Abstract
In this paper, we consider the orbital-reversibility problem for an n-dimensional vector field, which consists in determining if there exists a time-reparametrization that transforms the vector field into a reversible one. We obtain an orbital normal form that brings out the invariants that prevent the orbital-reversibility. Hence, we obtain a necessary condition for a vector field to be orbital-reversible. Namely, the existence of an orbital normal form which is reversible to the change of sign in some of the state variables. The necessary condition provides an algorithm, based on the vanishing of the orbital normal form terms that avoid the orbital-reversibility, that is applied to some families of planar and three-dimensional systems.
Similar content being viewed by others
Data Availability
The authors declare that data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
References
Algaba, A., Checa, I., Gamero, E., GarcÃa, C.: On orbital-reversibility for a class of planar dynamical systems. Commun. Nonlinear Sci. 20, 229–239 (2015)
Algaba, A., Freire, E., Gamero, E.: Hypernormal forms for equilibria of vector fields. Codimension on linear degeneracies. Rocky Mt. J. Math. 29, 13–45 (1999)
Algaba, A., Freire, E., Gamero, E., GarcÃa, C.: Quasihomogeneous normal forms. J. Comput. Appl. Math. 150, 193–216 (2002)
Algaba, A., Freire, E., Gamero, E., GarcÃa, C.: An algorithm for computing quasi-homogeneous formal normal forms under equivalence. Acta Appl. Math. 80, 335–359 (2004)
Algaba, A., Gamero, E., GarcÃa, C.: The integrability problem for a class of planar systems. Nonlinearity 22, 395–420 (2009)
Algaba, A., Gamero, E., GarcÃa, C.: The reversibility problem for quasi-homogeneous dynamical systems. Discrete Contin. Dyn. Syst. 33, 3225–3236 (2013)
Algaba, A., Gamero, E., GarcÃa, C.: The center problem: a view from the normal form theory. J. Math. Anal. Appl. 434, 680–697 (2016)
Algaba, A., GarcÃa, C., Giné, J.: Analytic integrability for some degenerate planar vector fields. J. Differ. Equ. 257, 549–565 (2014)
Algaba, A., GarcÃa, C., Giné, J.: Center conditions of a particular polynomial differential system with a nilpotent singularity. J. Math. Anal. Appl. 483, 123639 (2020)
Algaba, A., GarcÃa, C., Giné, J.: Orbital reversibility of planar vector fields. Mathematics 9, 14 (2021)
Algaba, A., GarcÃa, C., Teixeira, M.A.: Reversibility and quasi-homogeneous normal forms of vector fields. Nonlinear Anal. Theory 73, 510–525 (2010)
Berthier, M., Moussu, R.: Réversibilité et classification des centres nilpotents. Ann. I Fourier 44, 465–494 (1994)
Bruno, A.D.: Analysis of the Euler–Poisson equations by methods of power geometry and normal form. J. Appl. Math. Mech. 71, 168–199 (2007)
Buzzi, C.A., Llibre, J., Medrado, J.C.R.: Phase portraits of reversible linear differential systems with cubic homogeneous polynomial nonlinearities having a non-degenerate center at the origin. Qual. Theory Dyn. Syst. 7, 369–403 (2009)
Buzzi, C., Teixeira, M.A., Yang, J.: Hopf-zero bifurcations of reversible vector fields. Nonlinearity 14, 623–638 (2001)
Champneys, A.R.: Codimension-one persistence beyond all orders of homoclinic orbits to singular saddle centres in reversible systems. Nonlinearity 14, 87–112 (2001)
Chavarriga, J., Giacomini, H., Giné, J., Llibre, J.: Local analytic integrability for nilpotent centers. Ergod. Theory Dyn. Syst. 23, 417–428 (2003)
Christopher, C., Li, C.: Limit Cycles of Differential Equations. Birkhauser, Berlin (2007)
Chow, S.N., Hale, J.K.: Methods of Bifurcation Theory. Springer, New York (1982)
Cicogna, G., Gaeta, G.: Symmetry and Perturbation Theory in Nonlinear Dynamics. Springer, Berlin (1999)
Devaney, R.L.: Reversible diffeomorphisms and flows. Trans. Am. Math. Soc. 218, 89–113 (1976)
Gaeta, G.: Normal forms of reversible dynamical systems. Int. J. Theor. Phys. 33, 1917–1928 (1994)
Gasull, A., Torregrosa, J.: Center problem for several differential equations via Cherkas’ method. J. Math. Anal. Appl. 228, 322–343 (1998)
