Professional Documents
Culture Documents
Textbook Analysis of Chaotic Behavior in Non Linear Dynamical Systems Michal Piorek Ebook All Chapter PDF
Textbook Analysis of Chaotic Behavior in Non Linear Dynamical Systems Michal Piorek Ebook All Chapter PDF
https://textbookfull.com/product/non-linear-differential-
equations-and-dynamical-systems-1st-edition-luis-manuel-braga-da-
costa-campos/
https://textbookfull.com/product/intelligent-algorithms-for-
analysis-and-control-of-dynamical-systems-rajesh-kumar/
https://textbookfull.com/product/dynamical-systems-in-population-
biology-xiao-qiang-zhao/
https://textbookfull.com/product/dynamical-systems-in-population-
biology-second-edition-zhao/
Non-cooperative Stochastic Differential Game Theory of
Generalized Markov Jump Linear Systems 1st Edition
Cheng-Ke Zhang
https://textbookfull.com/product/non-cooperative-stochastic-
differential-game-theory-of-generalized-markov-jump-linear-
systems-1st-edition-cheng-ke-zhang/
https://textbookfull.com/product/spaces-of-dynamical-systems-2nd-
edition-pilyugin-s-yu/
https://textbookfull.com/product/spaces-of-dynamical-systems-2nd-
edition-sergei-yu-pilyugin/
https://textbookfull.com/product/analysis-and-data-based-
reconstruction-of-complex-nonlinear-dynamical-systems-using-the-
methods-of-stochastic-processes-m-reza-rahimi-tabar/
https://textbookfull.com/product/linear-systems-signal-
processing-and-hypercomplex-analysis-chapman-university-
november-2017-daniel-alpay/
Studies in Systems, Decision and Control 160
Michał Piórek
Analysis of Chaotic
Behavior in Non-
linear Dynamical
Systems
Models and Algorithms for Quaternions
Studies in Systems, Decision and Control
Volume 160
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
The series “Studies in Systems, Decision and Control” (SSDC) covers both new
developments and advances, as well as the state of the art, in the various areas of
broadly perceived systems, decision making and control-quickly, up to date and
with a high quality. The intent is to cover the theory, applications, and perspectives
on the state of the art and future developments relevant to systems, decision
making, control, complex processes and related areas, as embedded in the fields of
engineering, computer science, physics, economics, social and life sciences, as well
as the paradigms and methodologies behind them. The series contains monographs,
textbooks, lecture notes and edited volumes in systems, decision making and
control spanning the areas of Cyber-Physical Systems, Autonomous Systems,
Sensor Networks, Control Systems, Energy Systems, Automotive Systems,
Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace
Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power
Systems, Robotics, Social Systems, Economic Systems and other. Of particular
value to both the contributors and the readership are the short publication timeframe
and the world-wide distribution and exposure which enable both a wide and rapid
dissemination of research output.
123
Michał Piórek
Department of Computer Engineering
Wrocław University of Science
and Technology
Wrocław, Poland
This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To my Wife and my Family
Acknowledgements
Many people have contributed to this work. At first, I would like to express my
deepest gratitude to Prof. Ewa Skubalska-Rafajłowicz and Dr. Bartosz Jabłoski for
the countless support during the study on chaotic systems and quaternions. Their
guidance helped me in all the time of research and writing of this book.
I would like to express my appreciation to Prof. Józef Korbicz and Prof. Andrzej
Polański for the comments, which helped me to increase the value of this work.
I would also like to thank Prof. Ewaryst Rafajłowicz, Prof. Tyll Krueger, and Dr.
Łukasz Korus for very important advice and help during the research.
My sincere thanks also go to Prof. Konrad Wojciechowski, Dr. Henryk Josiński,
Dr. Adam Świtoński, and colleagues from Silesian University of Technology, who
provided me an opportunity to work with their team, and who gave access to the
great gait recordings data set. Without their precious support, it would not be
possible to conduct this research. Human motion data used in the experiments
presented in this book was obtained from the Centre for Research and Development
of Polish-Japanease Academy of Information Technology (www.bytom.pja.edu.pl).
