Fake saddles and their transition maps 0 0 footnotetext: 2020 AMS Subject Classification : 34D10, 34D20; Secondary: 34C05. 0 0 footnotetext: Key words and phrases : singularities, Poincaré and Dulac transition maps, asymptotic expansion, uniform flatness, stability 0 0 footnotetext: This work has be
David Marín · 2025
Read Fake saddles and their transition maps 0 0 footnotetext: 2020 AMS Subject Classification : 34D10, 34D20; Secondary: 34C05. 0 0 footnotetext: Key words and phrases : singularities, Poincaré and Dulac transition maps, asymptotic expansion, uniform flatness, stability 0 0 footnotetext: This work has be by David Marín (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-2509-12941.txt.
Introduction
David Marin studies fake saddles in 2025, odd rest points of flat flow fields whose linear part is zero. These points sit inside a worn flow box. They are hard to class because first-order tools give no grip.
Marin classes those with a non-zero second step. He gives the first term of a shared growth form for the return map. The map links two cross cuts of the matching fiber. Its sign settles whether the point holds or sheds nearby paths. The result adds a firm step for flat field work near worn points.
Catalog record
| CREATOR | David Marín |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-12941
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Fake saddles and their transition maps 0 0 footnotetext: 2020 AMS Subject Classification : 34D10, 34D20; Secondary: 34C05. 0 0 footnotetext: Key words and phrases : singularities, Poincaré and Dulac transition maps, asymptotic expansion, uniform flatness, stability 0 0 footnotetext: This work has been partially funded by the Ministry of Science, Innovation and Universities of Spain through the grant PID2021-125625NB-I00 and by the Agency for Management of University and Research Grants of Catalonia through the grant 2021SGR01015. David Marín Abstract We study degenerate singular points of planar vector fields inside a (degenerate) flow box. These kind of singularities are called fake saddles and their linear parts are always zero. We characterize fake saddles with non-zero second-order jet and we give the first term of a uniform asymptotic expansion of the Poincaré map between two transverse sections to their corresponding singular fiber, determining its stability. 1 Introduction and statements of main results Following [ 6 , 7 ] we define a fake saddle as a singular point having exactly two separatrices which are contained in a smooth invariant curve separating two hyperbolic sectors, see Figure 1 . This kind of singularities are also known as impassable grains. Another way to think about a fake saddle is that the corresponding vector field can be put in a form that is like a degenerate flow box, i.e. near the singularity the phase portrait consists of parallel fibers, all but one of which have no singular points, and the singular fiber (the union of the singular point with its two separatrices) has a semi-stable equilibrium point, see [ 2 ] . Taking two transverse sections \Sigma_{\alpha} and \Sigma_{\omega} to the singular fiber outside the singular point there is a transition map \Pi_{\alpha}^{\omega}:\Sigma_{\alpha}\to\Sigma_{\omega} which is well-defined on both sides of the singular fiber. The local phase portrait of a singularity with non-zero linear part (for the nilpotent case see for instance [ 1 , Theorem 3.5] ) is well known and prevents the singular point from being a fake saddle. Notice that the smoothness of the singular fiber is strictly necessary because the Hamiltonian vector field y\partial_{x}+x^{2}\partial_{y} has a nilpotent singularity at the origin with exactly two separatrices that are contained in the cuspidal curve y^{2}-\frac{2}{3}x^{3}=0 separating two hyperbolic sectors. The first non-trivial and generic case of a fake saddle arises when the second-order jet is non-zero. Our objective is to characterize these generic fake saddles and to analyze their transition maps in order to determine whether the behavior is attractive or repulsive on each side of the singular fiber. The techniques we employ allow us to derive the leading term of an asymptotic expansion that is uniform with respect to parameters. Figure 1: Fake saddle and the associated transition maps. This work is mainly motivated by the paper [ 2 ] , which provides a normal form for generic fake saddles that we recall in the sequel. First we can locate the fake saddle at the origin (0,0) . Secondly, if the vector field is at least of class \mathscr{C}^{3} there exists a local diffeomorphism that rectifies the germ of singular fiber, so that it is contained in the line y=0 in a certain coordinate system (x,y) , see [ 2 , Lemma 2.1] . Thanks to the hypothesis that the second-order jet of the fake saddle is non-zero, after a suitable rescaling of the coordinates we can assume that the fake saddle is defined by a system of differential equations \left\{\begin{array}[]{l}\dot{x}=x^{2}+axy+y^{2}+\mathrm{O}_{3}(x,y),\\ \dot{y}=y\big(cx+by+\mathrm{O}_{2}(x,y)\big),\end{array}\right. where a,b,c\in\mathbb{R} . This normal form motivates us to consider a smooth family of planar vector fields \{X_{\mu}\}_{\mu\in\Omega\subset\mathbb{R}^{n}} having the form X_{\mu}(x,y)=(x^{2}f_{1}(x,y;\mu)+a(\mu)xy+y^{2}f_{2}(x,y;\mu))\partial_{x}+\big(xg_{1}(x,y;\mu)+yg_{2}(y;\mu)\big)y\partial_{y}, (1) where f_{1}(x,y;\mu) , f_{2}(x,y;\mu) , g_{1}(x,y;\mu) and g_{2}(y;\mu) are \mathscr{C}^{K} -functions fulfilling f_{1}(0,0;\mu)=f_{2}(0,0;\mu)=1 . We consider the following invariants a(\mu),\quad b(\mu):=g_{2}(0;\mu),\quad c(\mu):=g_{1}(0,0;\mu) (2) of the family ( 1 ) and define the associated value d(\mu):=4(1-c(\mu))-(a(\mu)-b(\mu))^{2} , which will play a key role in the sequel. In this paper we extend the main theorems of [ 2 ] and we give simpler proofs of them using the results stated in [ 4 , 5 ] . Our first result gives a characterization of generic fake saddles inside the family ( 1 ), generalizing [ 2 , Theorem A] which only treats the case a=0 . Theorem A . If the invariants (a,b,c) given in ( 2 ) of a vector field X in the family ( 1 ) do not belong to \{d=0\}\cap\{a^{2}-b^{2}=4\} then the origin is a fake saddle if and only if, either d>0 , or c=1 and a=b . In both situations after blowing up the origin we have a single singular point on the exceptional divisor, which is a hyperbolic saddle of hyperbolicity ratio 1-c>0 in the first one and a semi-hyperbolic saddle in the second one. The hypothesis (a,b,c)\notin\{d=0\}\cap\{a^{2}-b^{2}=4\} in Theorem A (which is always verified when a=0 ) can not be removed as Example 3 will show. Let \Omega be an open subset of \mathbb{R}^{n} and let W any subset of \Omega . We recall (cf. [ 4 , Definition 1.2] ) that a \mathscr{C}^{K} -function f(s;\mu) defined in the intersection of (0,+\infty)\times\Omega with an open neighborhood of \{0\}\times\Omega\subset\mathbb{R}^{n+1} belongs to the flat class \mathcal{F}_{L}^{K}(W) if for each \mu_{0}\in W and \nu=(\nu_{0},\ldots,\nu_{n})\in\mathbb{N}^{n+1} with |\nu|=\nu_{0}+\cdots+\nu_{n}\leq K there exists \varepsilon>0 such that |\partial_{s}^{\nu_{0}}\partial_{\mu_{1}}^{\nu_{1}}\cdots\partial_{\mu_{n}}^{\nu_{n}}f(s;\mu)|\leq Cs^{L-\nu_{0}} for every s\in(0,\varepsilon) and \mu\in\mathbb{B}_{\varepsilon}(\mu_{0})\cap\Omega . We consider the transverse sections \Sigma_{\star}=\{x=\star\} to y=0 , for \star\in\{\alpha,\omega\} , where \alpha<0<\omega . We assume that f_{1}(x,0)>0 for all x\in[\alpha,\omega] and we consider the transition map \Pi_{\alpha}^{\omega}:\Sigma_{\alpha}\to\Sigma_{\omega} . Our second result extends [ 2 , Theorem B] , which only treats the case a=b=0 , \partial_{y}f_{1}=\partial_{y}g_{1}=0 , f_{2}=1 and g_{2}=0 , when the transition map is restricted to y>0 . In that theorem the authors characterize the stability of the transition map from the sign of the Cauchy principal value \mathrm{PV}\int_{\alpha}^{\omega}\frac{g_{1}(x,0;\mu)}{xf_{1}(x,0;\mu)}dx:=\lim\limits_{\varepsilon\to 0^{+}}\left(\int_{\alpha}^{-\varepsilon}\frac{g_{1}(x,0;\mu)}{xf_{1}(x,0;\mu)}dx+\int_{\varepsilon}^{\omega}\frac{g_{1}(x,0;\mu)}{xf_{1}(x,0;\mu)}dx\right). Theorem B . Assume that \mu_{0}\in\Omega and that the invariants ( 2 ) of the vector field X_{\mu_{0}} in ( 1 ) belong to \{d=4(1-c)-(a-b)^{2}>0\} . If \mu\approx\mu_{0} then the origin is a fake saddle of X_{\mu} and its transition map \Pi_{\alpha}^{\omega}:\Sigma_{\alpha}\to\Sigma_{\omega} satisfies \Pi_{\alpha}^{\omega}(y;\mu)=e^{\gamma_{\pm}(\mu)}y+\mathcal{F}_{1+\epsilon}^{K}(\{d>0\}) on \pm y\geq 0 , where \epsilon>0 and \gamma_{\pm}(\mu)=\mathrm{PV}\int_{\alpha}^{\omega}\frac{g_{1}(x,0;\mu)}{xf_{1}(x,0;\mu)}dx\pm\frac{\pi(2b(\mu)-c(\mu)(a(\mu)+b(\mu)))}{\sqrt{d(\mu)}}. Using that f_{1}(0,0)=1 and g_{1}(0,0)=c we can explicitly compute the previous Cauchy principal value using a convergent integral: \mathrm{PV}\int_{\alpha}^{\omega}\frac{g_{1}(x,0)}{xf_{1}(x,0)}dx=c\log\left|\frac{\omega}{\alpha}\right|+\int_{\alpha}^{\omega}\left(\frac{g_{1}(x,0)}{f_{1}(x,0)}-c\right)\frac{dx}{x}. \square Clearly the sign of \gamma_{\pm}(\mu) determines the stability of the singular fiber y=0 on the side \pm y>0 . In the case a=b=0 the two values \gamma_{\pm}(\mu) coincide with the principal value of the integral. Otherwise the stability of the two sides \pm y>0 can be different, see Example 3 . \square It is worth to be noticed that the writing of \Pi_{\alpha}^{\omega}(y;\mu) in the statement implies that it is of the form e^{\gamma_{\pm}(\mu)}y+\mathrm{o}(y) but also that the remainder term \mathrm{o}(y) is uniform with respect to the parameter \mu . This uniformity allows to address cyclicity problems and not just study stability. For clarity in the exposition we will omit the dependence on \mu when it is not essential. 2 Proofs of the main results Proof of Theorem A . A straightforward computation shows that Y=\frac{(u,uv)^{*}X}{u}=P(u,v)u\partial_{u}+Q(u,v)v\partial_{v} with P(0,0)=1 and Q(0,v)=-v^{2}+(b-a)v+c-1 . The point (u,v)=(0,0) is always a singular point of Y . If d<0 then Y has two