Giacomini, H., Giné, J., Llibre, J.: The problem of distinguishing between a center and a focus for nilpotent and degenerate analytic systems. J. Differ. Equ. 227, 406–426 (2006)
Giné, J., Maza, S.: The reversibility and the center problem. Nonlinear Anal. Theory 74, 695–704 (2011)
Han, M., Petek, T., Romanovski, V.G.: Reversibility in polynomial systems of ODE’s. Appl. Math. Comput. 338, 55–71 (2018)
Lamb, J.S.W., Lima, M.F.S., Martins, R.M., Teixeira, M.A., Yang, J.: On the Hamiltonian structure of normal forms at elliptic equilibria of reversible vector fields in \(\mathbb{R}^4\). J. Differ. Equ. 269, 11366–11395 (2020)
Lamb, J.S.W., Roberts, J.A.G.: Time-reversal symmetry in dynamical systems: a survey. Phys. D 112, 1–39 (1998)
Lamb, J.S.W., Roberts, J.A.G., Capel, H.W.: Conditions for local (reversing) symmetries in dynamical systems. Phys. A 197, 379–422 (1993)
Li, C., Llibre, J.: Quadratic perturbations of a quadratic reversible Lotka–Volterra system. Qual. Theory Dyn. Syst. 9, 235–249 (2010)
Llibre, J., Medrado, J.C.: Darboux integrability and reversible quadratic vector fields. Rocky Mt. J. Math. 35, 1999–2057 (2005)
Matveyev, M.V.: Lyapunov stability of equilibrium states of reversibles systems. Math. Notes+ 57, 63–72 (1993)
Matveyev, M.V.: Reversible systems with first integrals. Phys. D 112, 148–157 (1998)
Medrado, J.C., Teixeira, M.A.: Symmetric singularities of reversible vector fields in dimension three. Phys. D 112, 122–131 (1998)
Montgomery, D., Zippin, L.: Topological Transformation Groups. Interscience Publ, New York (1955)
Roberts, J.A.G., Quispel, G.R.W.: Chaos and time-reversal symmetry. Order and chaos in reversible dynamical systems. Phys. Rep. 216, 63–177 (1992)
Romanovski, V.G.: Time-reversibility in 2-dimensional systems. Open Syst. Inf. Dyn. 15, 359–370 (2008)
Sabatini, M.: Every period annulus is both reversible and symmetric. Qual. Theory Dyn. Syst. 16, 175–185 (2017)
Sevryuk, M.B.: Reversible Systems. Lecture Notes Mathematics, vol. 1211. Springer, Berlin (1986)
Teixeira, M.A.: Local reversibility and applications. In: Bruce, J.W., Tari, F. (eds.) Real and Complex Singularities. Research Notes in Mathematics, vol. 412, pp. 251–265. CRC, Boca Raton (2000)
Teixeira, M.A., Yang, J.: The center-focus problem and reversibility. J. Differ. Equ. 174, 237–251 (2001)
Tkhai, V.N.: The reversibility of mechanical systems. J. Appl. Math. Mech. 55, 461–469 (1991)
Wei, L., Romanovski, V., Zhang, X.: Generalized involutive symmetry and its application in integrability of differential systems. Z. Angew. Math. Phys. 68, 132 (2017)
Xiong, Y., Cheng, R., Li, N.: Limit cycle bifurcations in perturbations of a reversible quadratic system with a non-rational first integral. Qual. Theory Dyn. Syst. 19, 97 (2020)
Yagasaki, K., Wagenknecht, T.: Detection of symmetric homoclinic orbits to saddle-centres in reversible systems. Phys. D 214, 169–181 (2006)
Acknowledgements
This research was partly supported by the Ministerio de Ciencia e Innovación, fondos FEDER (project MTM2017-87915-C2-1-P), by the Ministerio de Ciencia, Innovación y Universidades, fondos FEDER (project PGC2018-096265-B-I00) and by the ConsejerÃa de EconomÃa, Innovación, Ciencia y Empleo de la Junta de AndalucÃa (projects P12-FQM-1658, UHU-1260150, TIC-130, FQM-276).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
A Technical Results on Orbital Normal Forms
We present now some technical results about the orbital normal form reduction procedure. Along this appendix, the vector field \({\mathbf{F}}\) is defined in (6). The first result can be found in [11].
Lemma 17
Let us consider \(k\in \mathbb {N}\) and a couple of near-identity transformations \(\Phi \) and \(\Psi \), with generators \({\mathbf{U}}=\sum _{j\ge 1}{\mathbf{U}}_j\) and \({\mathbf{V}}=\sum _{j\ge k}{\mathbf{V}}_j\) (with \({\mathbf{U}}_j,{\mathbf{V}}_j\in \mathcal {Q}_j^{\mathbf{t}}\)), respectively. Let also consider the near-identity transformation \(\Psi \circ \Phi \), and denote its generator by \({\mathbf{W}}\).