A special thanks to my family, my wife for her patience and indulgence, my
parents, siblings, and friends for supporting me spiritually. I hope we will have
more time for each other now.
vii
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 1
1.1 Problem Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 3
1.2 Proposed Methods and Algorithms, Carried Out Numerical
Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Novelties Presented in This Work . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Motivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Book Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Processes Described by Quaternion Models . . . . . . . . . . . . . . . . . . 7
2.1 Definition of the Processes Described by Quaternion Models . . . 7
2.2 Elements of Quaternion’s Algebra . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Quaternions Visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 Quaternions Averaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 Quaternions Random Generation . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Advantages and Disadvantages of Using Quaternions
Parametrization of Rotation . . . . . . . . . . . . . . . . . . . . . . . . . ... 16
3 Deterministic Chaos Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1 Dynamical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Chaos Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2.1 Positive Entropies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2.2 Strong Sensitivity to Initial Conditions . . . . . . . . . . . . . . 26
3.2.3 Strange Attractor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.4 Non-integer Fractal Dimension of the Attractor . . . . . . . . 29
3.3 Analysis of Chaos Basing on a Time Series . . . . . . . . . . . . . . . . 30
3.4 Time Delay Embedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4 Analysis of Chaos from Time Series - Existing Methods Survey . . . 35
4.1 Time Delay . . . . . . . . . . . . . . . . . . . . . . . ................ 35
4.2 Embedding Dimension . . . . . . . . . . . . . . . ................ 37
4.3 Reconstruction of the Phase Space . . . . . . . ................ 38
ix
x Contents
The main problem elaborated in this book can be defined as: Propose methods and
algorithms of nonlinear analysis of processes described by quaternion models allow-
ing for the detection of deterministic chaos presence.
Having that methodology in place one may answer the question whether basing
on the quaternion time series created from observations of the dynamical system
output is it possible to detect the deterministic chaos in the system?
For the performed research and the proposed methods the below assumptions
have been made:
1. As processes defined by quaternion models we treat the processes, which are ruled
by dynamical systems evolution and their output can be recorded in quaternion
parametrization form, especially quaternion time series.
2. In the scope of the research are stationary, autonomous dynamical systems, with-
out forcing.
3. The model of the system is created by time delay embedding. We consider in this
work the rotations space (a subset of R n space), which satisfies Takens embedding
theorem.
The book depicts a 5 years of studies on the deterministic chaos in the real processes,
which can be described by quaternions. The results of presented research can be
divided into two kinds: methodology and experiments.
From methodological point one may list the following algorithms have been pro-
posed:
• Mutual information method for quaternion time series - method for estimation of
time delay.
• False nearest neighbors method for quaternion time series - method for estimation
embedding dimension.
• The largest Lyapunov exponent estimation algorithm for quaternion domain.
• Correlation dimension estimation algorithm for the attractor in quaternion domain.
• kMeans clustering algorithm for quaternion data sets.
• Quaternions clustering validity indexes.
From the experimental point, validity and properties of presented methods have
been addressed in the experiments:
• Experiment of usage of proposed methods to differentiate random, chaotic and
periodic quaternion time series.
4 1 Introduction
The research presented in this book can be considered as a new one in the field
because of several reasons:
• As per the literature research, in the related work there is no existing direction of
analysis of chaos in the processes described by quaternions.
• Provided methods allow not only for a detection of deterministic chaos in quater-
nion time series but also for the classification of the nature of dynamical system,
which drives the quaternion time series. Using the proposed methodology three
types of the quaternion time series can be identified: random, periodic and chaotic.
• Proposed methods have been used to establish an analysis of chaos in real process -
gait sequences. The results of analysis for wide input data set have been compared
against the ‘classic’ methods designed for scalar time series.
• The additional value of the book is brought by the proposed framework of quater-
nions clustering. The framework includes the proposition of quaternions clustering
algorithm along with the measures of the quaternions clustering validity.
• Quaternions clustering framework has been tested on generated test data sets. The
quality of quaternions clustering has been tested against three parameters of the
input data set: real clusters number, overlapping degree and variance of the data
set.
1.4 Motivations
The entry point of analysis of chaos in the processes described by quaternion models
was an analysis of human gait recordings. Initially work was directed on nonlinear
analysis of gait time series in the context of human gait stability. The Author had
a great opportunity to work on this matter with the research team from Silesian
University of Technology, the owner of the massive gait recordings set recorded
using motion capture technique by system Vicon.
