A Comparison Principle for Bifurcation of Periodic Solutions of Hamiltonian Systems

Helene Cyris, Joanna Janczewska, Nils Waterstraat · 2025

Read A Comparison Principle for Bifurcation of Periodic Solutions of Hamiltonian Systems by Helene Cyris, Joanna Janczewska, Nils Waterstraat (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-24835.txt.

Introduction

Helene Cyris, Joanna Janczewska, and Nils Waterstraat give new criteria for local bifurcation of periodic answers in Hamiltonian systems, based on comparing spectral flow. Bifurcation means new periodic answers branch off as a setting changes, a central question for vibrating and orbital systems. Their contrast principle reduces detection to inspecting the system's coefficients.

Instead of heavy computation, a simple look at coefficients can reveal where new answers appear. Readers in dynamic systems get a useful test for Hamiltonian bifurcation. The method ties an abstract topological quantity, spectral flow, to concrete data written in the equations themselves.

Catalog record

CREATORHelene Cyris, Joanna Janczewska, Nils Waterstraat
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-24835

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Comparison Principle for Bifurcation of Periodic Solutions of Hamiltonian Systems Helene Cyris, Joanna Janczewska, Nils Waterstraat Abstract We obtain novel criteria for the existence of local bifurcation for periodic solutions of Hamiltonian systems by a comparison principle of the spectral flow. Our method allows to find the appearance of new solutions by a simple inspection of the coefficients of the system. 1 1 footnotetext: 2020 Mathematics Subject Classification: Primary 37J20; Secondary 37J46, 58J30 1 Introduction Let \mathcal{H}:I\times\mathbb{R}\times\mathbb{R}^{2n}\rightarrow\mathbb{R} be a C^{2} -map, where I=[0,1] denotes the unit interval, and assume that \mathcal{H}(\lambda,t,u) is 2\pi -periodic with respect to t as well as \nabla_{u}\mathcal{H}(\lambda,t,0)=0 for all (\lambda,t)\in I\times\mathbb{R} . We consider the Hamiltonian systems \left\{\begin{aligned} Ju^{\prime}(t)+\nabla_{u}&\mathcal{H}(\lambda,t,u(t))=0,\quad t\in[0,2\pi]\\ u(0)&=u(2\pi),\end{aligned}\right. (1) where J denotes the standard symplectic matrix \displaystyle J=\begin{pmatrix}0&-I_{n}\\ I_{n}&0\end{pmatrix}, (2) and note that u\equiv 0 is a solution of ( 1 ) for all \lambda\in I under the given assumptions. We say that \lambda_{0}\in I is a bifurcation point for ( 1 ), if in every neighbourhood of (\lambda_{0},0) in I\times C^{1}(S^{1},\mathbb{R}^{2n}) there is an element (\lambda,u) such that u\neq 0 is a solution of ( 1 ). There have been numerous attempts to study the existence of bifurcation points in this setting (cf., e.g., [ 11 ] , [ 9 ] , [ 2 ] and the references therein), and here we focus on an approach via the spectral flow that was introduced in [ 5 ] and applied to ( 1 ) in [ 6 ] and [ 10 ] . The spectral flow is defined for any path L=\{L_{\lambda}\}_{\lambda\in I} of selfadjoint Fredholm operators on a Hilbert space and, roughly speaking, it counts the net number of negative eigenvalues of L_{0} that become positive whilst the parameter \lambda traverses the interval I . In particular, a non-vanishing spectral flow means that there is an eigenvalue of L_{0} that crosses 0 for some \lambda in I . The application of the spectral flow in bifurcation theory