Skip to main content
Log in

The Periodic Orbits of a Dynamical System Associated with a Family of QRT-Maps

  • Published:
Qualitative Theory of Dynamical Systems Aims and scope Submit manuscript

Abstract

We study the QRT-maps associated with the family of biquadratic curves \(C_d(K)\) with equations \(x^2y^2 - dxy - 1 + K(x^2 + y^2) = 0\). With the Prime Number Theorem and the geometry of elliptic cubics we determine the periods of periodic orbits of the dynamical systems defined by these QRT-maps, and prove sensitivity to its initial conditions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13

Similar content being viewed by others

References

  1. Appel, P., Lacour, E.: Principes de la théorie des fonctions elliptiques. Gauthier-Villars, Paris (1897)

    MATH  Google Scholar 

  2. Bastien, G., Mañosa, V., Rogalski, M.: On periodic solutions of 2-periodic Lyness’ equations. Int. J. Bifurc. Chaos 23, 4 (2013)

    Article  MathSciNet  Google Scholar 

  3. Bastien, G., Rogalski, M.: Global behavior of the solutions of Lyness’ difference equation \(u_{n+2}u_n=u_{n+1}+a\). J. Differ. Equ. Appl. 10, 977–1003 (2004)

    Article  Google Scholar 

  4. Bastien, G., Rogalski, M.: On the algebraic difference equation \(u_{n+2}u_n=\psi (u_{n+1})\) in \({\mathbb{R}}^+_*\) related to a family of elliptic quartics in the plane. Adv. Differ. Equ. 3, 227–261 (2005)

    MATH  Google Scholar 

  5. Bastien, G., Rogalski, M.: A biquadratic system of two order one difference equations: periods, chaotic behavior of the associated dynamical system. Int. J. Bifurc. Chaos 22(11), 1 (2012a)

    Article  MathSciNet  Google Scholar 

  6. Bastien, G., Rogalski, M.: Periods of the solutions of a difference equation in \({\mathbb{R}}^+_*\) associated to a special QRT-map. Dyn. Cont. Discret. Impuls. Syst. Ser. A 20, 727–745 (2013)

    MathSciNet  MATH  Google Scholar 

  7. Bastien, G., Rogalski, M.: A QRT-system of two order one homographic difference equations: conjugation to rotations, periods of periodic solutions, sensitiveness to initial conditions. In: Alsedà i Soler, L., Cushing, J., Elaydi, S., Pinto, A. (eds.) Difference Equations, Discrete Dynamical Systems and Applications ICDEA, : Springer Proceedings in Mathematics and Statistics, vol. 180. Springer, Berlin (2012)

  8. Bastien, G., Rogalski, M.: On the algebraic difference equations \(u_{n+2} + u_n = \psi (u_{n+1})\) in \({{{\mathbb{R}}}}\), related to a family of elliptic quartics in the plane. J. Math. Anal. Appl. 326, 822–844 (2007)

    Article  MathSciNet  Google Scholar 

  9. Bastien, G., Rogalski, M.: QRT-familmies of degree four biquadratic curves each of them has genus zero, associated dynamical systems. To appear in Actes of the Congres ICDEA 2019, London (2019)

  10. Bastien, G., Rogalski, M.: Dynamical systems associated to QRT families of degree four biquadratic curves whose each curve is of genus zero. Sarajevo J. Math. (2019) (To appear)

  11. Bastien, G., Rogalski, M.: On some algebraic difference equation \(u_{n+2}u_n=\Psi (u_{n+1})\) in \({{{\mathbb{R}}}} _*^+\), related to families of conics or cubics: generalization of the Lyness’ equation. J. Math. Anal. Appl. 300, 303–333 (2004)

    Article  MathSciNet  Google Scholar 

  12. Bastien, G., Rogalski, M.: Behavior of orbits and periods of a dynamical system in \({{{\mathbb{R}}}}_*^2\) associated to a special QRT-map. Dynamics of Continuous, Discrete and Impulsive Systems (2019) (To appear)

