to chaos. For
ues of λ ∈ (0, 3), the population settles to a fixed size over λc > 3.569... (often referred to as the accumulation point),
the years. In particular, for λ ∈ (0, 1), the population will the system will evolve under an infinite number of points
eventually die, regardless the initial population, while for in a seemingly stochastic, or chaotic, way, although there
λ ∈ (1, 3) the population will quickly stabilize on the value are still certain isolated values of λ that show non-chaotic

325
bifurcation diagrams associated with this map, which are
obtained from two different initial conditions for the sys-
tem variable, namely x0 = 0.5 (top) and x0 = −0.5 (bot-
tom). As the reader can see, there are two missing branches
of the period-4 orbits in each diagram, so both pictures are
complementary each other.
Another interesting one-dimensional system is the
Singer map, described by:

xn+1 = µ(7.86xn − 23.31x2n + 28.75x3n − 13.30x4n ) (5)

where µ is the system parameter, which typically takes val-


ues on the interval (0.90, 1.08). In this case, the bifurcation
diagram, shown in Figure 8, exhibits the coexistence of a
stable fixed point until µ = 1.035... and a stable period-
2 orbit. Also, on the interval (0.99, 1.01) for µ a stable
period-1 and a stable period-2 orbits can also coexist. This
Figure 8. Bifurcation diagram of the Singer coexistence should however be understood as alternative,
map on the interval (0.9, 1.08) since each situation corresponds to different initial condi-
tions, so that the system can take only one of them at the
same time.
behavior, with orbits of period-3 (for λ = 3.825), period-6,
and so on. They can be seen, for instance, in Fig. 4 where a 3.3. Lyapunov Exponents
close up of Fig. 3 for λ ∈ (3.5, 4) is displayed.
The bifurcation diagram is a fractal: if you zoom in on One of the most important trademarks of chaos is the
the above-mentioned value λ = 3.825 and focus on one so-called sensitivity to initial conditions, meaning that two
branch of the three, the situation nearby looks like a shrunk initially closed arbitrary trajectories diverge exponentially
and slightly distorted version of the whole diagram (see Fig. over the time. The Lyapunov exponent (LE) of a dynamical
5). The same is true for all other non-chaotic points. This is system is the number that characterizes the rate of separa-
an example of the deep and ubiquitous connection between tion of these infinitesimally close trajectories.
chaos and fractals. Of course, the rate of separation can be different for dif-
Other example of chaotic one-dimensional map is the ferent orientations of initial separation vector, leading to as
Mandelbrot map, given by: many Lyapunov exponents as the number of dimensions of
the phase space. Thus, for one-dimensional systems, only
xn+1 = x2n + c (3) one LE is obtained. In the case of discrete systems given
by xn+1 = f (µ, xn ), the LE, denoted onwards as δ, can be
where c is the system parameter. The Mandelbrot map is computed as:
basically conjugate of the logistic map. Figure 6 shows its
m
bifurcation diagram for the system parameter c ∈ (−2, 0), 1 X
δ = lim Ln |f 0 (µ, x)| (6)
which resembles that of the logistic map as regarded in the m→∞ m
j=0
opposite direction. For c ∈ (−2.0, 0.25) the system exhibits
bounded orbits, while they diverge to infinity otherwise. where, in our computations, the limit to the infinity is re-
An interesting scenario appears when the branches in the placed by the summation over a large set of iterations. The
bifurcation diagram do depend on the initial values of the practical implication of the LE is that it is a good indicator
system variable, x0 . This happens, for instance, for the cu- of regular/chaotic behavior. In short, the LE is useful for
bic map, given by: distinguishing among the various types of orbits:

xn+1 = (1 − µ)xn + µx3n . (4) • δ < 0: this value is characteristic of dissipative or non-
conservative systems. Such systems exhibit asymp-
rThis system has two fixed points of the form x1,2 = totic stability; the more negative the exponent, the
µ−1 greater the stability. Superstable fixed points and su-
± , meaning that there is no single value to display perstable periodic points have a Lyapunov exponent of

the whole bifurcation diagram. Figure 7 displays the two δ = −∞.

