1: \documentclass[11pt,english,reqno]{amsart}
2: \usepackage[T1]{fontenc}
3: \usepackage[latin1]{inputenc}
4: \pagestyle{plain}
5: \usepackage[dvips]{graphicx}
6: \usepackage{amssymb}
7: \usepackage{amsfonts}
8: \usepackage{float}
9: \usepackage{babel}
10:
11:
12: \makeatletter
13:
14:
15: \renewcommand{\baselinestretch}{1.67}
16: \oddsidemargin=.0cm \evensidemargin=-.0cm \headsep=3cm
17: \textwidth=16cm \topmargin=-2.0cm \textheight=20.5cm
18: \footskip=2.0cm
19:
20: %\oddsidemargin 0cm \evensidemargin -0.7cm \headsep=.8cm
21: %\textwidth=16cm \topmargin=-1.3cm \textheight=24.5cm
22: %\footskip=.9cm
23:
24: %\pagestyle{plain}
25:
26:
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
29: %%%%%
30: %%%%% AUTHORS OWN DEFINITIONS
31: %%%%%
32: %%%%% including e.g.
33: %%%%%
34:
35: \theoremstyle{plain}
36: \newtheorem{cor}{Corollary}
37: \newtheorem{thm}{Theorem}
38: \newtheorem{prop}{Proposition}
39: \newtheorem{lem}{Lema}
40: \newtheorem{Counterexample}{Counterexample}
41: \newtheorem{Proper}{Property}
42: \theoremstyle{remark}
43:
44: \newtheorem*{Remarks}{Remarks}
45:
46: \theoremstyle{definition}
47: \newtheorem{defi}{Definition}
48: \newtheorem{ex}{Example}
49: \newtheorem{Note}{Remark}
50:
51: \newcommand{\dos}{\ensuremath{2^X_{U}}}
52: \newcommand{\C}{\mathbb{C}}
53: \newcommand{\qz}{\mathbb{Q}}
54: \newcommand{\rz}{\mathbb{R}}
55: \newcommand{\rc}{\mathbb{K}}
56: \newcommand{\N}{\mathbb{N}}
57: \newcommand{\gz}{\mathbb{Z}}
58: \newcommand{\lae}{\lambda (E)}
59: \newcommand{\cs}{\text{ cs}}
60: \newcommand{\tpc}{\underset{\pi}{\hat{\otimes}}}
61: \newcommand{\tspc}{\underset{ s,\pi}{\hat{\otimes}}}
62:
63:
64: %%%%%
65: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
66: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
67:
68: \makeatother
69:
70: \begin{document}
71:
72: \title{THE UNIVERSAL CARDINAL ORDERING OF FIXED POINTS }
73:
74:
75: \author{Jesús San Martín$^{a,b}$, MŞ Jose Moscoso$^{a}$, A. González Gómez$^{c}$}
76:
77:
78: \keywords{cardinal ordering, reflection of a sequence, iterated reflection,
79: period doubling cascade, dynamical symbolic.}
80:
81: \maketitle
82: {\small $a$ Departamento de Matemática Aplicada, E.U.I.T.I. Universidad
83: Politécnica de Madrid. 28012-Madrid, SPAIN.}{\small \par}
84:
85: {\small $b$ Departamento de Física Matemática y de Fluidos, Facultad
86: de Ciencias. Universidad Nacional de Educación a Distancia. 28040-Madrid,
87: SPAIN.}{\small \par}
88:
89: {\small $c$ Departamento de Matemática Aplicada a los Recursos Naturales,
90: E.T. Superior de Ingenieros de Montes. Universidad Politécnica de
91: Madrid. 28040-Madrid, SPAIN.}{\small \par}
92:
93: {\small Corresponding author:} {\small Jesús San Martín. e-mail:
94: jsm@dfmf.uned.es}{\small \par}
95:
96: \begin{abstract}
97: We present the theorem which determines, by a permutation, the cardinal
98: ordering of fixed points for any orbit of a period doubling cascade.
99: The inverse permutation generates the orbit and the symbolic sequence
100: of the orbit is obtained as a corollary. The problem present in the
101: symbolic sequences is solved. There, repeated symbols appear, for
102: example, the R (right), which cannot be distinguished among them as
103: it is not known which R is the rightmost of them all. Therefore, there
104: is a lack of information about the dynamical system. Interestingly
105: enough, it is important to point that this theorem needs no previous
106: information about any other orbit.
107: \end{abstract}
108:
109:
110:
111: \section{Introduction}
112:
113: The period doubling cascade ~\cite{Fei1,Fei2} is one of the most
114: common phenomena in dynamical systems and appears as one of the most
115: diverse fields of investigation. As time goes on, it has been found
116: in physical \cite{lete}, chemical \cite{kiss}, and biological \cite{giacomo}
117: models, to name but a few. Despite it being ubiquitous and to be perhaps
118: the most important mechanism of transition to chaos, it has not been
119: completely characterized yet. Its importance and the lack of a complete
120: comprehension of this phenomenon influences negatively the adecuate
121: comprehension of all of the underlying phenomena, which as has been
122: said above, is extraordinarily widespread. A common way to characterize
123: an orbit of a period doubling cascade, generated by iteration of an
124: unimodal map with a critical point in C, is giving a dynamical sequence.
125: So, the 8-periodic orbit has the symbolic sequence orbit CRLRRRLR,
126: where R and L indicate, respectively, right and left with respect
127: to the critical point C. Immediately, a problem can be seen: some
128: R cannot be distinguished from others, and the same happens with the
129: L. The R situated in the fifth position of the symbolic sequence,
130: is more to the right or to the left than the R in the sixth position?
131: The only thing that is known is that some points are to the left side
132: and others to the right of C, but where exactly are those points?
133: Obviously, this lack of information makes it difficult or impossible
134: to perform any calculation. There is another problem, symbolic sequences
135: are obtained recursively from the sequence of a former orbit of the
136: period doubling cascade; therefore, the symbolic sequence of $2^{k}$-periodic
137: orbit ($k$ arbitrary) is not known, so it cannot be used to prove
138: new theorems in dynamical systems. In this paper we will provide solutions
139: to these problems.
140:
141: Let us place this problem in its historical context, in order to see
142: its origin, evolution and pending issues. The classical work of Metropolis,
143: Stein and Stein ~\cite{metro} set the big basement of the study
144: of dynamical systems, demonstrating the universal behavior that emerged
145: from the unimodal map. In that pioneering work, it was demonstrated
146: that the transformations of an interval in itself generate limit sets
147: that have an universal structure. That is to say, they are independent
148: from the function itself and the only important thing is the function
149: to be unimodal. The limit sets are characterized by finite sequences
150: of the symbols R and L, depending on wether the successive iterates
151: are plotted right or left of the critical point of the unimodal function.This
152: finite sequence of symbols, the symbolic sequence, is repeated periodically
153: and represents a periodic orbit of the dynamical system. The underlying
154: universal behavior forces the ordering of the different orbits with
155: respect to their parameter value and also depending of their period.
156: This universal ordering was called to the attention of Feigenbaum.
157: In his famous works ~\cite{Fei1,Fei2} he found that the periodic
158: orbits duplicate their period according to a universal rate, as a
159: universal law, independently of the iterated function. This is the
160: well known Feigenbaum cascade or period doubling cascade. What originally
161: was a universal behavior, reflected as a symbolic sequence, was converted
162: into a numerical universal behavior allowing the forecast of the successive
163: bifurcations. The numerical and predictive character has converted
164: it into a powerful tool that has extended to continuous dynamical
165: systems. To reach this point it is necessary to build a return map.
