Birth and metamorphoses of resonances in the driven van der Pol oscillator
Jan Kyzioł, Andrzej Okniński · 2026
Read Birth and metamorphoses of resonances in the driven van der Pol oscillator by Jan Kyzioł, Andrzej Okniński (2026) 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-2605-26299.txt.
Introduction
Jan Kyzioł and Andrzej Okniński investigate the driven van der Pol oscillator, a classic nonlinear system, tracking how its resonances are born and change shape. They focus on one-to-one and one-to-three resonances, where the oscillator's rhythm locks to the driving force in simple ratios. Their analysis uses differential properties of implicit functions describing how response amplitude varies with frequency. The study maps the birth and metamorphoses of these resonances within that formalism.
Catalog record
| CREATOR | Jan Kyzioł, Andrzej Okniński |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 16458 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.26299
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Birth and metamorphoses of resonances in the driven van der Pol oscillator Jan Kyzioł, Andrzej Okniński Abstract The dynamics of the driven van der Pol oscillator are investigated. We study birth and metamorphoses of 1 : 1 1:1 and 1 : 3 1:3 resonances within the formalism of differential properties of amplitude-frequency response implicit functions. 1 Introduction We study the forced van der Pol equation in nondimensional form \dfrac{d^{2}z}{d\tau^{2}}-\mu\left(1-z^{2}\right)\dfrac{dz}{d\tau}+z=F\cos\left(\Omega\tau\right), (1) where \tau is a nondimensional time and \mu is a positive damping parameter. Equation ( 1 ) was introduced by van der Pol to describe vacuum-tube oscillations [ 1 ] (see also [ 2 ] for the non-forced equation). Investigations of vacuum-tube oscillations led to the early discovery of deterministic chaos, noticed by van der Pol and van der Mark in an electrical circuit described by Eq. ( 1 ) [ 3 ] . Equation ( 1 ) was investigated further by Cartwright and Littlewood [ 4 ] to confirm the presence of chaotic dynamics. The van der Pol oscillator displays several generic nonlinear phenomena; see for global bifurcation phenomena Refs. [ 5 , 6 , 7 , 8 , 9 , 10 ] . The first application of the van der Pol oscillators outside electrical engineering, carried out by van der Pol and van der Mark, was to describe the heartbeat [ 11 , 12 ] . Then, the van der Pol equation has been used to model self-sustained oscillations in physics, mechanics, engineering, and biology. In this paper, we study birth and metamorphoses of 1:1 and 1:3 resonances in the framework of the differential properties of amplitude-frequency response implicit functions [ 14 , 15 ] , continuing our previous work on van der Pol-Duffing equation [ 16 , 17 ] . 