First experiments were performed using classic methods of embedding for a uni-
variate time series. Having the full information of recorded rotation in time (e.g
Euler angles or unit quaternions), only one-dimensional subset of it was taken into
consideration e.g. time series consisted of rotations in one plane or quaternions angle
time series. At some point of time there has arisen a question whether the analysis of
the full rotation description will bring an advance? Hence the work on embedding of
1.4 Motivations 5
unit quaternions has started since the unit quaternion representation is an equivalent
of Euler angles one.
The problem tends to be more general motivation to answer the questions whether
embedding of multivariate time series can bring an advance against embedding on
univariate time series and whether deterministic chaos can occur in the processes
described by quaternions?
At the very beginning, to make the further reading, analysis and experiments clearer
we will focus on the definition of the processes described by quaternion models as
this is in the title of the book. The author would like to emphasize that any new
mathematical concept of quaternion processes is introduced here.
We assume the processes described by quaternion models are processes, which are
ruled by dynamical systems evolution and their output can be measured in quaternion
parametrization form, especially quaternion time series. Since that, it is crucial to
recall the theory of dynamical systems. This will be a subject of considerations of
next chapter.
The dynamical system concept includes three main components: phase space,
time and the time evolution law. A phase space X represents the possible states of
the system. Time can be a continuous variable or a discrete integer-valued variable.
The time evolution law, is a rule that allows us to determine the state of the system
at each moment of time t from its states at all previous times. An example of a
dynamical system is N first-order, autonomous, ordinary differential equations:
We will perform all observations and analysis of such kind of processes on a time
series recorded as samples measured in the output of the underlying system.
where s(n) is the value of nth sample of measurement of physical process x in time
t0 + nΔt and t0 is initial time.
In the presented research we will focus only on the processes, which are results of
those systems, for which the output can be recorded in quaternions parametrization
form. This is a novel approach and ultimately defines the scope of the processes
described by quaternion models. Hence, the main subject of further analysis will be
a unit quaternions time series representing the knowledge about the processes:
Q(n) = (q1 , q2 , . . . , q N )
(2.3)
= (w1 + i x1 + j y1 + kz 1 , . . . , w N + i x N + j y N + kz N )
where: w represents a real part and ρ = (x, y, z) is called a vector part (i, j and k
are equivalents of imaginary unit). The whole quaternion set is usually denoted as
H , which is due to it’s main founder Sir William Hamilton.
The details of quaternions algebra widely used in the parametrization of rotations
are well described in the related work (e.g. [5, 31, 56, 65, 84, 139] or [62]).
In the scope of our interests are unit quaternions which can parametrize the space
of rotations:
q = 1 (2.5)
→
where n ∈ R 3 is called an axis of the rotation and ϕ is called an angle of the rotation.
Quaternion conjugation is given as:
q∗
q −1 = (2.9)
q2
where reciprocal is defined for each q ∈ H \[0, (0, 0, 0)]. Quaternion multiplication
is defined as:
Operation of the quaternion’s logarithm, which is a setting the tangent space in the
point [1, (0, 0, 0)], is denoted by:
ϕ
logq = 0, n (2.11)
2
where n < π .
Let us suppose that r = (x, y, z) is a point in 3D space and qr represents the same
vector in quaternions form:
→
qr = [0, r ] = 0 + i x + j y + kz (2.12)
where q defines rotation angle and axis, as it was stated in Eq. 2.7. The whole
procedure is depicted by the concept diagram in the Fig. 2.1.
All time series considered in this book are unit quaternions time series and con-
cern of parametrization of rotations. All 3D rotations form S O(3), which is special
orthogonal group. Quaternions can be considered as one of parametrization of this
group.
The algorithms proposed in this work in most cases employ the distance function
between the two rotations parametrized by unit quaternion. Hence, it is important
to have well defined distance function (metric) between two elements of a data set.