166: If the return map is unimodal, then the dyanamical system will probably
167: show the universal behavior forecasted by Feigenbaum, and it will
168: be possible to determine where the successive bifurcations will appear.
169: Feigenbaum work is condensed into the Feigenbaum-Çvitanovic equation.
170: He conjectured the existence of only one relevant eigenvalue for this
171: equation. This eigenvalue would be responsible for the universal behavior
172: found by Feigenbaum. Later, these hypothesis got a rigorous mathematical
173: proof \cite{col}. Within this body of knowledge, there are two relevant
174: problems closely related that emerge from the Feigenbaum cascade:
175:
176: i) Given the symbolic sequence of a given periodic orbit, it is possible
177: to get the symbolic sequence of the doubled-period orbit. This is
178: a recursive process that necessarily requires the knowledge of the
179: former element. Therefore, it is not possible to know directly the
180: symbolic sequence of an arbitrary orbit and it is not possible to
181: use it in mathematical proofs. This rises the question of how to directly
182: determine the symbolic sequence of an arbitrary orbit without calculating
183: the previous sequences.
184:
185: ii) Related with the former point, another question arises. The symbolic
186: sequence indicates that the successive iterates are to the right (R)
187: or to the left (L) of the critical point of the iterated unimodal
188: function. But, how far to the right or to the left are these found
189: in relation to the others? Which is their relative position? Or more
190: rigorously stated: should iterates be enumerated, what their cardinals
191: would be?
192:
193: First, we will obtain the solution to the second problem, obviously
194: much more complicated; but a simple corollary, which is easy and fast
195: to prove, will provide the solution to the first problem.
196:
197: Observe that the points of the $2^{k}$-periodic orbit are fixed points
198: of $f^{2^{k}}$. The stability and bifurcations of the fixed points
199: and the flow in their neighborhoods play an important role in dynamical
200: systems theory. Positions of fixed points of mappings are not only
201: useful in nonlinear dynamics, they are also important in quamtum mechanics
202: ~\cite{and} and therefore in all related subjects. Hence, we expect
203: the theorems we are going to prove will be of use in these subjects.
204:
205: This paper is organized as follows. First, the definitions and notations
206: necessary to prove the theorems will be introduced. Later, a first
207: result will be obtained that establishes the relative position of
208: the points of the $2^{k+1}$-periodic orbit in function of the relative
209: position of the points of the $2^{k}$-periodic orbit. This result
210: will be used to prove a theorem that determines the cardinal ordering
211: of the points in the $2^{k+1}$-periodic orbit in terms of the cardinal
212: position of the points in the $2^{k}$-periodic orbit. Based on these
213: preceding results, the final goal is obtained: the theorem that determines
214: the cardinal ordering of the points in the $2^{k}$-periodic orbit,
215: for an arbitrary $k$, with no previous information. Corollaries and
216: reformulations of the theorems will be stated along the process. Examples
217: will also be shown to understand the geometrical meaning of the results
218: and to ease the use of the theorems by scientists and engineers.
219:
220: \medskip{}
221:
222:
223:
224: \section{Definitions and notation}
225:
226: Let $f:\mathrm{I}\subset\rz\to\mathrm{I}$ be an unimodal map,
227: where $C$ denotes its critical point. Let $O=\{ x_{0},x_{1},\ldots,x_{q-1}\}$
228: be a $q$-periodic orbit of $f$, where $x_{0}=C$, then the first
229: and the second iterates of $C$ determine the subinterval J$=[f^{2}(C),f(C)]$
230: when $f$ has a maximum, (J$=[f(C),f^{2}(C)]$ if $f$ has a minimum)
231: such that $O\subset[f^{2}(C),f(C)]\subset\mathrm{I}.$
232:
233: Given that the orbit $O$ does not have the natural order within the
234: interval J$=[f^{2}(C),f(C)]$ we introduce the following definition.
235:
236: \begin{defi} The set $\{ C_{(1,q)}^{*},\, C_{(2,q)}^{*},\ldots,C_{(q,q)}^{*}\}$
237: will denote the ascending (descending) cardinality order of the orbit
238: $O=\{ C,f(C),\ldots,f^{q-1}(C)\}$ when $f$ has a minimum (maximum)
239: in $C$. Given that $C_{(i,q)}^{*}\ i=1,\ldots,q$, is the point of
240: the orbit $O$ plot in the cardinal position $'i'$ when this orbit
241: is cardinally ordered, the point $C_{(i,q)}^{*}$ is defined as the
242: $i-$th cardinal of the $q-$period orbit. \end{defi} \begin{Note}\label{not1}
243: Notice that $f(C)=C_{(1,q)}^{*}$ and $f^{2}(C)=C_{(q,q)}^{*}.$ Therefore,
244: if $f$ has a maximum in $C$, it results in
245:
246: \[
247: f^{2}(C)=C_{(q,q)}^{*}<\ldots<C_{(2,q)}^{*}<C_{(1,q)}^{*}=f(C);\]
248: meanwhile that if $f$ has a minimum in $C$ , it results in
249:
250: \[
251: f(C)=C_{(1,q)}^{*}<\ldots<C_{(2,q)}^{*}<C_{(q,q)}^{*}=f^{2}(C)\]
252: \end{Note}
253:
254: \begin{defi}\label{def2} The natural number ${\sigma}_{(i,q)},\ i=1,\ldots,q$
255: \, will denote the number of iterations of $f$ such as that $f^{\sigma_{(i,q)}}(C)=C_{(i,q)}^{*},\ i=1,\ldots,q$.
256: \end{defi}
257:
258: The goal of this paper is to determine the number of iterations, $\sigma_{(i,q)},$
259: of $f$ from $C$, to get the point of the $q$-periodic orbit ($q=2^{p}$)
260: situated in the $i-$th cardinal position $C_{(i,q)}^{*}$.
261:
262: \begin{defi} We denote as {\Large {${\sigma}_{q}$}} the permutation
263: {\Large $\sigma_{q}$}$=(\sigma_{(1,q)},\sigma_{(2,q)},\ldots,\sigma_{(q,q)})$,
264: that is the $q$-tuple formed by the ${\sigma}_{(i,q)}$. \end{defi}
265:
266: \begin{defi} Let $\{ a_{1},a_{2},\ldots,a_{n}\}$ be a sequence of
267: real numbers. We define the \emph{Reflection} of the sequence $\{ a_{1},a_{2},\ldots,a_{n}\}$
268: with increment $\alpha$, denoted as $R(a_{1},\cdots,a_{n};\alpha),$
269: as the sequence of real numbers given by:
270:
271: \[
272: R(a_{1},\cdots,a_{n};\alpha)=\{ a_{1},\cdots,a_{n},a_{n}+\alpha,a_{n-1}+\alpha,\ldots,a_{1}+\alpha\}.\]
273: \end{defi}
274:
275: \begin{defi}\label{def5} Let $\{ a_{1},a_{2},\ldots,a_{n}\}$ be
276: a sequence of real numbers. We define the \emph{Iterated Reflection}
277: of the sequence $\{ a_{1},a_{2},\ldots,a_{n}\}$ with the descending
278: increment $2^{k}$, $k\in\N$, and we denote it as $\partial R(a_{1},\cdots,a_{n};2^{k})$,
279: as the sequence of real numbers given by:
280:
281: \[
282: \partial R(a_{1},\cdots,a_{n};2^{k})=R(R(\ldots(R(R(a_{1},\cdots,a_{n});2^{k});2^{k-1})\ldots;2^{1});2^{0})\]
283: \end{defi} \begin{Proper}\label{pro} The Reflection verifies that\[
284: R(a_{1}+b_{1},a_{2}+b_{2},\ldots,a_{n}+b_{n};\alpha+\beta)=R(a_{1},a_{2},\ldots,a_{n}+;\alpha)+R(b_{1},b_{2},\ldots,+b_{n};\beta)\]
285: \end{Proper}
286:
287:
288: \section{Theorems}
289:
290: \begin{lem}
291: \label{odd} Let $f:\mathrm{I}\to\mathrm{I}$ be an unimodal map,
292: depending on a parameter, that undergoes a period-doubling cascade.