326
Figure 10. Cobweb plot of the logistic map
for different values of the parameter: (top-
bottom, left-right): λ = 2.5; λ = 3.235; λ =
3.57; λ = 4

initial point for the system variable and the interval for the
system parameter. Other settings such as the number of it-
erations used for the calculation are internally handled at
the core of the numerical toolbox, according to a prescribed
error, which can be changed at will. Figure 9(top) depictes
the Lyapunov exponent of the logistic map for the system
parameter on the interval (0, 4). Comparison of this picture
with Figure 3 shows that negative values for the LE are an
indication of regular behavior, while positive values are as-
sociated with chaotic motion. Note also that the LE tends
Figure 9. (top) Lyapunov exponent of the lo- to −∞ for λ = 2 meaning the existence of a superstable
gistic map on the interval (0, 4); (bottom) fixed point. Figure 9(bottom) shows a magnification of the
close up on the interval (3.2, 4) figure on top for the interval (3.2, 4). We can see that the
LE is positive for λ > λc except by the existence of some
periodic windows, thus explaining very well the phenomena
previously observed in the bifurcation diagram.
• δ = 0: The orbit is a neutral fixed point (or an even-
tually fixed point). A Lyapunov exponent of zero in- 3.4. Cobweb Plot
dicates that the system is in some sort of steady state
mode. A physical system with this exponent is conser-
A cobweb plot is a graphical procedure especially suited
vative. Such systems exhibit Lyapunov stability.
to analyze the qualitative behaviour of one-dimensional iter-
• δ > 0: The orbit is unstable and chaotic. Nearby ated functions. Cobweb plots are useful because they allow
points, no matter how close, will diverge to any arbi- to determine the long-term evolution of an initial condition
trary separation. All neighborhoods in the phase space under repeated application of a map. Algorithmically, the
will eventually be visited. These points are said to be plot consists of a diagonal line y = x and a curve represent-
unstable. Although the system is deterministic, there ing the map y = f (x). To visualize the behaviour of the
is no order to the orbit that ensues. system from an initial point x0 for a given parameter value,
the following steps are to be applied:
Our program allows us to compute the Lyapunov expo-
nents in a very easy way; the initial input is given by the 1. Find the point on the function curve with an x-

327
Figure 11. Cobweb plot of the Singer map for
Figure 12. (left-right) 1D, 2D and 3D phase
µ=1
space graph of the systems: (top) logistic
map for λ = 4; (middle) cubic map for λ = 4;
(bottom) Singer map for µ = 1.07
coordinate of x0 . This leads to the coordinates
(x0 , f (x0 )).
2. Plot horizontally across from this point to the diagonal 3.5. Phase Space Graph
line. This yields the coordinates (f (x0 ), f (x0 )).
A very powerful strategy to analyze chaotic systems is
3. Plot vertically from the point on the diagonal to the
to use the so-called phase space graph. By this we mean
function curve. The coordinates (f (x0 ), f (f (x0 ))) are
a collection of pictures associated with the orbit of a given
hereby returned.
point x0 and embedded into different n-dimensional spaces.
4. Repeat from step 2 until reaching a prescribed amount For n = 1 we get the signal of the orbit, i.e. the sequence
of iterations. of iterates {xn }n over the time. Such sequence is usually
called the time series. For n = 2 the graph is obtained by
Figure 10 shows a matrix of four cobweb plots for the representing the sequence of iterates {xn+1 } vs. {xn }, for
logistic map and as many initial conditions. The first one, n = 3, we represent the sequence of iterates {xn+2 } vs.
for λ = 2.5 (top-left), displays an inward spiral leading to a {xn+1 } and {xn } and so on.
stable fixed point (conversely, an unstable fixed point corre- Figure 12 uses the phase space graph to analyze the
sponds to an outward spiral). A period-2 orbit is represented chaotic behavior of three discrete systems: the logistic map
by a rectangle, like in Fig. 10 (top-right) with λ = 3.235. (1) for λ = 4 (top), the cubic map (4) for λ = 4 (mid-
Greater period cycles produce further, more complex closed dle) and the Singer map (5), for µ = 1.07 (bottom). Left
loops. A chaotic orbit would show a “filled out” area, indi- column in Figure 12 shows the corresponding time series.
cating an infinite number of non-repeating values. Figs. 10 These pictures illustrate perfectly how the chaotic behavior
(bottom) show two cases of chaotic orbits, for λ = 3.57 looks like.
(left) λ = 4 (right). Note that the filled area is larger in the As shown in Section 3.3, a characteristic of chaos is that
second case, as shown in Fig. 4, where the filled vertical chaotic systems exhibit a great sensitivity to initial condi-
line of the diagram increases on the interval (0, 1) as the tions. A common source of such sensitivity to initial con-
parameter goes to 4 from the left. ditions is that the map represents a repeated folding and
Figure 11 depicts the cobweb plot of the Singer map for stretching of the space on which it is defined. For instance,
µ = 1. At this parameter value, the system stabilizes in a for the logistic map, the n = 2 phase space graph, rep-
periodic orbit (see also Fig. 8). However, the behavior of resented in Fig. 12(top,middle), gives a two-dimensional
this system strongly relies on our choice of the initial point phase diagram of the logistic map showing the quadratic
for the system variable, so other periodic behaviors for the curve of the difference equation (1). Similarly, central col-
same system parameter can also be obtained. umn of Fig. 12 displays the structure of the cubic curve and