Any distance function d used for handling rotation data, need to satisfy both usual
axioms for metrics and S O(3) properties [61]:
• d(R1 , R2 ) = 0 ↔ R1 = R2
• d(R1 , R2 ) = d(R2 , R1 )
• d(R1 , R3 ) ≤ d(R1 , R2 ) + d(R2 , R3 )
• function d respects the topology of S O(3)
• function d is left/right invariant or bi-invariant
where function d(R1 , R2 ) we call bi-invariant if d(R1 ∗ R2 , R1 ∗ R3 ) = d(R2 , R3 )
and d(R2 ∗ R1 , R3 ∗ R1 ) = d(R2 , R3 ).
A good survey of available quaternions distance functions for S O(3) can be found
in [61]. All of them are gathered in the Table 2.1.
The distance measure commonly used for the purposes of this research is d6
(functionally equivalent to d3 ). It has all required properties (metric axioms, bi-
variance, respecting S O(3) topology). It was used in some of the related papers [63,
64] and provided very good results in practical applications. The physical interpreta-
tion of this measure is a length of geodesic curve on the quaternion unit sphere. It can
be also considered as the amount of energy or rotation needed to rotate quaternion
q1 to the rotation defined by quaternion q2 .
2.3 Quaternions Visualization 11
While designing the algorithms for analysis of chaos in the quaternions domain, it
is very beneficial to visualize their output to see the physical results of work. Proper
quaternions time series visualization can increase the understanding and intuition
about the investigated matter.
There are a few quaternion’s visualization techniques, they are presented in
[56, 57] or [68]. In the case of this book, the subject of analysis are rotational data
so the visualization approach should highlight the rotation direction instead of just
representing the unit quaternion. In the light of this, the author sees the visualiza-
tion proposed by Dr. Bartosz Jabłoński in his works [65] and [67] as a best choice.
The Author would like to thank Dr. Bartosz Jabłoński for the access to his software
implementing mentioned visualization approach and possibility to plot quaternions
data figures presented in this book.
The selected visualization is common for all types of parametrization for rotational
space, it can present the rotations written in unit quaternions form as well as Euler
angles and rotation matrix. The data originally recorded as Euler angles are presented
using selected method in the Fig. 2.2.
The selected concept utilizes directed arrows representing data items. For each
data point the placement and orientation of the corresponding arrow describes the
rotation. Each data point is visualized as a series of three consecutive rotations around
axes (x, y, z).
In the Fig. 2.2 the gray arrow represents reference (base) orientation. The red
arrow depicts the rotation of base orientation around X ax. Similarly green and blue
arrows depict the analogical rotations of basis point around Y and Z axes. A bit more
complex rotation is depicted by the yellow element.
In the case of presence a massive data set it might be beneficial to replace the
arrows with dots for the better readability.
Very crucial aspect of the methodology proposed in this research is also a selection
of quaternion averaging method employed in algorithms. Apart from selection, very
important is to be aware of advantages and disadvantages of used method. A good
survey of the available methods is presented in [93]. The motivation of Markley’s
paper was to propose the optimal weighted unit quaternions average. In case of the
algorithms presented in this work it is sufficient to have non-weighted average but
having the weighted one algorithm is more than comfortable since one can always
set all the weights to one.
The simplest procedure of estimation weighted quaternions average can be defined
as: n −1 n
q simple = wi wi qi (2.14)
i=1 i=1
This approach has two consequences. The first one is the estimated average is not a
unit quaternion, however this can be easily fixed by dividing quaternion by its norm.
The second is more important: q and −q represent the same rotation and hence that
there is 2:1 mapping over quaternion group. In result changing the sign of any qi
shouldn’t change the average value but Eq. 2.14 does not hold this property.
As it was found out in the paper [100], the problem of averaging of quaternions
can be solved with help of attitude matrices description. The attitude matrix is related
to the quaternion by the formula:
A(q) = w2 − ρ2 I3X 3 + 2ρρ T − 2w[ρ X ] (2.15)
where ρ and w are a vector and scalar parts of a quaternion (Eq. 2.4), I3 is a 3X 3
identity matrix and [ρ X ] is a cross product matrix defined by
⎡ ⎤
0 −z y
[ρ X ] = ⎣ z 0 −x ⎦ (2.16)
−y x 0
2.4 Quaternions Averaging 13
n
q Mar kley = arg min wi A(q) − A(qi )2F (2.17)
q∈S3 i=1
Q example = {(0, 0, 30); (0, 0, 60); (−30, 0, 60); (30, 0, 60); (0, 0, 90);
(2.18)
(0, 0, −30); (0, 0, −60); (−30, 0, −60); (30, 0, −60); (0, 0, −90)}
For this short quaternion data set two averages have been calculated and com-
pared in the Table 2.2. First average was computed using Q simple algorithm and then
normalized. The second average was computed using Q Mar kley . The example time
series along with computed average rotations are also presented on the unit quater-
nions sphere in the Fig. 2.3. All rotations form the sample set are marked as violet
when the both averages are marked as pink and light blue.