293: Let $\{ C,f(C)=C'\}$ be $2-$periodic superstable orbit of the cascade,
294: where $C$ is the critical point of $f$. Let $O'\equiv\{ C,y_{1},\cdots,y_{2^{p}-1}\}$
295: and $O\equiv\{ C,x_{1},\cdots,x_{2^{p+1}-1}\}$ be the $2^{p}-$periodic
296: and $2^{p+1}-$periodic superstable orbits of $f$, respectively.
297: Then
298: \begin{enumerate}
299: \item [(i)] The relative position of the points $O_{\text{odd}}\equiv\{ x_{1},x_{3},x_{5},\ldots x_{2^{p+1}-3},C'\}$
300: with respect to $C'$ is the same as the relative position of the
301: points $O'\equiv\{ C,y_{1},\ldots,y_{2^{p}-1}\}$ with respect to
302: $C$. The points are visited in the same order in both cases.
303: \item [(ii)] The relative position of the points $O_{\text{even}}\equiv\{ C,x_{2},x_{4},\ldots,x_{2^{p+1}-2}\}$
304: with respect to $C$ is the same as the relative postion of the points
305: of the orbit $O'\equiv\{ C,y_{1},\ldots,y_{2^{p}-1}\}$ with respect
306: to $C$ after being conjugated by homotecy with respect to $C.$
307: \end{enumerate}
308: \end{lem}
309: \begin{proof}
310: Let us suppose, without loss of generality, that $f$ has a maximum
311: at the critical point $C.$ Therefore, $f^{2}$ will have a minimum
312: at the critical point $C$ and a maximum at $C'.$ Consequently, there
313: is a neighborhood of $C,$ $\mathrm{I}_{C}$, and a neighborhood of
314: $C',$ $\mathrm{I}_{C'}$, such that $f_{|_{\mathrm{I}_{C}}}^{2}$
315: and $f_{|_{\mathrm{I}_{C'}}}^{2}$ are unimodals (see fig. \ref{fig:1a-and-1b}a).%
316: \begin{figure}
317: \begin{tabular}{cc}
318: \includegraphics[width=0.45\textwidth]{figuras/fig1a}&
319: \includegraphics[width=0.45\textwidth]{figuras/fig1b}\\
320: (a)&
321: (b)\\
322: \end{tabular}
323:
324:
325: \caption{\label{fig:1a-and-1b}(a) $f$ is an unimodal map, and $f^{2}$is
326: also unimodal in the neighbourhood of $I_{C}$ and $I_{C^{\prime}}$.
327: The graphs of $f^{2^{2}}$are reproduced in the neighbourhood of $I_{C}$
328: and $I_{C^{\prime}}$, being conjugated of each other. (b) Graph of
329: $f^{2^{2}}$.}
330: \end{figure}
331:
332:
333: As $O\equiv\{ C,x_{1},\ldots,x_{2^{p+1}-1}\}$ is the $2^{p+1}-$periodic
334: superstable orbit of $f$, it turns out that $O_{even}\equiv\{ C,x_{2},x_{4},\ldots,x_{2^{p+1}-2}\}$
335: and $O_{odd}\equiv\{ x_{1},x_{3},\ldots,x_{2^{p+1}-1}\}$ are the
336: $2^{p}-$periodic superstable orbits of $f_{|_{\mathrm{I}_{C}}}^{2}$
337: and $f_{|_{\mathrm{I}_{C'}}}^{2}$ respectively.
338:
339: As $f$ undergoes a period-doubling cascade so do $f_{|_{\mathrm{I}_{C}}}^{2}$
340: and $f_{|_{\mathrm{I}_{C'}}}^{2}$ and viceversa. Therefore, when
341: $f_{|_{\mathrm{I}_{C}}}^{2}$ goes through a bifurcation and it goes
342: from having a $2^{p-1}-$periodic superstable orbit to having a $2^{p}-$periodic
343: superstable orbit, then $f$ goes from having a $2^{p}-$periodic
344: superstable orbit to having a $2^{p+1}-$periodic superstable orbit.
345: Furthermore, $f_{|_{\mathrm{I}_{C'}}}^{2}$ and $f_{|_{\mathrm{I}_{C}}}^{2}$
346: undergo bifurcations simultaneously.
347:
348: It is well known that if two unimodal maps, with a maximum at their
349: critical points, undergo a period double cascade, then they will have
350: the same symbolic sequence. Meanwhile, if one of them has a maximum
351: and the other a minimum, those symbolic sequences will be conjugated
352: \cite{Gil}.
353:
354: Observe that as $f_{|_{\mathrm{I}_{C'}}}^{2}$ is unimodal at $\mathrm{I}_{C'}$,
355: if we take $x=x_{2^{p+1}-1}$ then \[
356: (f^{2}(x_{2^{p+1}-1}))'=(f(\underbrace{f(x_{2^{p+1}-1})}_{C}))'=f'(C)f'(x_{2^{p+1}-1})=0.\]
357: Therefore, $x_{2^{p+1}-1}$ is an extremum of $f_{|_{\mathrm{I}_{C'}}}^{2}$
358: and since $f_{|_{\mathrm{I}_{C'}}}^{2}$ is unimodal, and thus it
359: has one single extremum at $C'$, it results in $C'=x_{2^{p+1}-1}$.
360:
361: In line with the former discussion:
362: \begin{enumerate}
363: \item [(i)] follows because $f$ and $f_{|_{\mathrm{I}_{C'}}}^{2}$ are
364: unimodal, with maxima at $C$ and $C'$, respectively.
365: \item [(ii)] is also deduced since $f$ and $f_{|_{\mathrm{I}_{C}}}^{2}$
366: are unimodal with a maximum and minimum at $C$, respectively. The
367: conjugation of the orbit $O'\equiv\{ C,y_{1},\ldots,y_{2^{p}-1}\}$
368: with respect to $C$, is geometrically a homotecy of $O'$ with respect
369: to $C$; it is the equivalent to the homotecy of $O_{odd}$ with respect
370: to $C'$ and then a translation to $C$.
371:
372:
373: (See figures \ref{fig:1a-and-1b} and \ref{fig:2a-and-2b}).
374:
375: \end{enumerate}
376: \end{proof}
377: %
378: \begin{figure}
379: \begin{tabular}{cc}
380: \includegraphics[width=0.45\textwidth]{figuras/fig2a}&
381: \includegraphics[width=0.45\textwidth]{figuras/fig2b}\\
382: (a)&
383: (b)\\
384: \end{tabular}
385:
386:
387: \caption{\label{fig:2a-and-2b}(a) Cardinals of the points of the $2^{3}$-periodic
388: superstable orbit (b) $O_{even}$ reproduces $O_{odd}$ conjugated.