  13. Cima, A., Gasull, A., Mañosa, V.: Studying discrete dynamical systems through differential equations. J. Differ. Equ. 244(3), 630–648 (2008)

    Article  MathSciNet  Google Scholar 

  14. Duistermaat, J.: Discrete Integrable Systems. QRT Maps and Elliptic Surfaces. Springer, Berlin (2010)

    Book  Google Scholar 

  15. Esch, J., Rogers, T.D.: The screenaver map: dynamics on elliptic curves arising from polygonal folding. Discret. Comput. Geom. 25, 477–502 (2001)

    Article  Google Scholar 

  16. Jogia, D., Roberts, J.A.G., Vivaldi, F.: An algebraic approach to integrable maps of the plane. J. Phys. A 39, 1133–1149 (2006)

    Article  MathSciNet  Google Scholar 

  17. Perrin, D.: Géométrie algébrique. Une introduction. CNRS Editions, EDP Sciences (2001)

  18. Quispel, G.R.W., Roberts, J.A.G., Thompson, C.J.: Integrable mappings and soliton equations. Phys. Lett. A 126, 419–421 (1988)

    Article  MathSciNet  Google Scholar 

  19. Robin, G.: Estimation de la fonction de Tchebychef \(\Theta \) sur le k-ième nombre premier et grandes valeurs de \(\omega (n)\) nombre de diviseurs premiers de n. Acta Arithmetica XLII, 367–389 (1983)

    Article  Google Scholar 

  20. Silverman, J.H., Tate, J.: Rational points on elliptic curves. Springer, Berlin (1992)

    Book  Google Scholar 

  21. Tenenbaum, G.: Introduction à la théorie analytique et probabiliste des nombres. Cours spécialisé no. 1, Collection Société Mathématique de France, Paris (1995)

  22. Zeeman, E.C.: Geometric unfolding of a difference equation. Hertford College, Oxford, Unpublished paper. Reprinted as a Preprint of the Warwick Mathematics Institute, 2008 (1996)

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Marc Rogalski.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix 1: 4 and 6 are Minimal Periods

(a) We will find a necessary and sufficient condition for 4 being a minimal period.

Let \({\mathcal {V}}\) and \({\mathcal {H}}\) be the involutions defined by vertical [resp. horizontal] alignment of two points on the same curve \({\mathcal {C}}_d(K)\), so that we have \(T_d={\mathcal {V}}\circ {\mathcal {H}}\) and \(T_d^{-1}={\mathcal {H}}\circ {\mathcal {V}}\) (excepted for the points \(M_0\), \(M_1\), \(N_0\), \(N_1\)).

From Proposition 11 or Corollary 12, we see that \(T_d\) is 4-periodic on \({\mathcal {C}}_d(K)\) if and only if (with an exception that we will see later) \(A_0:= (m,m)\) is 4-periodic for \(T_d\) (m is given by relation (7)), or the same condition for \(B_0:=(-m,-m)\). The condition \(T_d^4(A_0)=A_0\) can be written

$$\begin{aligned} {\mathcal {V}}\circ {\mathcal {H}}\circ {\mathcal {V}}\circ {\mathcal {H}}(A_0)={\mathcal {H}}\circ {\mathcal {V}}\circ {\mathcal {H}}\circ {\mathcal {V}}(A_0). \end{aligned}$$
(61)

But these two points are symmetric with respect to the diagonal (because the curve is symmetric). So they can be equal if and only if they are both equal to \(A_0\) or to \(B_0\). If they are equal to \(A_0\), we would have \(T_d^2(A_0)=A_0\), and \(T_d\) would be 2-periodic, which is false. So we must have the equality

$$\begin{aligned} T_d(A_0)=T_d^{-1}(B_0). \end{aligned}$$
(62)

If we put \(T_d(A_0)=(X,Y)\) and \(T_d^{-1}(B_0)=(X_1,Y_1)\), the condition (61) is \(X-X_1=0\) and \(Y-Y_1=0\), and the easy (with Maple) computation gives

$$\begin{aligned} d=\frac{\sqrt{2}(K^2+1)}{\sqrt{K^2-1}}. \end{aligned}$$
(63)

As an example, for \(d=5\) and \(K=3\), \(T_5\) is 4-periodic on the curve \({\mathcal {C}}_5(3)\).