From this simple comparison one may see the quaternion averages calculated
using two algorithms represent the same rotation: both occurred on the north pole
of unit quaternion sphere. The averages are equal to the value but with different
sign. Moreover, the Markley’s algorithm seems to provide the average with some
numerical error, which is not relevant in the presented test case but can result with
higher numerical discrepancies in further quaternions processing (e.g. when the
quaternions time series is longer or the average is being computed multiple times
in one algorithm). This numerical error can be resolved during normalization of
algebraic mean. For the purposes of the algorithms presented in the book the simple
algorithm has been used. However, the author sees the testing of the application of
different quaternions averages against its performance can be a interesting subject
for the further research.
distribution of variance 0.005 with defined quaternion mean. The clusters centers
set Q 5C (in Euler angles form) is defined as:
Q 5C = {(0, 0, 30); (0, 0, 60); (−30, 0, 60); (30, 0, 60); (0, 0, 90); } (2.21)
In case of some further presented algorithms there is also a purpose of the exper-
iment against a different degree of randomly generated data overlapping. As an
overlapping degree we will consider the distance between two adjacent clusters cen-
ters on the unit hypersphere. For example, in the data randomly generated around
two clusters, the sample procedure of changing the amount of overlapping in the data
set can be defined as following:
Q 10dist = {(0, 0, 30); (0, 0, 40)}
Q 20dist = {(0, 0, 30); (0, 0, 50)}
Q 30dist = {(0, 0, 30); (0, 0, 60)}
Q 40dist = {(0, 0, 30); (0, 0, 70)}
Q 50dist = {(0, 0, 30); (0, 0, 80)}
Q 60dist = {(0, 0, 30); (0, 0, 90)}
Q 70dist = {(0, 0, 30); (0, 0, 100)}
16 2 Processes Described by Quaternion Models
Fig. 2.6 Visualization of 3 datasets with variable amount of overlapping clusters. Left - Q 10dist ,
Middle - Q 40dist , Right - Q 70dist
Summarizing, the changing of the data overlapping degree between two(or more)
clusters is a increasing or decreasing of the value of one of its Euler’s angle, which
results in changing the distance between the same rotations represented on the
unit quaternion hypersphere. The example data sets with different data overlapping
amount are presented in the Fig. 2.6
All of the reviewed quaternions properties will be used in the proposed algorithms
designed to investigate chaos properties in a quaternion domain. The analysis of
chaos in rotational data will be performed on unit quaternions parametrization due
to their advantages against the other available parametrization.
Closing the chapter about processes described by quaternion models and quater-
nions properties it is worth to recall the main advantages and disadvantages of using
quaternions as a rotation representation.
The main merit of using quaternions parametrization of rotations is they do not
suffer from gimbal lock and coordinates singularities. This property is an outcome
from the Hairy Ball Theorem [124]. The main conclusion from this theorem is that,
no matter how a point is moving continuously around the sphere, there is no such area
where it will be moving in a direction, which is prohibited because of coordinates
singularity presence.
One may say that the representation of a rotation in a quaternion form is consistent
in the sense it requires only four numbers to describe a one rotation and only one
constraint - unit magnitude. This fact implies also a better computational efficiency
and stability than matrix of coordinates. Multiplying quaternions is less complex and
more computationally efficient than multiplying operation on matrices and provides
better numerical stability. It is also worth to underline that even if there occurred
2.6 Advantages and Disadvantages of Using Quaternions Parametrization of Rotation 17
the numerical drift away of unit magnitude, it can be resolved by renormalizing the
given quaternion [65].