389: $O_{odd}$ is the same orbit observed in fig. \ref{fig:1a-and-1b}b.}
390: \end{figure}
391:
392:
393: \begin{lem}
394: \label{lem:tres-dos}Let $f:\mathrm{I}\to\mathrm{I}$ be an unimodal
395: map, depending on a parameter, that, as a function of the parameter,
396: goes through a period doubling cascade. If \[
397: C\,\mathrm{I}_{1}\,\mathrm{I}_{2}\ldots\mathrm{I}_{2^{p}-1}\quad\text{ where }\ \mathrm{I}_{i}=\mathrm{R}\,\text{ or }\,\mathrm{L}\quad i=1,\ldots,2^{p}-1\]
398: is the symbolic sequence of the $2^{p}-$periodic superstable orbit
399: of the cascade, then \[
400: C\,\mathrm{R}\,\overline{\mathrm{I}}_{1}\,\mathrm{R}\,\overline{\mathrm{I}}_{2}\,\mathrm{R}\ldots\mathrm{R}\,\overline{\mathrm{I}}_{2^{p}-1}\mathrm{R}\]
401: is the symbolic sequence of the $2^{p+1}-$periodic superstable orbit.
402: Where
403:
404: \[
405: \overline{\mathrm{I}}_{i}=\left\{ \begin{array}{ll}
406: \mathrm{R} & \hbox{if}\ \mathrm{I}_{i}=\mathrm{L}\\
407: \mathrm{L} & \hbox{if}\ \mathrm{I}_{i}=\mathrm{R}\end{array}\right.\qquad\, i=1,\ldots,2^{p}-1.\]
408:
409: \end{lem}
410: \begin{proof}
411: The points of the $2^{p+1}-$periodic superstable orbit $O\equiv\{ C,x_{1},\cdots,x_{2^{p+1}-1}\}$
412: can be split into the orbits $O_{\text{even}}\equiv\{ C,x_{2},\cdots,x_{2^{p+1}-2}\}$
413: and $O_{\text{odd}}\equiv\{ x_{1},\cdots,x_{2^{p+1}-1}\}$. Where,
414: $O_{\text{even}}$ is the $2^{p}-$periodic superstable orbit of $f_{|_{\mathrm{I}_{C}}}^{2}$
415: and $O_{\text{odd}}$ is the $2^{p}-$periodic superstable orbit of
416: $f_{|_{\mathrm{I}_{C'}}}^{2}.$
417:
418: By lemma, ~\ref{odd} the $2^{p}-$periodic superstable orbit of
419: $f,$ after being conjugated by homotecy with respect to $C$, is
420: transformed into $O_{\text{even}}$. As the homotecy with respect
421: to $C$ generates the conjugation $\mathrm{R}\longleftrightarrow\mathrm{L}$,
422: that is to say, it transforms $\mathrm{I}_{i}$ into $\overline{\mathrm{I}}_{i}\quad i=1,\ldots,2^{p}-1$,
423: and as the symbolic sequence of the $2^{p}-$periodic superstable
424: orbit is \[
425: C\,\mathrm{I}_{1}\,\mathrm{I}_{2}\ldots\mathrm{I}_{2^{p}-1},\]
426: it results that the symbolic sequence of $O_{\text{even}}$ is
427:
428: \[
429: C\,\mathrm{\overline{I}}_{1}\,\mathrm{\overline{I}}_{2}\ldots\mathrm{\overline{I}}_{2^{p}-1}.\]
430:
431:
432: On the other hand, all of the points of $O_{\text{odd}}$ are in the
433: neighborhood of $\mathrm{I}_{C'}$, situated to the right of $C$;
434: therefore, each time that a point of $O_{\text{odd}}$ is visited,
435: it will appear as a $\mathrm{R}$ in the sequence of $O$. Given that
436: the orbit of $O$ is built by alternating points of $O_{\text{odd}}$
437: and $O_{\text{even}},$ if an $\mathrm{R}$ is put after every symbol
438: of the symbolic sequence of $O_{\text{even}}$ , then the symbolic
439: sequence of $O$ will be obtained. That is to say, as the symbolic
440: sequence of $O_{\text{even}}$ is \[
441: C\,\mathrm{\overline{I}}_{1}\,\mathrm{\overline{I}}_{2}\ldots\mathrm{\overline{I}}_{2^{p}-1}\]
442: the sought for symbolic sequence results in \[
443: C\,\mathrm{R}\,\mathrm{\overline{I}}_{1}\,\mathrm{R}\,\mathrm{\overline{I}}_{2}\,\mathrm{R}\ldots\mathrm{R}\,\mathrm{\overline{I}}_{2^{p}-1}\,\mathrm{R}\]
444:
445: \end{proof}
446: As a trivial case, if the superstable orbit of period $1$, with symbolic
447: sequence $C$, is taken it results that the sequence of the $2$-periodic
448: superstable orbit is $C\,\mathrm{R}$ as it is well-known.
449:
450: \begin{ex}\label{pri1} While using the symbolic sequence of the
451: $2^{3}$-periodic superstable orbit, which is $C\,\mathrm{R}\,\mathrm{L}\,\mathrm{R}\,\mathrm{R}\,\mathrm{R}\,\mathrm{L}\,\mathrm{R},$
452: we are going to obtain the symbolic sequence of the $2^{4}$-periodic
453: superstable orbit using the technique of MSS \cite{metro} and later
454: the lemma ~\ref{lem:tres-dos}.
455:
456: \begin{enumerate}
457: \item [i)] To get the symbolic sequence of the $2^{4}$-periodic superstable
458: orbit according to MSS, the following steps must be taken:
459:
460: \begin{itemize}
461: \item [a)] To write consecutively twice the symbolic sequence: \[
462: C\,\mathrm{R\, L\, R\, R\, R\, L\, R}\, C\,\mathrm{R\, L\, R\, R\, R\, L\, R}\]
463:
464: \item [b)] Change the second $C$ to a $\mathrm{R}$ if the $\mathrm{R-}$parity
465: of the orginal orbit is even or $\mathrm{L}$ otherwise. As in this
466: case the $\mathrm{R-}$parity is odd we have to change the second
467: $C$ for a $\mathrm{L}$ , resulting in \[
468: C\,\mathrm{R\, L\, R\, R\, R\, L\, R\,\mathbf{L}\, R\, L\, R\, R\, R\, L\, R}\]
469:
470: \end{itemize}
471: \item [ii)] Using the Lemma~\ref{lem:tres-dos} to obtain the symbolic
472: sequence of the $2^{4}$-periodic superstable orbit, the following
473: steps must be taken:
474:
475: \begin{itemize}
476: \item [a)] Conjugate the symbolic sequence of the original orbit: \[
477: C\,\mathrm{R\, L\, R\, R\, R\, L\, R}\quad\Longrightarrow\quad C\,\mathrm{L\, R\, L\, L\, L\, R\, L}\]
478:
479: \item [b)] To add a $\mathrm{R}$ after every symbol of the conjugated
480: sequence
481:
482:
483: \[
484: C\,\mathrm{R\, L\, R\, R\, R\, L\, R\, L\, R\, L\, R\, R\, R\, L\, R}\]
485: obtaining the same results.
486:
487: \end{itemize}
488: \end{enumerate}
489: \end{ex}
490:
491: In what follows, we are going to use the Lemma ~\ref{odd} to determine
492: the cardinals of the points of the $2^{p+1}-$periodic superstable
493: orbit when the cardinals of the points of the $2^{p}-$periodic superstable
494: orbit are known.