An exception happens when the two denominators (which are the same ones) of \(X-X_1\) and \(Y-Y_1\) are 0, that is when

$$\begin{aligned} m^{12}-5dm^{10}+3(d^2+1)m^8-(d^3-2d)m^6+3(d^2+1)m^4+3dm^2+1=0.\nonumber \\ \end{aligned}$$
(64)

But it is easy to see that relations (64) and (62) are incompatible. So this exception is not a curve on which \(T_d\) is 4-periodic.

It is also interesting to note that the condition that \({\mathcal {H}}(A_0)=N_0\), so that \(Y=\infty \), id est \(T_d(A_0)=V\), is exactly the relation (61), and by symmetry it is also the relation for having \(T_d^{-1}(B_0)=H\). So we find again the previous exception.

(b) Now we will find a condition for 6 being a minimal period.

With the same reasoning we find that the condition is

$$\begin{aligned} ({\mathcal {V}}\circ {\mathcal {H}})^{3\circ }(A_0)=({\mathcal {H}}\circ {\mathcal {V}})^{3\circ }(A_0), \end{aligned}$$
(65)

so that, by symmetry, these to points must be equal to \(A_0\) or to \(B_0\). If it would \(A_0\), \(T_d\) would be 3-periodic, and 6 would not be minimal. So the good condition is

$$\begin{aligned} T_d^2(A_0)=T_d^{-1}(B_0). \end{aligned}$$
(66)

It is now easy, with the use of Maple, and by computing \(T_d^2(A_0)=(X_2,Y_2)\), to find the condition for 6 being a minimal period (at least, it is a sufficient condition)

$$\begin{aligned} K^2d^4-(K^4-1)(K^2+1)d^2+3(K^2+1)^4=0, \end{aligned}$$
(67)

which gives easily d as a function of K

$$\begin{aligned} d=\frac{(K^2+1)\sqrt{K^2-1\pm \sqrt{K^4-14K^2+1}}}{K\sqrt{2}}. \end{aligned}$$
(68)

As an example, for \(d=4\sqrt{3+2\sqrt{3}}\) and \(K=2+\sqrt{3}\), we obtain a curve on which 6 is a minimal period.

Appendix 2: On the Forms of the Curves \({\mathcal {C}}_d(K)\) (Proof of Lemma 5)

The plan of proof of Lemma 5 is simple :

(1) The curves are starlike with respect the point (0, 0) : we put \(x=\rho u\) and \(y=\rho v\), with a unit vector (uv). If \(uv\not =0\) we have with equation of \({\mathcal {C}}_d(K)\) a quadratic equation in \(\rho ^2\) with a unique positive solution, so \(\rho =\pm \alpha \): the curve is starlike. If \(uv = 0\), \(\rho =\pm \frac{1}{\sqrt{K}}\) if \(K>0\), and if \(K\le 0\) and \(uv=0\), there is no solution: the curves \({\mathcal {C}}_d(K)\) does not cut the axes.

(2) There is no double point in finite distance (see Lemma 6).

(3) If \(K>0\), one see the inclusion \({\mathcal {C}}_d(K)\subset B\Big ((0,0),\sqrt{\frac{1+d^2/4}{K}}\Big )\): one has \(\displaystyle x^2+y^2=\frac{1+dxy-x^2y^2}{K}\le \frac{1+d^2/4}{K}\). So, from the three previous points, one see that the curves are homeomorphic to circles for \(K>0\).