In case of disadvantages of using quaternions representation of rotation, the main
one is that they are less intuitive in understanding and usually some additional oper-
ations are required to see the unit quaternions underlings rotation. Moreover, not all
mathematical operations can be implemented in the quaternions world. For example
the Euclidean vector space interpolation methods are not available and there is a need
of having the equivalents in the quaternions domain. One may say the drawback is
also a dual representation of rotation. Two quaternions q and −q represent exactly
same rotation as it was shown in the quaternions averaging section.
Chapter 3
Deterministic Chaos Properties
The deterministic chaos term and its properties is tightly connected with dynamical
systems theory. Good references of dynamical systems theory can be found in [75,
82, 101, 135, 143].
As it was well summarized in [75] there are a few approaches of modern dynamical
systems theory e.g. Ergodic theory, Topological dynamics, The theory of smooth
dynamical systems and Hamiltonian dynamics.
We assume that in the scope of this book are smooth dynamical systems. As it
was already mentioned in the previous chapter the dynamical system includes three
components:
• phase space
• time
• time evolution law
We assume the phase space represents all possible states of the system. Let’s assume
it is a compact Riemannian manifold M , a compact manifold with Riemannian metric
and sufficiently differentiable structure, which can be rewritten as bounded subset in
Rn . A state of the system at a given time, denoted as X (t), we call the single point in
the manifold M
X (t) = x(1) , x(2) , . . . , x(d ) (3.1)
where X (t) ∈ M .
By the time we mean a real value (t ∈ R) for the continuous time dynamical
systems or an integer (t ∈ Z) for discrete time systems. The time evolution law
describes how a particular point in manifold M moves in this manifold with time
passing.
The time evolution law is a rule, which describes where the system state is in M
after time t. This rule can be described e.g. by a flows or maps [33].
dX (t)
= F(X (t)) (3.2)
dt
where X (t) is a state vector and F : M → M is a vector field differentiable as often
as required.
Having the vector field F defined in Eq. 3.2 one may see time variable t does not
appear explicitly. This kind of systems not explicitly dependent on time are called
autonomous and they are the subject of this research. The analysis and embedding
of forced systems is also a subject of the related work e.g. [128–130].
Dynamical evolution of the system is an initial value problem, which resolution
determine what will happen with the initial system’s state X (0) after time t
Having the time evolution law F and initial conditions: X (0), one may resolve
the state of the system in the future time t > 0
During the evolution, the state of the system draws a path in the phase space called
trajectory or orbit. An example trajectory is presented in the Fig. 3.1.
For the autonomous dynamical system the solution of the initial value problem
of Eq. 3.3 exists and is unique once the vector field F is Lipshitz continuous [75].
The Lipschitz continuity can be defined as following. Let U be an open set in n .
A vector field F(X , t) on n is is said to be Lipschitz on U if does exist a constant
L such that
F(X1 , t) − F(X2 , t) ≤ L X1 − X2 (3.5)
is a state vector.
The behavior of the trajectory of dynamical system’s evolution depends both on
the form of F or M and on the initial condition.
It is worth to emphasize, that there exist some sets of initial conditions, which
leads to the same asymptotic behavior of trajectories. Such kind of a set of initial
conditions is called the basin of attraction. The attracting subset of the phase space
to which the trajectories tend to is called an attractor [72]. The examples of attractors
are fixed point attractor and limited cycle attractor presented in the Fig. 3.2.
The dynamical systems can be divided into two subgroups:
• Linear dynamical systems
• Nonlinear dynamical systems
Linear dynamical systems satisfy the superposition principle and they can be
described by the equations using linear operators. Nonlinear dynamical systems do
not satisfy the superposition principle and to describe them nonlinear operators are
used.
It is worth to underline that in practical applications very often the continuous time
dynamical system can be replaced by discrete time system by sampling in discrete
moments in time. It can be also done using Poincare section method ([96]), which
reduces the dimensionality of the system from N to N − 1.
Fig. 3.2 Example of dynamical systems attractors. Left - fixed point attractor, Right - limited cycle
attractor
Another random document with
no related content on Scribd:
alli no se la hauia dado, la qual
dezia en esta manera.
CARTA DE FLAMIANO Á
VASQUIRAN
Verdaderamente, Vasquiran, tus
cartas me desatinan porque
quando miro en ellas el
encarecimiento de tu daño me
parece grande, quando considero
la causa dél lo juzgo pequeño.