495:
496: \begin{thm}
497: \label{orbit} Let $f:\mathrm{I}\to\mathrm{I}$ be an unimodal map,
498: depending on a parameter, undergoing a period doubling cascade. Let
499: $\{ C_{(1,2^{p})}^{*},\, C_{(2,2^{p})}^{*},\ldots,C_{(2^{p},2^{p})}^{*}\}$
500: be the cardinals of the points of the $2^{p}$-periodic superstable
501: orbit of the cascade, given by $f^{{\sigma}_{(i,2^{p})}}(C)=C_{(i,2^{p})}^{*}$
502: $,\quad i=1,\ldots,2^{p}.$ Then the permutation {\Large $\sigma_{2^{p+1}}$},
503: given by \[
504: \text{{\Large{$\sigma_{2^{p+1}}$}}}=(\sigma_{(1,2^{p+1})},\ldots,\sigma_{(2^{p+1},2^{p+1})})=R(2{\sigma}_{(i,2^{p})}-1,\ldots,2{\sigma}_{(2^{p},2^{p})}-1;1)\]
505: determines the cardinals $\{ C_{(1,2^{p+1})}^{*},\, C_{(2,2^{p+1})}^{*},\ldots,C_{(2^{p},2^{p+1})}^{*},\ldots,C_{(2^{p+1},2^{p+1})}^{*}\}$
506: of the points of the $2^{p+1}$-periodic superstable orbit by means
507: of $f^{\sigma_{(i,2^{p+1})}}(C)=C_{(i,2^{p+1})}^{*}$ $\quad i=1,\ldots,2^{p+1}.$
508: \end{thm}
509: \begin{proof}
510: Let $\{ C_{(1,2^{p+1})}^{*},\, C_{(2,2^{p+1})}^{*},\ldots,C_{(2^{p},2^{p+1})}^{*},C_{(2^{p}+1,2^{p+1})}^{*}\ldots,C_{(2^{p+1},2^{p+1})}^{*}\}$
511: be the cardinals of the $2^{p+1}$-periodic superstable orbit given
512: by $O\equiv\{ C,x_{1},x_{2},\ldots,x_{2^{p+1}-1}\}.$
513:
514: By Lemma ~\ref{odd}, the $2^{p}$ first cardinals of $O$, given
515: by $\{ C_{(1,2^{p+1})}^{*},\ldots,C_{(2^{p},2^{p+1})}^{*}\}$, are
516: the cardinals of the orbit $O_{\text{odd}}$ of $O$ (See figures
517: \ref{fig:2a-and-2b}).
518:
519: Furthermore, by Lemma ~\ref{odd}, they are visited in the same order
520: as the orbit of period $2^{p}$, with two nuances:
521: \begin{enumerate}
522: \item [(a)] the points of $O_{\text{odd}}$ are reached by means of $f^{2}$.
523: \item [(b)] the points of $O_{\text{odd}}$ are reached from the critical
524: point $C'$, and not from $C.$
525: \end{enumerate}
526: Therefore, this results in
527:
528: \[
529: (f^{2})^{{\sigma}_{(i,2^{p})}}(C')=f^{2{\sigma}_{(i,2^{p})}}(C')=C_{(i,2^{p+1})}^{*}\qquad i=1,\ldots,2^{p}\]
530:
531:
532: Furthermore, as
533:
534: \[
535: f^{2^{p+1}-1}(C)=C'\]
536: results in \[
537: f^{2{\sigma}_{(i,2^{p})}}(C')=f^{2{\sigma}_{(i,2^{p})}}(f^{2^{p+1}-1}(C))=f^{2{\sigma}_{(i,2^{p})}-1}(C)\]
538: that is \[
539: f^{2{\sigma}_{(i,2^{p})}-1}(C)=C_{(i,2^{p+1})}^{*}\qquad i=1,\ldots,2^{p}\]
540:
541:
542: On the other hand, the cardinals $\{ C_{(2^{p}+1,2^{p+1})}^{*},\, C_{(2^{p}+2,2^{p+1})}^{*}\ldots,C_{(2^{p+1},2^{p+1})}^{*}\}$
543: are the cardinals of $O_{\text{even}}$(See figures \ref{fig:2a-and-2b}).
544: By the Lemma ~\ref{odd}, as it was shown in the proof, $O_{\text{even}}$
545: is obtained applying a homotecy to $O_{\text{odd}}$ with respect
546: to ${C}'$ and then translating to $C$. Therefore, as \[
547: O_{\text{odd}}=\{ f^{2{\sigma}_{(1,2^{p})}}(C'),f^{2{\sigma}_{(2,2^{p})}}(C'),\ldots,f^{2{\sigma}_{(2^{p},2^{p})}}(C')\}\]
548: applying the homotecy with respect to ${C}'$ results in \[
549: O_{\text{even}}=\{ f^{2{\sigma}_{(2^{p},2^{p})}}(C'),f^{2{\sigma}_{(2^{p}-1,2^{p})}}(C')\ldots,f^{2{\sigma}_{(1,2^{p})}}(C')\}\]
550:
551:
552: After translation from $C'$ to $C$ results in
553:
554: \[
555: O_{\text{even}}=\{ f^{2{\sigma}_{(2^{p},2^{p})}}(C),f^{2{\sigma}_{(2^{p}-1,2^{p})}}(C)\ldots,f^{2{\sigma}_{(1,2^{p})}}(C)\}\]
556:
557:
558: Joining the cardinals of $O_{\text{odd}}$ and $O_{\text{even}}$
559: to retrieve the $2^{p+1}$-periodic orbit finally results in \begin{multline*}
560: \{ C_{(1,2^{p+1})}^{*},\ldots,C_{(2^{p},2^{p+1})}^{*},C_{(2^{p}+1,2^{p+1})}^{*},\ldots,C_{(2^{p+1},2^{p+1})}^{*}\}=\\
561: =\left\{ \, f^{2{\sigma}_{(1,2^{p})}-1}(C),\ldots,f^{2{\sigma}_{(2^{p},2^{p})}-1}(C),\, f^{2{\sigma}_{(2^{p},2^{p})}}(C),\ldots,f^{2{\sigma}_{(1,2^{p})}}(C)\right\} \end{multline*}
562: that is \[
563: f^{{\sigma}_{(i,2^{p+1})}}(C)=C_{(i,2^{p+1})}^{*}\,,\qquad i=1,\ldots,2^{p+1}\]
564: with \begin{multline*}
565: \text{\Large{$\sigma_{2^{p+1}}$}}=(\sigma_{(1,2^{p+1})},\ldots,\sigma_{(2^{p+1},2^{p+1})})=\\
566: =(2{\sigma}_{(1,2^{p})}-1,2{\sigma}_{(2,2^{p})}-1,\ldots,2{\sigma}_{(2^{p},2^{p})}-1,2{\sigma}_{(2^{p},2^{p})},\ldots,2{\sigma}_{(1,2^{p})})=\\
567: =R(2{\sigma}_{(1,2^{p})}-1,\ldots,2{\sigma}_{(2^{p},2^{p})}-1;1)\end{multline*}
568:
569: \end{proof}
570: \begin{ex}\label{prim} Observing the $2^{2}$-periodic orbit in
571: fig. \ref{fig:1a-and-1b}b results that \[
572: \sigma_{2^{2}}=\begin{pmatrix}1 & 2 & 3 & 4\\
573: 1 & 3 & 4 & 2\end{pmatrix}\ \begin{tabular}{c}
574: $\longleftarrow$ \text{indicates the position 'i' of $C_{{(i,2^{2})}}^{*}$} \\
575: $\longleftarrow\ \text{number of interations},\sigma_{(i,2^{2})},\text{to reach \mbox{$C_{{(i,2^{2})}}^{*}$}}$ \end{tabular}{\Longrightarrow}\ f^{\sigma_{(i,2^{2})}}(C)=C_{(i,2^{2})}^{*}.\]
576: Using the theorem \ref{orbit} and having in mind {\Large $\sigma_{2^{2}}$},
577: it is obtained
578:
579: \[
580: \sigma_{2^{3}}=\begin{pmatrix}1 & 2 & 3 & 4 & 5 & 6 & 7 & 8\\
581: 1 & 5 & 7 & 3 & 4 & 8 & 6 & 2\end{pmatrix}\ \begin{tabular}{c}
582: $\longleftarrow$ \text{ position 'i' of $C_{{(i,2^{3})}}^{*}$} \\
583: $\longleftarrow\ \text{ iterations},\sigma_{(i,2^{3})},\text{ to reach\mbox{\text{ $C_{{(i,2^{3})}}^{*}$}} }$ \end{tabular}{\Longrightarrow}\ f^{\sigma_{(i,2^{3})}}(C)=C_{(i,2^{3})}^{*}\]
584: that, it can be seen, is the permutation obtained from the ${2^{3}}$-periodic
585: orbit in fig{\Large .} \ref{fig:2a-and-2b}a. \end{ex}
586:
587: ~
588:
589: Finally, with the following theorem, we will reach our original goal
590: of determining how many iterations of $f$ over $C$ are necessary
591: to reach the point of the $2^{p}$-periodic orbit situated at the
592: cardinal position $'i'$, denoted by $C_{(i,2^{p})}^{*},\quad i=1,\ldots,2^{p}.$
593: The number of iterations needed to reach these points will be determined
594: by a permutation, whose inverse will determine the orbit and its symbolic
595: sequence.