328
the quartic curve in middle and bottom rows, respectively.
We can also embed the same sequence in a three- Acknowledgments
dimensional phase space, in order to investigate the deeper This research has been supported by the Computer Science
structure of the map. Fig. 12(top,right), demonstrates this, National Program of the Spanish Ministry of Education and
showing how initially nearby points begin to diverge, partic- Science, Project Ref. #TIN2006-13615 and the University
ularly in those regions corresponding to the steeper sections of Cantabria.
of the plot. Note also the complicated structure of the cu-
bic and Singer maps, given by the right column of Fig. 12,
middle and bottom respectively. References

4. Conclusions and Further Remarks [1] A. Dhooge, W. Govaerts and Y.A. Kuznetsov, “Matcont: A
Matlab package for numerical bifurcation analysis of ODEs”.
ACM Transactions on Mathematical Software, 29(2) (2003)
This paper focuses on the problem of analyzing one- pp. 141-164.
dimensional discrete dynamical systems and their wealth
of different chaotic behaviors. In this work, a new Matlab [2] A. Dhooge, W. Govaerts and Y.A. Kuznetsov, “Numerical
toolbox and GUI have been introduced, aimed at providing continuation of fold bifurcations of limit cycles in MAT-
end-users with a powerful computer tool to perform such CONT”. Proceedings of CASA’2003. Lecture Notes in Com-
analysis. This tool has proven to be extremely useful in puter Science, 2657 (2003) pp. 701-710.
order to carry out a number of computations (fixed points,
[3] A. Gálvez, “Numerical-symbolic Matlab program for the
their stability, bifurcation diagrams, Lyapunov exponents,
analysis of three-dimensional chaotic systems”. Proceedings
cobweb plots, phase space graphs for various dimensions) of CASA’2007. Lectures Notes in Computer Science, 4488
in an efficient and simple way. (2007) pp. 211-218.
As above-mentioned, one practical application of 1D-
maps is that they exhibit most of the classical phenomena [4] A. Gálvez and A. Iglesias, “Symbolic/numeric analysis of
associated with chaos. The advantage of using discrete chaotic synchronization with a CAS”. Future Generation
maps is that they don’t require the sophisticated mathemat- Computer Systems, 25(5) (2007) pp. 727-733.
ical machinery needed for continuous systems. In fact, dis-
crete systems are often used as a simplification of chaotic [5] W. Govaerts and B. Sautois, “Phase response curves, delays
and synchronization in Matlab”. Proceedings of CASA’2006.
flows via the so-called Poincaré map, given by the intersec-
Lectures Notes in Computer Science, 3992 (2006) pp. 391-
tion of a periodic orbit in the state space of a continuous dy-
398.
namical system with a certain lower dimensional subspace,
called the Poincaré section, transversal to the flow of the [6] J.M. Gutiérrez, A. Iglesias, J. Guémez and M.A. Matı́as,
system. More precisely, one considers a periodic orbit with “Suppression of chaos through changes in the system vari-
initial conditions on the Poincaré section and observes the ables through Poincaré and Lorenz return maps”. Interna-