Then, \({\mathbf{U}}\) and \({\mathbf{W}}\) agree up to quasi-homogeneous degree \(k-1\) (i.e., \(\mathcal {J}^{k-1}\left( {\mathbf{U}} \right) =\mathcal {J}^{k-1}({\mathbf{W}})\)), and the k-degree quasi-homogeneous terms are
Next results show that the orbital normal form reduction procedure can also be carried out by changing the order of the temporal and spatial transformations for each degree, that is, first performing the near-identity transformation and then the time-reparametrization.
Lemma 18
Let us consider \(\mu _{k}\in \mathcal {P}_{k}^{\mathbf{t}}\), where \(k\in \mathbb {N}\), and \({\mathbf{U}}=\sum _{j\ge 1} {\mathbf{U}}_j\), with \({\mathbf{U}}_j\in \mathcal {Q}_j^{\mathbf{t}}\). Then, there exists \(\lambda =\sum _{j\ge k}\lambda _j\), with \(\lambda _j \in \mathcal {P}_j^{\mathbf{t}}\), such that
where the lowest-degree quasi-homogeneous term of \(\lambda \) is \(\lambda _k = \mu _{k} \).
Proof
Let us define \(T_{\mathbf{U}}^{(0)}\left( {\mathbf{F}} \right) := {\mathbf{F}}\), and
In [2] it is shown that the transformed system is given by
Let us denote the successive Lie derivatives along the vector field \({\mathbf{U}}\) of the scalar function \(\mu _{k}\) by \(\mu ^{[j]}_{k}\), i.e.,
Using induction, it is easy to prove that
Then,
as claimed, where we have introduced \(\lambda := \sum _{j=0}^l \frac{1}{j!} (-1)^j \mu ^{[j]}_{k} \). Moreover, it is easy to show that \(\lambda _k = \mu _{k} \). \(\blacksquare \)
Lemma 19
Let us consider \(\rho _{k} \in \mathcal {P}_k^{\mathbf{t}}\), where \(k\in \mathbb {N}\), and \(\mu =\sum _{j\ge 1} \mu _j\), where \(\mu _j\in \mathcal {P}_j^{\mathbf{t}}\). Then, there exists \(\eta =\sum _{j\ge 1} \eta _j\), with \(\eta _j \in \mathcal {P}_j^{\mathbf{t}}\), such that:
Moreover, \(\mathcal {J}^{r+k-1}\left( \eta \right) =\mathcal {J}^{r+k-1}(\mu )\) (i.e., \(\eta \) and \(\mu \) agree up to quasi-homogeneous degree \(r+k-1\)), and the \((r+k)\)-degree quasi-homogeneous term is \(\eta _{r+k} = \mu _{r+k} - \nabla \rho _{k} \cdot {\widetilde{\mathbf{F}}}_r\).
Proof
Firstly, we will show that, for each \(l\in \mathbb {N}\), there exists \(\eta ^{(l)}=\sum _{ j \ge l(r+k)} \eta ^{(l)}_j\), with \( \eta ^{(l)}_j \in \mathcal {P}_j^{\mathbf{t}}\), satisfying:
We proceed by induction on l.
-
If \(l=1\), using (3) we obtain
$$\begin{aligned} T_{\rho _{k} {\mathbf{F}}} \left( (1+\mu ){\mathbf{F}} \right)= & {} \left[ (1+\mu ){\mathbf{F}},\rho _{k} {\mathbf{F}} \right] \\= & {} \left( \rho _{k} \left( \nabla \mu \cdot {\mathbf{F}} \right) - (1+\mu ) \left( \nabla \rho _{k} \cdot {\mathbf{F}} \right) \right) {\mathbf{F}}=\eta ^{(1)}{\mathbf{F}}, \end{aligned}$$where we have denoted \(\eta ^{(1)} := \rho _{k} \left( \nabla \mu \cdot {\mathbf{F}} \right) - (1+\mu ) \left( \nabla \rho _{k} \cdot {\mathbf{F}} \right) \). Observe that the principal part of \(\eta ^{(1)}\) is \(\eta _{r+k}^{(1)} = - \nabla \rho _{k} \cdot {\widetilde{\mathbf{F}}}_r\).