596:
597: Notice that the orbit is arbitrary and no previous information of
598: any other orbit is needed, in contrast with the previous theorem where
599: the former orbit cardinals were needed.
600:
601: \begin{thm}
602: \label{fuerte} Let $f:\mathrm{I}\to\mathrm{I}$ be an unimodal function
603: undergoing a period doubling cascade and the $2^{p}$-periodic orbit
604: of the cascade with cardinals $\{ C_{(1,2^{p})}^{*},\ldots,C_{(2^{p},2^{p})}^{*}\}$.
605: Let the permutation {\Large $\sigma_{2^{p}}$}$=(\sigma_{(1,2^{p})},\,\sigma_{(2,2^{p})},\ldots,\sigma_{(2^{p},2^{p})})\qquad p\in\N,\quad p\geq2$
606: be determined by\[
607: \text{{\Large{$\sigma_{2^{p}}$}}}=\partial R(1,2^{p-1}+1;2^{p-2})\qquad p\in\N,\quad p\geq2,\]
608: then \[
609: f^{\sigma_{(i,2^{p})}}(C)=C_{(i,2^{p})}^{*}\qquad\, i=1,\ldots,2^{p}.\]
610:
611: \end{thm}
612: \begin{proof}
613: From remark \ref{not1} the $2-$periodic orbit satisfies \[
614: f^{1}(C)=C_{(1,2)}^{*}\quad\text{ and }\quad f^{2}(C)=C_{(2,2)}^{*}\]
615: Using the definition \ref{def2} it results in \[
616: C_{(1,2)}^{*}=f^{\sigma_{(1,2)}}(C)\quad\text{ and }\quad C_{(2,2)}^{*}=f^{\sigma_{(2,2)}}(C)\]
617: therefore, it results in \[
618: (\sigma_{(1,2)},\sigma_{(2,2)})=(1,2).\]
619: We are going to prove the formula by induction over $p$
620: \begin{itemize}
621: \item [(i)] Case $p=2.$ The orbit of period $2^{2}$ is $\{ C,f(C),f^{2}(C),f^{3}(C)\}\equiv\{ C,\, C_{(1,2^{2})}^{*}\, C_{(2^{2},2^{2})}^{*},f^{3}(C)\}$,
622: where remark \ref{not1} has been used. Since the orbit of period
623: $2$ is given by $(\sigma_{(1,2)},\sigma_{(2,2)})=(1,2)$, applying
624: the Theorem ~\ref{orbit} results in \begin{multline*}
625: \text{{\Large{$\sigma_{2^{2}}$}}}=(\sigma_{(1,2^{2})},\sigma_{(2,2^{2})},\sigma_{(3,2^{2})},\sigma_{(2^{2},2^{2})})=R(2\sigma_{(1,2)}-1,2\sigma_{(2,2)}-1;1)=(1,3,4,2)=\\
626: =\partial R(1,3;1)=\partial R(1,2^{2-1}+1;2^{0}).\end{multline*}
627:
628: \item [(ii)] Let us suppose that the theorem is valid for $p=k$ and check
629: wether this is true for $p=k+1,$ that is to say that \[
630: (\sigma_{(1,2^{k+1})},\sigma_{(2,2^{k+1})},\ldots,\sigma_{(2^{k+1},2^{k+1})})=\partial R(1,2^{k}+1;2^{k-1}).\]
631:
632:
633:
634: Using the theorem ~\ref{orbit} and the property \ref{pro} we have
635: \begin{multline}
636: (\sigma_{(1,2^{k+1})},\sigma_{(2,2^{k+1})},\ldots,\sigma_{(2^{k+1},2^{k+1})})=R(2\sigma_{(1,2^{k})}-1,\ldots,2\sigma_{(2^{k},2^{k})}-1;1)=\\
637: =2R(\sigma_{(1,2^{k})},\sigma_{(2,2^{k})},\ldots,\sigma_{(2^{k},2^{k})};0)+R(\underbrace{-1,\ldots,-1}_{2^{k}};1).\end{multline}
638: By the induction hypothesis, and using recursively the property \ref{pro},
639: the equality $(1)$ is transformed into \begin{multline}
640: (\sigma_{(1,2^{k+1})},\sigma_{(2,2^{k+1})},\ldots,\sigma_{(2^{k+1},2^{k+1})})=2R(\partial R(1,2^{k-1}+1;2^{k-2});0)+R(\underbrace{-1,\ldots,-1}_{2^{k}};1)=\\
641: =2R\underbrace{(R(\cdots R(R}_{k-1}(1,2^{k-1}+1;2^{k-2});2^{k-3})\cdots;1);0)+R(\underbrace{-1,\ldots,-1}_{2^{k}};1)=\\
642: =R\underbrace{(R(\cdots R(R}_{k-1}(2,2(2^{k-1}+1);2^{k-1});2^{k-2};)\cdots;2);0)+R(\underbrace{-1,\ldots,-1}_{2^{k}};1).\end{multline}
643: Using again recursively the property \ref{pro} the equality (2)
644: is rewritten as
645:
646: \begin{multline}
647: (\sigma_{(1,2^{k+1})},\sigma_{(2,2^{k+1})},\ldots,\sigma_{(2^{k+1},2^{k+1})})=\\
648: =R(\underbrace{R(\cdots R(R}_{k-1}(2,2(2^{k-1}+1);2^{k-1});2^{k-2};)\cdots;2);0)+R(\underbrace{R(\cdots R(R}_{k-1}(-1,-1;0);0)\cdots;0);1)=\\
649: =R(\underbrace{R(\cdots R(R}_{k-1}(2-1,2(2^{k-1}+1)-1;2^{k-1});2^{k-2};)\cdots;2);1)=\\
650: =\partial R(1,2^{k}+1;2^{k-1})\end{multline}
651:
652:
653: \end{itemize}
654: \end{proof}
655: \begin{ex}\label{seg} In the example \ref{prim}, {\Large {$\sigma_{2^{3}}$}}
656: has been obtained from {\Large {$\sigma_{2^{2}},$}} both of them
657: corresponding respectively to the $2^{3}$-periodic orbit and to the
658: $2^{2}$-periodic orbit. With the theorem ~\ref{fuerte} it is not
659: necessary to know information about previous orbits to obtain the
660: results desired, let us see it in the following example.