point at which this orbits first returns to the section, thus tional Journal of Bifurcation and Chaos, 6 (1996) pp. 1351-
the name first recurrence map. The transversality of the 1362.
Poincaré section basically means that periodic orbits start-
ing on the subspace flow through it and not parallel to it. [7] J.M. Gutiérrez and A. Iglesias, “A Mathematica package for
the analysis and control of chaos in nonlinear systems”. Com-
Such Poincaré map can be interpreted as a discrete
puters in Physics, 12(6) (1998) pp. 608-619.
dynamical system with a state space that is one dimen-
sion smaller than the original continuous dynamical sys- [8] A. Iglesias, J.M. Gutiérrez, J. Guémez and M.A. Matı́as,
tem. Because it preserves many properties of periodic and “Chaos suppression through changes in the system variables
quasiperiodic orbits of the original system and has a lower and numerical rounding errors”. Chaos, Solitons and Frac-
dimensional state space it is often used for analyzing the tals, 7(8) (1996) pp. 1305-1316.
original system. From this point of view, the present pro-
gram can also be applied to analyze continuous systems, [9] A. Iglesias, “A new scheme based on semiconductor lasers
and might be used in combination with a previous program with phase-conjugate feedback for cryptographic communi-
cations”. Lectures Notes in Computer Science, 2510 (2002)
reported in [3] and developed for continuous 3D flows.
pp. 135-144.
Future works include the extension of this program to
the case of muldimensional discrete systems, and the im- [10] A. Iglesias and A. Gálvez, “Analyzing the synchronization
plementation of especialized routines for the control of of chaotic dynamical systems with Mathematica: Part I”. Pro-
chaos [6, 7, 8] and the synchronization of chaotic systems ceedings of CASA’2005. Lectures Notes in Computer Science,
[9, 10, 11] for discrete maps through the Poincaré section. 3482 (2005) pp. 472-481.

329
[11] A. Iglesias and A. Gálvez, “Analyzing the synchronization of
chaotic dynamical systems with Mathematica: Part II”. Pro-
ceedings of CASA’2005. Lectures Notes in Computer Science,
3482 (2005) pp. 482-491.

[12] A. Iglesias and A. Gálvez, “Revisiting some control schemes


for chaotic synchronization with Mathematica”. Lectures
Notes in Computer Science, 3516 (2005) pp. 651-658.

[13] H.A. Lauwerier, “One-dimensional iterative maps”. In:


Chaos. Holden, A.V. (ed.). Manchester University Press,
Manchester (1986).

[14] The Mathworks Inc: Using Matlab. Natick, MA (1999).

[15] R.M. May, “Simple mathematical models with very compli-


cated dynamics”. Nature, 261 (1976) pp. 459-467.

[16] E. Ott, Chaos in Dynamical Systems. Cambridge University


Press, Cambridge (1993).

[17] H. Sarafian, “A closed form solution of the run-time of a


sliding bead along a freely hanging slinky”. Proceedings of
CASA’2004. Lecture Notes in Computer Science, 3039 (2004)
pp. 319-326.

[18] W. Zhou, D.J. Jeffrey and G.J. Reid, “An algebraic method
for analyzing open-loop dynamic systems”. Proceedings of
CASA’2005. Lecture Notes in Computer Science, 3516 (2005)
pp. 586-593.

330

You might also like