can be outlined as follows. Let H be a real Hilbert space and f:I\times H\rightarrow\mathbb{R} a C^{2} -map such that 0\in H is a critical point of all f_{\lambda}:H\rightarrow\mathbb{R} , i.e., \nabla f_{\lambda}(0)=0 , \lambda\in I . We say that \lambda^{\ast}\in I is a bifurcation point of the equation \displaystyle\nabla f_{\lambda}(u)=0 (3) if in any neighbourhood of (\lambda^{\ast},0) in I\times H there is some solution (\lambda,u) of ( 3 ) such that u\neq 0 . The Hessian L_{\lambda} of f_{\lambda} at the critical point 0 is a selfadjoint operator on H , and L_{\lambda^{\ast}} has to be non-invertible if \lambda^{\ast} is a bifurcation point of ( 3 ). If we require in addition that the operators L_{\lambda} are Fredholm, then the spectral flow of the path L=\{L_{\lambda}\}_{\lambda\in I} is defined and its non-triviality indicates that there is some \lambda^{\ast}\in I such that L_{\lambda^{\ast}} has a non-trivial kernel made by an eigenvalue that crosses 0 when \lambda passes \lambda^{\ast} . The main theorem of [ 5 ] is as follows: Theorem 1.1 . Let H be a real separable Hilbert space and f:I\times H\rightarrow\mathbb{R} a C^{2} -map such that 0\in H is a critical point of all functionals f_{\lambda}:H\rightarrow\mathbb{R} . Let the Hessians L_{\lambda} of f_{\lambda} at 0 be Fredholm operators and assume that L_{0} and L_{1} are invertible. If \sfl(L)\neq 0\in\mathbb{Z} for the path L=\{L_{\lambda}\}_{\lambda\in I} , then there is a bifurcation point for ( 3 ) in (0,1) . Let us note that the spectral flow of L is just the difference of the Morse indices of L_{0} and L_{1} if the latter are finite. Thus Theorem 1.1 yields the well-known existence of a bifurcation point when the Morse indices are finite and change from L_{0} to L_{1} (cf. [ 3 ] , [ 9 ] ). Since the work [ 5 ] the spectral flow has been applied to study bifurcation problems for various kinds of differential equations with infinite Morse indices by several authors. Unfortunately, the spectral flow often is notoriously difficult to compute if possible at all. Pejsachowicz and the third author invented in [ 10 ] a comparison principle that can be used to estimate the spectral flow, which can be good enough to show that it does not vanish and so makes Theorem 1.1 applicable (cf. [ 17 ] ). For example, it allows to find bifurcation points for ( 1 ) just from the eigenvalues of the family of Hessians of the Hamiltonian \mathcal{H} at 0\in\mathbb{R}^{2n} , which we now want to outline. Let H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n}) be the linear space of all functions u:[0,2\pi]\rightarrow\mathbb{R}^{2n} such that \displaystyle u(t)=c_{0}+\sum^{\infty}_{k=1}{(a_{k}\sin(kt)+b_{k}\cos(kt))}, (4) where c_{0},a_{k},b_{k}\in\mathbb{R}^{2n} , k\in\mathbb{N} , and \sum^{\infty}_{k=1}{k(|a_{k}|^{2}+|b_{k}|^{2})}<\infty (cf. [ 1 ] ). This is a Hilbert space with respect to the scalar product \displaystyle\langle u,v\rangle_{H^{\frac{1}{2}}}=2\pi\langle c_{0},\tilde{c}_{0}\rangle+\pi\sum^{\infty}_{k=1}{k(\langle a_{k},\tilde{a}_{k}\rangle+\langle b_{k},\tilde{b}_{k}\rangle)}, (5) where \tilde{c}_{0} and \tilde{a}_{k},\tilde{b}_{k} are the Fourier coefficients of v\in H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n}) as in ( 4 ). Let now \displaystyle Q:H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n})\times H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n})\rightarrow\mathbb{R} be