(4) If \(K<0\), it is easy to find the asymptotes of the curve \({\mathcal {C}}_d(K)\). \(\square \)

Appendix 3: Previous Particular Results of the Authors About QRT-Maps

The invention of QRT-maps was originally in [18], for physical reasons, and the essential exemple was on [22], but these papers were not easy to find. So the first papers ([3, 11] and [4]) were slightly different. They studied the now called “symmetric special QRT-maps”, where the family of biquadratic curves \({\mathcal {C}}(K)\) with equations \(Q_1(x,y)-KQ_2(x,y)=0\) were symmetric with respect the diagonal, and with \(Q_2(x,y)=xy\); and the QRT-map was defined by the following way: if \({\mathcal {C}}(K)\) is the curve passing through the point M, we cut it in \(M_1\) by the horizontal line which contains M, and then T(M) is the symmetric of \(M_1\) with respect the diagonal (it is on \({\mathcal {C}}(K)\)).

These cases correspond to the study of difference equations of the form

$$\begin{aligned} u_{n+2}u_n=\frac{au_{n+1}^2+bu_{n+1}+c}{du_{n+1}^2+eu_{n+1}+f}. \end{aligned}$$
(69)

In [8] we studied the case of difference equations \(u_{n+2}+u_n=\psi (u_{n+1})\) which is associated with “symmetric QRT-maps” defined by families of the forms \(Q_1(x,y)+K=0\).

In each of these papers, we determine the possible periods of periodic orbits, prove the density of periodic points and not periodic points, and a form of sensitivity to initial conditions.

In the other works, we studied classical QRT-maps, associated with a couple of difference equations of the form \(u_{n+1}u_n=f(v_n), v_{n+1}v_n=g(u_{n+1})\). We give some examples :

$$\begin{aligned}&[7] \quad u_{n+1}u_n=c+\frac{d}{v_n},\,\,v_{n+1}v_n \\&\quad =c+\frac{d}{v_{n+1}} \quad \mathrm{with\,\,curves} \quad xy(x+y)+xy+d-Kxy=0. \\&[5]\,\, u_{n+1}u_n=v_n^2-bv_n+c,\,v_{n+1}v_n\\&\quad =u_{n+1}^2-au_{n+1}+c\,\, \mathrm{with\,\,curves}\,\, x^2+y^2-ax-by+c-Kxy=0. \\&[6]\,\, u_{n+1}u_n=av_n+b,\,v_{n+1}v_n\\&\quad =u_{n+1}+\frac{b}{u_{n+1}}\,\, \mathrm{with\,\,curves}\,\, xy^2+x^2+ay+b-Kxy=0. \end{aligned}$$

In [2] we studied the 2-periodic Lyness’ equation \(u_{n+2}u_n=u_{n+1}+a_n\), with \(a_n\) 2-periodic.

In [12] we study the particular case of the dynamical systems \((x,y)\mapsto T_d(x,y)= (X,Y)\) given by

$$\begin{aligned} (X,Y)=\Big (\frac{1}{x}\frac{dy^2-20y+16}{y^2-5y+d}, \frac{1}{y}\frac{dX^2-20X+16}{X^2-5X+d}\Big ), \end{aligned}$$

with curves \({\mathcal {C}}_d(K)\) equations of them are

$$\begin{aligned} x^2y^2-5xy(x+y)+d(x^2+y^2)-20(x+y)+16-Kxy=0. \end{aligned}$$

In the present paper we begin study of non-special QRT-dynamical systems, associated with QRT-families of curves with equations \(Q_1(x,y)-KQ_2(x,y)=0\) with \(Q_2(x,y)\) not of the form xy.

In fine, in [9] and [10] we present examples of QRT-families of degree four, but such that each of the curves of the family has genus zero, and studied the correspondent dynamical systems.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Bastien, G., Rogalski, M. The Periodic Orbits of a Dynamical System Associated with a Family of QRT-Maps. Qual. Theory Dyn. Syst. 19, 57 (2020). https://doi.org/10.1007/s12346-020-00393-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s12346-020-00393-2

Keywords

Navigation