2 Nonlinear resonances via KBM method We apply the Krylov-Bogoliubov-Mitropolsky (KBM) perturbation approach [ 13 ] to Eq. ( 1 ). We assume a resonance in form z_{\alpha}\left(\tau\right)=A\cos\left(\alpha\Omega\tau+\varphi\right)+\varepsilon z_{1}\left(A,\varphi,\tau\right)+\ldots, (2) where the amplitude A and frequency \Omega fulfil the amplitude-frequency response equation \mathcal{F}_{\alpha}\left(A,\Omega,\mu,F\right)=0. (3) In what follows we consider \alpha=1 and \alpha=\frac{1}{3} for 1:1 and 1:3 resonances, respectively. In the case of primary resonance, we have \begin{array}[]{lll}\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)&=&A^{2}\left(\mu^{2}\Omega^{2}\left(1-\tfrac{1}{4}A^{2}\right)^{2}+\left(1+\tfrac{3}{4}\lambda A^{2}-\Omega^{2}\right)^{2}\right)-F^{2},\end{array} (4) while for the 1:3 resonance we get \begin{array}[]{lll}\mathcal{F}_{1:3}\left(A,\Omega,\mu,F\right)&=&\Omega^{4}\mu^{4}\left(A^{2}-4+\tfrac{2F^{2}}{\left(\Omega^{2}-1\right)^{2}}\right)^{2}+\Omega^{2}\mu^{2}\left(\tfrac{4}{3}\Omega^{2}-12\right)^{2}\\ &&-\tfrac{\Omega^{4}A^{2}F^{2}\mu^{4}}{\left(\Omega^{2}-1\right)^{2}}\quad\left(\Omega\neq 1\right).\end{array} (5) 3 Singular and critical points of the amplitude profiles Singular points of an implicit function L\left(A,\Omega,\underline{c}\right)=0 , where \underline{c}=c_{1},c_{2},\ldots are parameters, are solutions of equations \displaystyle L\left(A,\Omega,\underline{c}\right) \displaystyle= \displaystyle 0, (6a) \displaystyle\frac{\partial L\left(A,\Omega,\underline{c}\right)}{\partial\Omega} \displaystyle= \displaystyle 0, (6b) \displaystyle\frac{\partial L\left(A,\Omega,\underline{c}\right)}{\partial A} \displaystyle= \displaystyle 0. (6c) Solutions of Eqs. ( 6 ), if exist, are of form A=A_{\ast} , \Omega=\Omega_{\ast} , \underline{c}=\underline{c}_{\ast} . It follows from Eqs. ( 6b ) and ( 6c ), that in the neighborhood of a singular point \left(A_{\ast},\Omega_{\ast}\right) of the function L\left(A,\Omega,\underline{c}_{\ast}\right)=0 , neither of the functions \Omega=f\left(A\right) nor A=g\left(\Omega\right) is single-valued. Moreover, equations ( 6a ) and ( 6b ) are conditions for extremas of the function A=g\left(\Omega\right) , while equations ( 6a ), ( 6c ) are conditions for vertical tangencies of the function A=g\left(\Omega\right) or extremas of the function \Omega=f\left(A\right) . 3.1 The case of resonance 1:1 To compute singular points of the amplitude-frequency implicit function for the 1:1 resonance, we substitute in ( 6 ) L\left(A,\Omega,\underline{c}\right)=\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right) . Solutions of Eqs. ( 7 ) \displaystyle\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right) \displaystyle= \displaystyle 0 (7a) \displaystyle\frac{\partial\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)}{\partial A} \displaystyle= \displaystyle 0 (7b) \displaystyle\frac{\partial\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)}{\partial A} \displaystyle= \displaystyle 0 (7c) read A=2,\ \Omega=1,\ F=0,\quad\left(\mu\neq 0\right) (8) and \hskip-6.0pt\left.