the continuous extension of the bounded bilinear form \displaystyle\tilde{Q}(u,v)=\int^{2\pi}_{0}{\langle Ju^{\prime}(t),v(t)\rangle\,dt} (6) from the dense subspace H^{1}(S^{1},\mathbb{R}^{2n}) to H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n}) , where we recall that H^{1}(S^{1},\mathbb{R}^{2n}) consists of all absolutely continuous functions u:[0,2\pi]\rightarrow\mathbb{R}^{2n} having a square integrable derivative. We consider the map \displaystyle f:I\times H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n})\rightarrow\mathbb{R},\quad f_{\lambda}(u)=\frac{1}{2}\,Q(u,u)+\int^{2\pi}_{0}{\mathcal{H}(\lambda,t,u(t))\,dt}. (7) It is a well-known result (see [ 11 ] , [ 4 ] ) that f is C^{2} if there are constants a,b\geq 0 and r>1 such that \displaystyle\begin{split}|\nabla_{u}\mathcal{H}(\lambda,t,u)|&\leq a+b|u|^{r},\\ |D_{u}\nabla_{u}\mathcal{H}(\lambda,t,u)|&\leq a+b|u|^{r},\quad(\lambda,t,u)\in I\times\mathbb{R}\times\mathbb{R}^{2n}.\end{split} Moreover, \langle\nabla_{u}f_{\lambda},v\rangle_{H^{\frac{1}{2}}}=Q(u,v)+\int^{2\pi}_{0}{\langle\nabla_{u}\mathcal{H}(\lambda,t,u(t)),v(t)\rangle\,dt},\quad v\in H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n}), and thus the critical points of f_{\lambda} are the weak solutions of the Hamiltonian system ( 1 ). In particular, \ u\equiv 0\in H^{\frac{1}{2}}(S^{1},\mathbb{R}^{2n}) is a critical point of f_{\lambda}, and the associated Hessian is given by \displaystyle\langle L_{\lambda}u,v\rangle_{H^{\frac{1}{2}}}=Q(u,v)+\int^{2\pi}_{0}{\langle A_{\lambda}(t)u(t),v(t)\rangle\,dt}, (8) where A_{\lambda}(t) is the Hessian matrix of the functional \mathcal{H}_{\lambda}(t,\cdot):\mathbb{R}^{2n}\rightarrow\mathbb{R} at 0 . Note that the kernel of L_{\lambda} consists of the solutions of the linear equation \left\{\begin{aligned} Ju^{\prime}(t)&+A_{\lambda}(t)u(t)=0,\quad t\in[0,2\pi]\\ u(0)&=u(2\pi).\end{aligned}\right. (9) The Hessians L_{\lambda} are Fredholm operators (cf. [ 14 , Lemma 3.1] , [ 16 ] ) and thus we see from Theorem 1.1 that there is a bifurcation point of ( 1 ) if ( 9 ) only has the trivial solution for \lambda=0,1 and \sfl(L)\neq 0 . As said before, obtaining the number \sfl(L) usually is a challenging task (cf. [ 6 ] ) but estimating its value can be enough to show its non-triviality and thus to make Theorem 1.1 applicable. This was done in a joint work of Pejsachowicz and the third author [ 10 ] as follows (cf. [ 8 ] ). We let \mu_{1}(\lambda,t)\leq\cdots\leq\mu_{2n}(\lambda,t) be the eigenvalues of the matrices A_{\lambda}(t) for \lambda\in I , t\in[0,2\pi] , and set \displaystyle\alpha_{\lambda}:=\inf_{t\in[0,2\pi]}{\mu_{1}(\lambda,t)} (10) as well as \displaystyle\beta_{\lambda}:=\sup_{t\in[0,2\pi]}{\mu_{2n}(\lambda,t)}. (11) Thus \alpha_{\lambda} is a lower bound for the spectra of the matrix family \{A_{\lambda}(t)\}_{t\in[0,2\pi]} and \beta_{\lambda} is an upper bound. Theorem 1.2 . Let \mathcal{H}:I\times\mathbb{R}\times\mathbb{R}^{2n}\rightarrow\mathbb{R} be a family of Hamiltonians as above as well as \alpha_{1},\beta_{0} as in ( 10 ) and ( 11 ). Assume that ( 9 ) only has the trivial solution for \lambda=0 and \lambda=1 . (i) If there is some k\in\mathbb{Z} such that \beta_{0}<k\leq\alpha_{1} , then there is a bifurcation point of ( 1 ). (ii) If there are only finitely many \lambda\in I such that ( 9 ) has a non-trivial solution, then the

Read the complete text