-
Now, suppose that (20) holds for \(l-1\), that is, there exists \(\eta ^{(l-1)}=\sum _{ j \ge (l-1)(r+k)} \eta ^{(l-1)}_j\), with \( \eta ^{(l-1)}_j \in \mathcal {P}_j^{\mathbf{t}}\), such that
$$\begin{aligned} T_{\rho _{k} {\mathbf{F}}}^{(l-1)} \left( (1+\mu ){\mathbf{F}} \right) = \eta ^{(l-1)}{\mathbf{F}}. \end{aligned}$$Using again (3), we obtain
$$\begin{aligned} T_{\rho _{k} {\mathbf{F}}}^{(l)} \left( (1+\mu ){\mathbf{F}} \right) = \left[ T_{\rho _{k} {\mathbf{F}}}^{(l-1)} \left( (1+\mu ){\mathbf{F}} \right) , \rho _{k} {\mathbf{F}} \right] = \left[ \eta ^{(l-1)}{\mathbf{F}}, \rho _{k} {\mathbf{F}} \right] = \eta ^{(l)}{\mathbf{F}}, \end{aligned}$$where we have introduced \(\eta ^{(l)} := \rho _{k} \left( \nabla \eta ^{(l-1)}\cdot {\mathbf{F}} \right) - \eta ^{(l-1)} \left( \nabla \rho _{k} \cdot {\mathbf{F}} \right) \). We notice that the principal part of \(\eta ^{(l)}\) has quasi-homogeneous degree \( l(r+k)\).
By the principle of induction, equality (20) holds for all \(l\in \mathbb {N}\). Hence, we have:
where we have denoted \(\eta :=\mu +\sum _{l\ge 1}\frac{1}{l!}\eta ^{(l)}\). To show that the principal part of \(\eta \) is \(\eta _{r+k} = \mu _{r+k} - \nabla \rho _{k} \cdot {\widetilde{\mathbf{F}}}_r\), it is enough to recall that the principal part of \(\eta ^{(1)}\) is \(\eta _{r+k}^{(1)} = - \nabla \rho _{k} \cdot {\widetilde{\mathbf{F}}}_r\). \(\blacksquare \)
The following two lemmas deal with vector fields having a Lie symmetry up to order \(r+m\).
Lemma 20
Let us consider \(m\in \mathbb {N}\) and assume that there exists \({\mathbf{U}}=\sum _{j=1}^{m}{\mathbf{U}}_j\), with \({\mathbf{U}}_j\in \mathcal {Q}_j^{\mathbf{t}}\), such that
Then, for any \(\mu =\sum _{j=1}^{m}\mu _j\), with \(\mu _j\in \mathcal {P}_j^{\mathbf{t}}\), \(l\ge 0\), and \(k=l+1,\dots ,l+m\), we have:
for some \(\mu _{j}^{(l)} \in \mathcal {P}_{j}^{{\mathbf{t}}}\). Moreover, \(\mu _{{j}}^{(l)}\) depends univocally on \({\mathbf{U}}_1,\dots ,{\mathbf{U}}_{{j}-l},\nu _1,\dots ,\nu _{{j}-l},\mu _1,\dots ,\mu _{{j}-l}\).
Proof
We use induction on l.
-
For \(l=0\), the result is trivial because \(\left( T_{\mathbf{U}}^{(0)}\left( \mu {\mathbf{F}} \right) \right) _{r+k}=\left( \mu {\mathbf{F}} \right) _{r+k}=\sum _{j=1}^{k}\mu _{j}{\mathbf{F}}_{r+k-j}\).
-
Assume that the result is true for \(l-1\), being \(l>0\). Then:
$$\begin{aligned} \left( T_{\mathbf{U}}^{(l)}\left( \mu {\mathbf{F}} \right) \right) _{r+k}= & {} \sum _{{j}=1}^{k-l} \left[ \left( T_{\mathbf{U}}^{(l-1)}\left( \mu {\mathbf{F}} \right) \right) _{r+k-{j}},{\mathbf{U}}_{j} \right] \\= & {} \sum _{{i}=l}^{k-1} \sum _{{j}=1}^{k-{i}} \left[ \mu _{{i}}^{(l-1)} {\mathbf{F}}_{r+k-{i}-{j}},{\mathbf{U}}_{j} \right] \\= & {} \sum _{{i}=l}^{k-1} \left( \sum _{{j}=1}^{k-{i}} \left( \nabla \mu _{{i}}^{(l-1)} \cdot {\mathbf{U}}_{j} \right) {\mathbf{F}}_{r+k-{i}-{j}} + \mu _{{i}}^{(l-1)} \sum _{{j}=1}^{k-{i}} \left[ {\mathbf{F}}_{r+k-{i}-{j}},{\mathbf{U}}_{j} \right] \right) \\= & {} \sum _{{i}=l}^{k-1} \left( \sum _{{j}=1}^{k-{i}} \left( \nabla \mu _{{i}}^{(l-1)} \cdot {\mathbf{U}}_{j} \right) {\mathbf{F}}_{r+k-{i}-{j}} \right) + \sum _{{i}=l}^{k-1}\mu _{{i}}^{(l-1)} \left[ {\mathbf{F}},{\mathbf{U}} \right] _{r+k-{i}} . \end{aligned}$$As \(\mathcal {J}^{r+m}\left( \left[ {\mathbf{F}},{\mathbf{U}} \right] \right) =\mathcal {J}^{r+m}\left( \nu {\mathbf{F}} \right) \), then \(\left[ {\mathbf{F}},{\mathbf{U}} \right] _{r+k-{i}} = \left( \nu {\mathbf{F}} \right) _{r+k-{i}} = \sum _{{j}=1}^{k-{i}} \nu _{j} {\mathbf{F}}_{r+k-{i}-{j}}\) for \(k=l+1,\dots ,l+m\). Hence:
$$\begin{aligned} \left( T_{\mathbf{U}}^{(l)}\left( \mu {\mathbf{F}} \right) \right) _{r+k}= & {} \sum _{{i}=l}^{k-1}\sum _{{j}=1}^{k-{i}} \left( \left( \nabla \mu _{{i}}^{(l-1)}\cdot {\mathbf{U}}_{j} \right) + \mu _{i}^{(l-1)}\nu _{j} \right) {\mathbf{F}}_{r+k-{i}-{j}}\\= & {} \sum _{{j}=l+1}^{k}\left( \sum _{{i}=l}^{{j}-1} \left( \left( \nabla \mu _{{i}}^{(l-1)}\cdot {\mathbf{U}}_{{j}-{i}} \right) + \mu _{i}^{(l-1)}\nu _{{j}-{i}} \right) \right) {\mathbf{F}}_{r+k-{j}}, \end{aligned}$$and the result holds by defining \(\mu _{{j}}^{(l)}:=\sum _{{i}=l}^{{j}-1} \left( \left( \nabla \mu _{{i}}^{(l-1)}\cdot {\mathbf{U}}_{{j}-{i}} \right) + \mu _{i}^{(l-1)}\nu _{{j}-{i}} \right) \in \mathcal {P}_{j}^{{\mathbf{t}}}\), that depends univocally on \({\mathbf{U}}_1,\dots ,{\mathbf{U}}_{{j}-l},\nu _1,\dots ,\nu _{{j}-l},\mu _1,\dots ,\mu _{{j}-l}\).
\(\blacksquare \)
Lemma 21
Let us consider \(m\in \mathbb {N}\) and assume that there exists \({\mathbf{U}}=\sum _{j=1}^{m+1}{\mathbf{U}}_j\), with \({\mathbf{U}}_j\in \mathcal {Q}_j^{\mathbf{t}}\), such that
Then:
-
(a)
For any \(\mu = \sum _{j=1}^{m+1}\mu _j\), with \(\mu _j\in \mathcal {P}_j^{\mathbf{t}}\), there exists \(\lambda = \sum _{j=1}^{m+1}\lambda _j\), with \(\lambda _j\in \mathcal {P}_j^{\mathbf{t}}\), such that
$$\begin{aligned} \mathcal {J}^{r+m+1} \left( {\mathbf{U}}{\, }_{ **}{\, }\left( (1+\mu ){\mathbf{F}} \right) \right) = \mathcal {J}^{r+m+1} \left( \left[ {\mathbf{F}},{\mathbf{U}} \right] +(1+\lambda ){\mathbf{F}} \right) . \end{aligned}$$(21) -
(b)
For any \(\lambda = \sum _{j=1}^{m+1}\lambda _j\), with \(\lambda _j\in \mathcal {P}_j^{\mathbf{t}}\), there exists \(\mu = \sum _{j=1}^{m+1}\mu _j\), with \(\mu _j\in \mathcal {P}_j^{\mathbf{t}}\), such that (21) holds.
Proof
As \([{\mathbf{F}},{\mathbf{U}}]_{r+k}=\left( \nu {\mathbf{F}} \right) _{r+k}\) for \(k=1,\dots , m\), it can be shown that
for all \(l\ge 0\), \(k=l+1,\dots ,l+m\). Using that
we obtain
From Lemma 20, we get:
Notice that \(\mu _{j}^{(l)}\) depends univocally on \({\mathbf{U}}_1,\dots ,{\mathbf{U}}_{{j}-1},\nu _1,\dots ,\nu _{{j}-1},\mu _1,\dots ,\mu _{{j}-1}\). On the other hand, we have:
To prove item (a), it is enough to define \(\lambda _j=\mu _{j}+ \sum _{l=1}^{j-1}\mu _{j}^{(l)}\), for \(j=1,\dots , k\).
To prove item (b), it is enough to solve the equation \(\lambda _j=\mu _j+ \sum _{l=1}^{j-1}\mu _{j}^{(l)}\) (this can be done because \(\mu _{j}^{(l)}\) depends univocally on \({\mathbf{U}}_1,\dots ,{\mathbf{U}}_{{j}-1},\nu _1,\dots ,\nu _{{j}-1},\mu _1,\dots ,\mu _{{j}-1}\)) and define \(\lambda _j=\mu _{j}+ \sum _{l=1}^{j-1}\mu _{j}^{(l)}\), for \(j=1,\dots , k\). \(\blacksquare \)
B The Range of \(\overline{\mathcal {NL}}^{(N)}\)
In this appendix, we show that \({\mathrm {Range}}(\overline{\mathcal {NL}}^{(N)})={\mathrm {Range}}(\overline{\mathcal {L}}^{(N)})\). This is a consequence of Propositions 24 and 25.
The following propositions are necessary to prove Propositions 24 and 25.
Proposition 22
Let us consider the vector field \({\mathbf{F}}\) given in (6). Then, for each \({\widetilde{\mathbf{U}}}\in \bigoplus _{j=1}^m\mathcal {R}_j^{\mathbf{t}}\) and \({\widetilde{\lambda }}\in \bigoplus _{j=1}^{m}\mathcal {O}_j^{\mathbf{t}}\) (\(m\in \mathbb {N}\)), there exist \({\widetilde{\mathbf{V}}}\in \bigoplus _{j=1}^{m}\widehat{\mathcal {R}}_j^{\mathbf{t}}\) and \({\widetilde{\nu }}\in \bigoplus _{j=1}^{m}\widehat{\mathcal {O}}_j^{\mathbf{t}}\) such that
Proof
Let us denote
Observe that \(\kappa \le m\), otherwise \({\widetilde{\mathbf{V}}}={\widetilde{\mathbf{U}}}\) and \({\widetilde{\nu }}={\widetilde{\lambda }}\).
From (11), we can write \({\widetilde{\lambda }}_{\kappa } = {\widetilde{\lambda }}_{\kappa }^{(1)} + {\widetilde{\lambda }}_{\kappa }^{(2)}\), where \({\widetilde{\lambda }}_{\kappa }^{(1)} = \nabla {\overline{\rho }}\cdot {\widetilde{\mathbf{F}}}_r \in {\mathrm {Range}}\left( \ell ^{(\texttt {e})}_{\kappa -r} \right) \) (\({\overline{\rho }}\in \mathcal {E}_{\kappa -r}^{\mathbf{t}}\)) and \({\widetilde{\lambda }}_{\kappa }^{(2)} \in \widehat{\mathcal {O}}_{\kappa }^{\mathbf{t}}\). We observe that \({\overline{\rho }}=0\) if \({\widetilde{\lambda }}_{\kappa }^{(1)} = 0\), and \({\overline{\rho }}\in \mathcal {E}_{\kappa -r}^{\mathbf{t}}{\setminus } {\mathrm {Ker}}\left( {\ell }_{\kappa -r} \right) \) otherwise.
Moreover, by (13), we can write \({\widetilde{\mathbf{U}}}_{\kappa } = {\widetilde{\mathbf{U}}}_{\kappa }^{(1)} + {\widetilde{\mathbf{U}}}_{\kappa }^{(2)}\), where \({\widetilde{\mathbf{U}}}_{\kappa }^{(1)} = {\overline{\zeta }}{\widetilde{\mathbf{F}}}_r\in {\mathrm {Ker}}\left( \ell ^{(\texttt {e})}_{\kappa -r} \right) {\widetilde{\mathbf{F}}}_r\) (\({\overline{\zeta }}\in {\mathrm {Ker}}\left( \ell ^{(\texttt {e})}_{\kappa -r} \right) \)) and \({\widetilde{\mathbf{U}}}_{\kappa }^{(2)} \in \widehat{\mathcal {R}}_{\kappa }^{\mathbf{t}}\).
Let us denote \({\widetilde{\mathbf{F}}}\) the reversible part of \({{\mathbf{F}}}\) and define \({\widetilde{\mathbf{V}}}:= {\widetilde{\mathbf{U}}}- ({\overline{\rho }}+{\overline{\zeta }}){\widetilde{\mathbf{F}}}\), where
From Lemma 5, we obtain:
To complete the proof, it is enough to observe that:
\(\blacksquare \)
Proposition 23
Let us assume that the vector field \({\mathbf{F}}\) given in (6) is \(R_x\)-reversible up to order \(r+N-1\), with \(N\in \mathbb {N}\). Then, for all \({\widetilde{\mathbf{U}}}\in \bigoplus _{j=1}^N \widehat{\mathcal {R}}_j^{\mathbf{t}}\) and \({\widetilde{\mu }}\in \bigoplus _{j=1}^N \widehat{\mathcal {O}}_j^{\mathbf{t}}\) such that \(({\widetilde{\mathbf{U}}},{\widetilde{\mu }})\) belongs to the domain of \(\overline{\mathcal {NL}}^{(N)}\), there exists \(\lambda \in \bigoplus _{j=1}^{N}\mathcal {P}_j^{\mathbf{t}}\) such that
Proof
We use induction on N. If \(N=1\), the result is trivial. Moreover
and it is enough to take \(\lambda _1={\widetilde{\mu }}_1\).