\begin{array}[]{l}5\mu^{2}A^{6}-44\mu^{2}A^{4}+\left(-192+112\mu^{2}\right)A^{2}+256-64\mu^{2}=0\\ 8\Omega^{2}+16\mu^{2}-8\mu^{2}A^{2}+\mu^{2}A^{4}-32=0\\ \left(\mu^{4}-5\mu^{2}\right)A^{4}+\left(-120+38\mu^{2}-8\mu^{4}\right)A^{2}-104\mu^{2}+160+16\mu=5^{3}F^{2}\end{array}\right\} (9) Solution ( 8 ) corresponds to an isolated point, while solution ( 9 ) represents self-intersections. Figure 1: Amplitude-frequency response functions \mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)=0 . \mu=0.2 ; F=0.05,0.10,0.153\,618,0.20,0.50 (LightBlue, LightGreen, LightRed, Green, Blue). LightRed dot marks an isolated point. Moreover, solutions of Eqs. ( 10 ) for vertical tangencies \displaystyle\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right) \displaystyle= \displaystyle 0 (10a) \displaystyle\frac{\partial\mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)}{\partial A} \displaystyle= \displaystyle 0 (10b) are \hskip-5.0pt\left.\begin{array}[]{l}2\mu^{4}A^{12}+\left(-3F^{2}\mu^{4}-16\mu^{4}\right)A^{10}+\left(28F^{2}\mu^{4}+32\mu^{4}\right)A^{8}\\ +\left(-80F^{2}\mu^{4}+32F^{2}\mu^{2}\right)A^{6}+\left(64F^{2}\mu^{4}-128F^{2}\mu^{2}\right)A^{4}+128F^{4}=0\\ 2A^{6}\mu^{2}-\left(3F^{2}\mu^{2}+8\mu^{2}\right)A^{4}+16\mu^{2}F^{2}A^{2}-16F^{2}\mu^{2}+32F^{2}=16F^{2}\Omega^{2}\end{array}\right\} (11) Figure 2: Amplitude-frequency response functions \mathcal{F}_{1:1}\left(A,\Omega,\mu,F\right)=0 . \mu=0.2 ; F=0.05,0.15 (LightBlue, Purple). LightRed dots mark singular points. Dashed lines show vertical tangencies. 3.2 The case of resonance 1:3 Since the implicit function \mathcal{F}_{1:3}\left(A,\Omega,\mu,F\right)=0 depends on A^{2} , \Omega^{2} , \mu^{2} , and F^{2} only we write Eq. ( 5 ) in a simpler form \hskip-4.0ptL_{1:3}\left(Y,X,m,f\right)=X^{2}m^{2}\left(Y-4+\tfrac{2f}{\left(X-1\right)^{2}}\right)^{2}+Xm\left(\tfrac{4}{3}X-12\right)^{2}-\tfrac{X^{2}Yfm^{2}}{\left(X-1\right)^{2}}, (12) where Y=A^{2} , X=\Omega^{2} , m=\mu^{2} , f=F^{2} , and L_{1:3}\left(A^{2},\Omega^{2},\mu^{2},F^{2}\right)=\mathcal{F}_{1:3}\left(A,\Omega,\mu,F\right) . Singular points of the implicit function L_{1:3}\left(Y,X,m,f\right)=0 are solutions of the following equations \displaystyle L_{1:3}\left(Y,X,m,f\right) \displaystyle= \displaystyle 0, (13a) \displaystyle\frac{\partial L_{1:3}\left(Y,X,m,f\right)}{\partial X} \displaystyle= \displaystyle 0, (13b) \displaystyle\frac{\partial L_{1:3}\left(Y,X,m,f\right)}{\partial Y} \displaystyle= \displaystyle 0. (13c) We note here that there are two kinds of singular points in these systems: isolated points and self-intersections. The isolated points are signatures of the birth of a new branch of solution, while self-intersections correspond to rupture of existing branches. Figure 3: Amplitude-frequency response functions L_{1:3}\left(Y,X,m,f\right)=0 . \mu=0.2 ; F=0.2 , 0.6 , 3 , 8.7 , 12 , 12.25 , 12.26 (LightBlue, LightGreen, Sienna, Purple, Brown, Green, Blue). LightRed dots mark isolated points. Dashed lines show vertical tangencies. . Moreover, vertical tangencies are computed from \displaystyle L_{1:3}\left(Y,X,m,f\right) \displaystyle= \displaystyle 0, (14a) \displaystyle\frac{\partial L_{1:3}\left(Y,X,m,f\right)}{\partial Y} \displaystyle= \displaystyle 0. (14b) Vertical tangencies show the range of a resonance, while self-intersections correspond to rupture of existing branches. We solve Eqs. ( 13 ) to get the following singular solutions X=9,Y=4,f=0,\text{\quad}\left(m\text{ arbitrary}\right) (15a) which is an isolated point, and another solution \left\{\begin{array}[]{l}Y=\tfrac{-5X^{3}+73X^{2}+12mX^{2}-243X-81}{3mX^{2}},\\ f=4\tfrac{-4354X^{3}+11\,732X^{2}+17\,010X+3969+315X^{4}+108X^{4}m-360X^{3}m-324mX^{2}}{315mX^{2}},\\ 175X^{5}+\left(-216m-3535\right)X^{4}+\left(720m+22\,498\right)X^{3}\\ \qquad\hskip 10.0pt+\left(648m-40\,194\right)X^{2}-31\,185X-5103=0,\end{array}\right. (15b) with arbitrary m . Condition that discriminant of equation for X in ( 15b ) vanishes yields m=-28 , m=0 , and \displaystyle 24\,564\,384m^{4}+1317\,273\,111m^{3}+104\,720\,016\,996m^{2} (16) \displaystyle-1475\,275\,114\,404m-178\,499\,175\,680=0. A positive solution of Eq. ( 16 ) is m_{cr}=12.\,009\,279 . Accordingly, equation for X in Eq. ( 15b ) has one positive root for m<m_{cr} , corresponding to an isolated point, and three positive roots for m>m_{cr} , corresponding to two isolated points and one self-intersection. To compute vertical tangencies, we solve equations ( 14 ) obtaining \displaystyle Y=\tfrac{8X^{2}-16X-3f+8}{2\left(1-X\right)^{2}}, (17a) \displaystyle 64X^{6}-1408X^{5}+10\,176X^{4}+\left(-27\,904-144mf\right)X^{3} (17b) \displaystyle+\left(35\,776+288mf\right)X^{2}+\left(63mf^{2}-21\,888-144mf\right)X+5184=0. In the case when two vertical tangencies overlap, we have a singular point – an isolated point. This happens when Eq. ( 17b ) has a double root, i.e. when a discriminant, D\left(m,f\right) , of Eq. ( 17b ) vanishes. Computing the discriminant, we get D=\left(\begin{array}[]{l}987\,614\,208m^{4}f^{4}-4514\,807\,808f^{3}m^{4}+5159\,780\,352f^{2}m^{4}\\ +27\,\allowbreak 205\,113\,600m^{3}f^{4}-30\,\allowbreak 059\,237\,376m^{3}f^{3}-1056\,\allowbreak 502\,185\,984f^{2}m^{3}\\ +761\,\allowbreak 014\,517\,760fm^{3}-472\,696\,875f^{5}m^{2}+96\,\allowbreak 981\,192\,000m^{2}f^{4}\\ -3830\,\allowbreak 568\,505\,344m^{2}f^{3}-11\,576\,\allowbreak 272\,748\,544m^{2}f^{2}-60\,542\,\allowbreak 932\,746\,240fm^{2}\\ -22\,265\,\allowbreak 110\,462\,464m^{2}-55\,\allowbreak 535\,398\,912f^{3}m-11\,861\,\allowbreak 638\,512\,640mf^{2}\\ +2635\,718\,\allowbreak 350\,340\,096mf+2216\,615\,\allowbreak 441\,596\,416m+377\,132\,\allowbreak 488\,327\,168f\\ -55\,169\,095\,\allowbreak 435\,288\,576\end{array}\right) (18) The plot of D\left(m,f\right)=0 is shown in Fig. 4 . Figure 4 shows values of parameters f , m for which the function L\left(Y,X,m,f\right)=0 has singular (isolated) points (red and blue curves). Values of f and m that lead to physical values for X and Y are on the red interval. Figure 4: Manifold of singular (isolated) points. Physical points lie on the LightRed interval. . 4 Computational results In this Section, we numerically solve Eq. ( 1 ) and compare the results with analytical predictions. 4.1 The case of resonance 1:1 The solution ( 8 ) of Eqs. ( 7 ) describes a well-known limit cycle of the unforced van der Pol equation [ 10 ] . This solution is a good approximation to the limit cycle for small \mu , while for large values of \mu a good approximation can be found in [ 18 , 19 ] . It follows from Figs. 1 and 2 that for F<0.153\,618 the