661:
662: Applying the theorem~\ref{fuerte}, it results in \[
663: \text{\Large{$\sigma_{2^{3}}$}}=\partial R(1,2^{2}+1;2^{1})=\partial R(1,5;2).\]
664: Using definition \ref{def5}, it is obtained \[
665: \partial R(1,5;2)=R(R(1,5;2);2^{0})=R(1,5,7,3,;1)=(1\ 5\ 7\ 3\ 4\ 8\ 6\ 2)\]
666: that coincides with the obtained in the example \ref{prim} \end{ex}
667: The theorem ~\ref{fuerte} has the following reformulation
668:
669: \begin{thm}
670: \label{refu} Let $f:\mathrm{I}\to\mathrm{I}$ be an unimodal function
671: undergoing a period doubling cascade and let $O=\{ x_{1},x_{2},\ldots,x_{2^{p}}\}$
672: be the $2^{p}$-periodic orbit of the cascade with cardinals $\{ C_{(1,2^{p})}^{*},\ldots,C_{(2^{p},2^{p})}^{*}\}$.
673: Let the permutation $\text{{\Large{$\sigma_{2^{p}}$}}}=(\sigma_{(1,2^{p})},\,\sigma_{(2,2^{p})},\ldots,\sigma_{(2^{p},2^{p})})\quad p\in\N,\quad p\geq2$
674: be determined by\[
675: \text{{\Large{$\sigma_{2^{p}}$}}}=(\sigma_{(1,2^{p})},\,\ldots,\sigma_{(2^{p},2^{p})})=\partial R(1,2^{p-1}+1;2^{p-2})\quad p\in\N,\ p\geq2,\]
676: then
677:
678: \[
679: C_{(i,2^{p})}^{*}=x_{\sigma_{(i,2^{p})}}\]
680:
681: \end{thm}
682: \begin{proof}
683: By definition $f^{\sigma_{(i,2^{p})}}(C)=x_{\sigma_{(i,2^{p})}}$
684: and by theorem \ref{fuerte} we get \[
685: f^{\sigma_{(i,2^{p})}}(C)=C_{(i,2^{p})}^{*}\qquad\text{ with }\ \text{{\Large{$\sigma_{2^{p}}$}}}=\partial R(1,2^{p-1}+1;2^{p-2})\quad p\in\N,\ p\geq2\]
686: then \[
687: C_{(i,2^{p})}^{*}=x_{\sigma_{(i,2^{p})}}\qquad\text{ with }\ \text{{\Large{$\sigma_{2^{p}}$}}}=\partial R(1,2^{p-1}+1;2^{p-2})\quad p\in\N,\ p\geq2.\]
688:
689: \end{proof}
690: \begin{cor}
691: \label{cor1} With the conditions of the theorem \ref{refu}, for
692: every $p\geq2$ with $p\in\N$ the following holds
693:
694: \[
695: f^{i}(C)=C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}\quad i=1,2,\ldots,2^{p}\]
696: where $\text{{\Large{$\sigma_{2^{p}}^{-1}$}}}=(\sigma_{(1,2^{p})}^{-1},\,\sigma_{(2,2^{p})}^{-1},\ldots,\sigma_{(2^{p},2^{p})}^{-1}),$
697: is the inverse permutation of $\text{{\Large{$\sigma_{2^{p}}$}}}=(\sigma_{(1,2^{p})},\,\sigma_{(2,2^{p})},\ldots,\sigma_{(2^{p},2^{p})})$.
698: \end{cor}
699: \begin{proof}
700: By the Theorem \ref{fuerte} it follows $f^{\sigma_{(j,2^{p})}}(C)=C_{(j,2^{p})}^{*}$
701: with {\Large {$\sigma_{2^{p}}$}}$=\partial R(1,2^{p-1}+1;2^{p-2})$.
702: If it is taken $j=\sigma_{(i,2^{p})}^{-1}$ then by definition of
703: inverse permutation it results in \[
704: \sigma_{(j,2^{p})}=\sigma_{(\sigma_{(i,2^{p})}^{-1},2^{p})}=i\]
705: and therefore $f^{i}(C)=C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}$
706: with {\Large $\sigma_{2^{p}}^{-1}$} the inverse of {\Large {$\sigma_{2^{p}}$}}{\Large \par}
707: \end{proof}
708: \begin{ex}\label{ter} Taking in mind the results from example \ref{prim},
709: the inverse permutation of {\Large {$\sigma_{2^{3}}$}} is \[
710: \text{\Large{$\sigma_{2^{3}}^{-1}$}}=\left(\begin{array}{cccccccc}
711: 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8\\
712: 1 & 8 & 4 & 5 & 2 & 7 & 3 & 6\end{array}\right)\]
713: that determines in which order are reached the cardinals as the orbit
714: is visited; that is to say, those are reached in the following order
715: (see fig. \ref{fig:2a-and-2b}a)
716:
717: \begin{multline*}
718: \underbrace{C_{(\sigma_{(1,2^{3})}^{-1},2^{3})}^{*}}_{1}\longrightarrow\underbrace{C_{(\sigma_{(2,2^{3})}^{-1},2^{3})}^{*}}_{8}\longrightarrow\underbrace{C_{(\sigma_{(3,2^{3})}^{-1},2^{3})}^{*}}_{4}\longrightarrow\underbrace{C_{(\sigma_{(4,2^{3})}^{-1},2^{3})}^{*}}_{5}\longrightarrow\underbrace{C_{(\sigma_{(5,2^{3})}^{-1},2^{3})}^{*}}_{2}\longrightarrow\\
719: \longrightarrow\underbrace{C_{(\sigma_{(6,2^{3})}^{-1},2^{3})}^{*}}_{7}\longrightarrow\underbrace{C_{(\sigma_{(7,2^{3})}^{-1},2^{3})}^{*}}_{3}\longrightarrow\underbrace{C_{(\sigma_{(2^{3},2^{3})}^{-1},2^{3})}^{*}}_{6}\longrightarrow\underbrace{C_{(\sigma_{(1,2^{3}),2^{3})}^{-1}}^{*}}_{1}\end{multline*}
720:
721:
722: That is to say, the corollary \ref{cor1} indicates that the inverse
723: permutation determines the order in which the orbit is visited. \end{ex}
724:
725: \begin{cor}
726: \label{sec} With conditions of theorem \ref{refu}, the symbolic
727: sequence of the orbit is \[
728: (\mathrm{I}_{(\sigma_{(1,2^{p})}^{-1},2^{p})},\,\mathrm{I}_{(\sigma_{(2,2^{p})}^{-1},2^{p})},\ldots,\mathrm{I}_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})})\]
729: with \[
730: \left\{ \begin{array}{ll}
731: \mathrm{I}_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}=C\\
732: \mathrm{I}_{(\sigma_{(i,2^{p})}^{-1},2^{p})}=\left\{ \begin{array}{ll}
733: \mathrm{R} & \mbox{si $C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}$ is to the right of $C_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}^{*}$ }\\
734: \mathrm{L} & \mbox{otherwise.}\end{array}\right.\end{array}\right.\]
735:
736: \end{cor}
737: \begin{proof}
738: In the orbit with cardinals $\{ C_{(\sigma_{(1,2^{p})}^{-1},2^{p})}^{*},\ldots,C_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}^{*}\},$
739: as $C=C_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}^{*}$, we mark $C_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}^{*}$
740: as symbol $'\mathrm{C}'$. Futhermore, if \, $C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}$
741: is to the right (left) of \, $C_{(\sigma_{(2^{p},2^{p})}^{-1},2^{p})}^{*}$
742: we mark it with $\mathrm{R}$\, ($\mathrm{L}$) and the symbolic
743: sequence of the orbit is thus obtained.