Pero en esta carta tuya postrera
he conocido en las cosas que me
escribes lo que te engañas, en
especial en quererte hazer ygual
en el martirio con Petrarca y
Garcisanchez. Si supiesses de
quan lexos vas errado,
maravillarte yas por cierto. Los
tiros de su combate muy lexos
hizieron los golpes de donde los
tuyos dan. De virgines y martires
ganaron ellos la palma si bien lo
miras, que no de confessores de
sus vitorias como tú hazes. Si
gozo ellos han hauido, en la
muerte lo habrian; que en la vida
nunca lo houieron. Mi dolor
sintieron e tu gozo ignoraron.
Claro está segun muestran las
liciones del uno e los sonetos del
otro, e quanto ambos escriuieron,
porque de ninguno dellos leemos
sino pesares en la vida, congoxas
y dolores en la muerte; desseos,
sospiros, ansias apassionadas,
cuydados e disfauores e
desesperados pensamientos;
quando quexando, quando
plañendo, quando pidiendo la
muerte, quando aborreciendo la
vida. Destos misterios dexaron
llenos de tinta sus papeles e de
lastimas su memoria, estos
hizieron sus vidas llenos de pena
e sus fines tan doloridos; con
estos que son los males do mis
males se engendran, con estos
que fueron martirizados como yo
lo soy; verdad es que de dias
vencieron como tú a quien de
amor y fe vencidos los tuvo e los
hizo viuir desseando la muerte
con mas razon que tú la desseas.
Assi que mira lo que por la boca
escriuiendo publicaron e
conoceras lo que en el alma
callando encubierto suffrieron, e
mira si hallarás en ellos vn dia de
victoria como tú plañes doze años
de gloria que dizes que perdiste.
Yo digo que los ganaste, mas
hate parecido a ti que la fortuna te
era obligada a tenerte queda la
rueda en la cumbre del plazer; yo
te prometo que si de sus bienes
no te houiera hecho tan contento,
que de sus males no fueras tan
quexoso sin razon, como estos e
yo lo somos. Tambien me
escriues como soñaste que viste
en vision tu alegria, tus placeres,
tu descanso, tu consentimiento, tu
esperança, tu memoria, tu
desseo; beato tú que primero las
gozaste en la vida y en la muerte
las ensueñas, yo te prometo que
avnque mi placer, ni mi alegria, ni
mi descanso, ni mi
contentamiento, ni mi esperança
yo los encontrasse a medio dia,
que no los conociesse pues que
nunca los vi; mi desseo y mi
memoria no me los cale soñar,
que velando me hazen soñar la
muerte sin dormir cada hora.
Tambien me escribes que viste á
Violina e te habló, e quexaste
dello, ¿qué te pudo hazer
viuiendo que muerta no te quiere
oluidar? No me alegraré yo de lo
que tú, que ni agora en vida ni
despues de mis dias acabados de
mi tuuo memoria ni terná, no digo
de verme que es impossible, mas
avn de pensar si soy en el mundo.
Contentate pues, recobra tu
juyzio, no des mas causa para
que las gentes te juzguen, no
corrompas la reputacion de tu
fama, ni el agudeza de tu ingenio
con tan flaca causa, dando lugar
a tu dolor que de pesar te haya de
tener tal que á ti pierdas e a mi no
ayudes, pues que vees que mi
vida penando se consume; sino te
voy a ver es por la necesidad que
tengo que a verme vengas. Lo
qual te pido que hagas tanto
caramente quanto rogartelo
puedo, porque avnque soledad
busques para tu descanso, la
compañia de mis sospiros te la
dará, e con la mucha confianza
que de ti tengo quedo con tu vista
esperando la respuesta glosando
esta cancion:
Aqui yaze
todo el bien que mal me haze.
No tardará la vitoria
de mi morir en llegar,
pues que yo vi este lugar
qu'era tan lleno de gloria
quanto agora de pesar.
Yo vi en toda esta riuera
mill arboles de alegria,
veola agora vazia
de plazer de tal manera
que me da la fantasia
qu'el dolor de su memoria
ya no dexará tardar
mi morir de no llegar
para darme tanta gloria
quanto m'a dado pesar.