Now, let us consider \(N>1\) and assume that there exists \(\lambda = \sum _{j=1}^{N-1} \lambda _j\), such that
The hypothesis \(\mathcal {J}^{r+N-1} \left( \overline{{\widetilde{\mathbf{U}}}{\, }_{ **}{\, }\left( (1+{\widetilde{\mu }}){\mathbf{F}} \right) } \right) ={\mathbf{0}}\) yields. If we take \(\lambda =\bar{\lambda }+\tilde{\lambda }\), where \(\bar{\lambda }=\sum _{j=1}^{N-1}\bar{\lambda }_j\), \(\bar{\lambda }_j\in \mathcal {E}_j^{\mathbf{t}}\) and \(\tilde{\lambda }=\sum _{j=1}^{N-1}\tilde{\lambda }_j\), \(\tilde{\lambda }_j\in \mathcal {O}_j^{\mathbf{t}}\) we obtain
Hence \(\mathcal {J}^{r+N-1} \left( \left[ {\mathbf{F}},{\widetilde{\mathbf{U}}} \right] +{\widetilde{\lambda }}{\mathbf{F}} \right) = {\mathbf{0}}\) and from Lemma 21 item (a), the statement holds for the value N. Consequently, we have
\(\blacksquare \)
Next result shows that \({\mathrm {Range}}\left( \overline{\mathcal {NL}}^{(N)} \right) \subset {\mathrm {Range}}\left( \overline{\mathcal {L}}^{(N)} \right) \).
Proposition 24
Let us assume that the vector field \({\mathbf{F}}\) given in (6) is \(R_x\)-reversible up to order \(r+N-1\), with \(N\in \mathbb {N}\). Then for all generators \(({\widetilde{\mathbf{U}}},{\widetilde{\mu }})\), belonging to the domain of the nonlinear operator \(\overline{\mathcal {NL}}^{(N)}\), there exists \(({\widetilde{\mathbf{V}}},{\widetilde{\nu }})\), belonging to the domain of the linear operator \(\overline{\mathcal {L}}^{(N)}\), such that \(\overline{\mathcal {NL}}^{(N)}\left( {\widetilde{\mathbf{U}}},{\widetilde{\mu }} \right) =\overline{\mathcal {L}}^{(N)} \left( {\widetilde{\mathbf{V}}},{\widetilde{\nu }} \right) \).
Proof
By Proposition 23 there exists \(\lambda \in \bigoplus _{j=1}^{N}\mathcal {P}_j^{\mathbf{t}}\) such that
If we take \(\lambda =\bar{\lambda }+\tilde{\lambda }\), where \(\bar{\lambda }=\sum _{j=1}^{N-1}\bar{\lambda }_j\), \(\bar{\lambda }_j\in \mathcal {E}_j^{\mathbf{t}}\) and \(\tilde{\lambda }=\sum _{j=1}^{N-1}\tilde{\lambda }_j\), \(\tilde{\lambda }_j\in \mathcal {O}_j^{\mathbf{t}}\) we obtain
As \({\mathbf{F}}\) is \(R_x\)-reversible up to order \(r+N-1\), from Lemma 5(a) we obtain \(\mathcal {J}^{r+N} \left( \overline{(1+\bar{\lambda }){\mathbf{F}}} \right) ={\mathbf{0}}\) and by Proposition 22 there exist \({\widetilde{\mathbf{V}}}\in \bigoplus _{j=1}^{N}\widehat{\mathcal {R}}_j^{\mathbf{t}}\) and \({\widetilde{\nu }}\in \bigoplus _{j=1}^{N}\widehat{\mathcal {O}}_j^{\mathbf{t}}\) such that
Hence \(\mathcal {J}^{r+N} \left( \overline{\left[ {\mathbf{F}},{\widetilde{\mathbf{U}}} \right] +(1+{\widetilde{\mu }}){\mathbf{F}}} \right) =\mathcal {J}^{r+N} \left( \overline{\left[ {\mathbf{F}},{\widetilde{\mathbf{V}}} \right] +{\widetilde{\nu }}{\mathbf{F}}} \right) \).
The hypothesis \(\mathcal {J}^{r+N-1} \left( \overline{{\widetilde{\mathbf{U}}}{\, }_{ **}{\, }\left( (1+{\widetilde{\mu }}){\mathbf{F}} \right) } \right) ={\mathbf{0}}\) yields, since \((\tilde{{\mathbf{U}}},\tilde{\mu })\) belongs to the domain of \(\overline{\mathcal {NL}}^{(N)}\). Therefore \(\mathcal {J}^{r+N-1}\left( \overline{\left[ {\mathbf{F}},{\widetilde{\mathbf{V}}} \right] +{\widetilde{\nu }}{\mathbf{F}}} \right) ={\mathbf{0}}\) and we prove that \(({\widetilde{\mathbf{V}}},{\widetilde{\nu }})\) belongs to the domain of \({\overline{\mathcal {L}}^{(N)}_{\scriptstyle {\left\{ {\widetilde{\mathbf{F}}}_r,\dots ,{\widetilde{\mathbf{F}}}_{r+N-1} \right\} }}}\), and \(\overline{\mathcal {NL}}^{(N)}\left( {\widetilde{\mathbf{U}}},{\widetilde{\mu }} \right) =\overline{\mathcal {L}}^{(N)}\left( {\widetilde{\mathbf{V}}},{\widetilde{\nu }} \right) \).
\(\blacksquare \)
In the following result, we prove that \({\mathrm {Range}}\left( \overline{\mathcal {L}}^{(N)} \right) \subset {\mathrm {Range}}\left( \overline{\mathcal {NL}}^{(N)} \right) \).
Proposition 25
Let us assume that the vector field \({\mathbf{F}}\) given in (15) is \(R_x\)-reversible up to order \(r+N-1\), with \(N\in \mathbb {N}\). Then for all generators \(\left( {\widetilde{\mathbf{V}}},{\widetilde{\nu }} \right) \) belonging to the domain of \(\overline{\mathcal {L}}^{(N)}\), there exists \(\left( {\widetilde{\mathbf{U}}},{\widetilde{\mu }} \right) \) belonging to the domain of the nonlinear operator \(\overline{\mathcal {NL}}^{(N)}\) such that \(\overline{\mathcal {L}}^{(N)}(\tilde{{\mathbf{V}}},\tilde{\nu })=\overline{\mathcal {NL}}^{(N)}({\widetilde{\mathbf{U}}},{\widetilde{\mu }})\).
Proof
We will prove that
As \(\left( {\widetilde{\mathbf{V}}},{\widetilde{\nu }} \right) \) belongs to the domain of the operator \(\overline{\mathcal {L}}^{(N)}_{\scriptstyle {\left\{ {\widetilde{\mathbf{F}}}_r,\dots ,{\widetilde{\mathbf{F}}}_{r+N-1} \right\} }}\), then we obtain \(\mathcal {J}^{r+N-1}\left( \overline{[{\mathbf{F}},{\widetilde{\mathbf{V}}}]+{\widetilde{\nu }}{\mathbf{F}}} \right) ={\mathbf{0}}\). Also, as \({\mathbf{F}}\) is \(R_x\)-reversible up to order \(r+N-1\), from Lemma 5 items (b) and (c), we get \(\mathcal {J}^{r+N-1}\left( [{\mathbf{F}},{\widetilde{\mathbf{V}}}]+{\widetilde{\nu }}{\mathbf{F}} \right) ={\mathbf{0}}\).
From Lemma 21 item (b), for \(m=N\), \(\lambda ={\widetilde{\nu }}\) there exists \(\mu =\sum _{j=1}^{N}\mu _j\), with \(\mu _j\in \mathcal {P}_j^{\mathbf{t}}\), such that
In particular, as \({\mathbf{F}}\) is \((N-1)\)-\(R_x\)-reversible, we obtain \(\mathcal {J}^{r+N-1}\left( {\overline{\mathbf{F}}} \right) ={\mathbf{0}}\) and
By Proposition 8 there exist \({\widetilde{\mathbf{U}}}\in \bigoplus _{j=1}^N\widehat{\mathcal {R}}_j^t\) and \({\widetilde{\mu }}\in \bigoplus _{j=1}^N\widehat{\mathcal {O}}_{j}^{\mathbf{t}}\) such that
Therefore \(({\widetilde{\mathbf{U}}},{\widetilde{\mu }})\) belongs to the domain of nonlinear operator \(\overline{\mathcal {NL}}^{(N)}\). On the other hand
\(\blacksquare \)
Rights and permissions
About this article
Cite this article
Algaba, A., Checa, I., Gamero, E. et al. Characterizing Orbital-Reversibility Through Normal Forms. Qual. Theory Dyn. Syst. 20, 38 (2021). https://doi.org/10.1007/s12346-021-00478-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12346-021-00478-6