744: \end{proof}
745: \begin{ex} Since $C=C_{(\sigma_{(2^{3},2^{3})}^{-1},2^{3})}^{*}=C_{(6,2^{3})}^{*}$
746: then $C_{(7,2^{3})}^{*}$ and $C_{(2^{3},2^{3})}^{*}$ are to the
747: left of $C_{(6,2^{3})}^{*}$ and they are assigned symbol $\mathrm{L}$
748: according to corollary \ref{sec}. On the other hand $C_{(1,2^{3})}^{*},\ldots,C_{(5,2^{3})}^{*}$
749: are to the right of $C_{(6,2^{3})}^{*}$ and they are assigned $\mathrm{R}$
750: according to corollary \ref{sec}. As the inverse permutation, according
751: example \ref{ter}, is \[
752: C_{(1,2^{3})}^{*},\ C_{(8,2^{3})}^{*},\ C_{(4,2^{3})}^{*},\ C_{(5,2^{3})}^{*},\
753: C_{(2,2^{3})}^{*},\ C_{(7,2^{3})}^{*},\ C_{(3,2^{3})}^{*},\ C_{(6,2^{3})}^{*}\]
754: doing the substitution results in $\mathrm{R}$ $\mathrm{L}$ $\mathrm{R}$
755: $\mathrm{R}$ $\mathrm{R}$ $\mathrm{L}$ $\mathrm{R}$ $\mathrm{C}$
756: or, equivalently, $\mathrm{C}$ $\mathrm{R}$ $\mathrm{L}$ $\mathrm{R}$
757: $\mathrm{R}$ $\mathrm{R}$ $\mathrm{L}$ $\mathrm{R}$.\end{ex}
758:
759: The corollary \ref{sec} has the following reformulation:
760:
761: \begin{cor}
762: With the conditions of theorem \ref{refu}, \[
763: C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}=x_{i}\]
764:
765: \end{cor}
766: \begin{proof}
767: By theorem \ref{refu} it is known that $C_{(j,2^{p})}^{*}=x_{\sigma_{(j,2^{p})}}.$
768: \, Then, if we take $j=\sigma_{(i,2^{p})}^{-1}$ then \[
769: C_{(\sigma_{(i,2^{p})}^{-1},2^{p})}^{*}=x_{(\sigma_{(\sigma_{(i,2^{p})}^{-1},2^{p})})}=x_{i}\]
770:
771: \end{proof}
772:
773: \section{Conclusions}
774:
775: Theorem \ref{fuerte} determines the cardinal ordering of fixed points
776: for any orbit of a period doubling cascade through its associated
777: permutation {\Large $\sigma_{2^{p}}$}. This result is reached without
778: using previous information, hence the importance of this result. In
779: the corollaries that follow from this theorem it is shown, for instance,
780: that the symbolic sequence of any $2^{p}$-periodic orbit is obtained
781: from {\Large $\sigma_{2^{p}}$} just giving the number of the bifurcation
782: $p$.
783:
784: Let us point out that topological analysis, in particular, templates,
785: is a powerful tool to study dynamical systems \cite{Gil98}. Characterization
786: by templates of dynamical systems is not limited to discrete systems
787: \cite{lete94}. Symbolic dynamics plays an important role in these
788: techniques. With the complete characterization presented in this paper,
789: we complete the information about the dynamics of the system.
790:
791:
792:
793: \begin{thebibliography}{10}
794: \bibitem{Fei1} \textsc{Mitchell J. Feigenbaum:} \textit{Quantitative
795: Universality for a Class of Nonlinear Transformations}, Journal of
796: Statistical Physics 19 (1978) 25-52.
797:
798: \bibitem{Fei2} \textsc{Mitchell J. Feigenbaum:} \textit{The Universality
799: Metric Properties for Nonlinear Transformations}, Journal of Statistical
800: Physics 19 (1978) 25-52M.
801:
802: \bibitem{lete} \textsc{Christophe Letellier, Mounia Bennoud and Gilles
803: Martel:} \textit{Intermittency and period-doubling cascade on tori
804: in a bimode laser model Chaos}, Solitons \& Fractals, Volume 33, Issue
805: 3, August 2007, Pages 782-794.
806:
807: \textsc{\bibitem{kiss}István Z. Kiss, Zoltán Kazsu and Vilmos Gáspár:}
808: \textit{Tracking unstable steady states and periodic orbits of oscillatory
809: and chaotic electrochemical systems using delayed feedback control.
810: Chaos,} Volume 16, 2006, 033109.
811:
812: \bibitem{giacomo} \textsc{Innocenti, Giacomo; Morelli, Alice; Genesesio,
813: Roberto; Torcini, Alessandro:} \textit{Dynamical phases of the Hindmars-Rose
814: neuronal model: Studies of the transition from bursting to spiking
815: chaos. Chaos}, Dec 2007, Vol 17, Issue 4, DOI:10.1063/1.2818153.
816:
817: \bibitem{metro} \textsc{N. Metroplois, M.L. Stein, and P.R. Stein:}
818: \textit{On Finite Limit Sets for Transformations on the Unite Interval},
819: Journal of Combinatorial Theory 15 (1973) 25-44.
820:
821: \bibitem{col} \textsc{P. Collet, J.P. Eckmann, O.E. Landford :}\textit{Universal
822: Properties of Maps on an interval}, Commun. Math. Phys. 76 (1980)
823: 211-254.
824:
825: \bibitem{and} \textsc{Andrianov, A. A.; Cannata, F.; Dedonder, J.-P.;
826: Ioffe, M. V.:} Local realizations of q-oscillators in quantum mechanics
827: Physics Letters A, 217 (1996), p. 7-14.
828:
829: \bibitem{Gil} \textsc{Gilmor}, R., Lefranc, M.: The toplogy of chaos,
830: John Wiley \& Sons 2002.
831:
832: \bibitem{Gil98}\textsc{Gilmor, R.:} Topological analysis of chaotic
833: dynamical systems. Review of Modern Physics (1998), Vol 70, Issue
834: 4, Pages 1455-1529.
835:
836: \bibitem{lete94}\textsc{Letellier, C.; Dutertre, P.; Maheu, B.:}
837: Unstable periodic orbits and templates of the Rössler system: Toward
838: a systematic topological characterization. Chaos (1995), Vol 5, Issue
839: 1, Pages 271-282.
840: \end{thebibliography}
841:
842: \end{document}
843: