A simple criterion for identifying slow-fast relaxation oscillations Journal: Applied Mathematics and Computation
Lucas MacQuarrie, Pei Yu · 2025
Read A simple criterion for identifying slow-fast relaxation oscillations Journal: Applied Mathematics and Computation by Lucas MacQuarrie, Pei Yu (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-2509-13893.txt.
Introduction
Lucas MacQuarrie and Pei Yu sharpen a test for relaxation oscillations in a 2025 dynamical systems paper. Slow-fast systems can produce relaxation oscillations, long slow drifts punctuated by sudden jumps, but the standard geometric theory does not apply to every model. Earlier work by Zhang, Wahl, and Yu, and by Yu and Wang, gave simple checkable conditions that certify such oscillations in those harder cases.
MacQuarrie and Yu present seven models that fail the existing conditions yet still display recurrent behavior, showing the tests are too narrow. Motivated by those counterexamples, they generalize the conditions into new, still simple, hypothetical criteria for recognizing slow-fast relaxation oscillations, and lay out arguments aimed at growing them into a fuller mathematical theory.
Catalog record
| CREATOR | Lucas MacQuarrie, Pei Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-13893
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A simple criterion for identifying slow-fast relaxation oscillations Journal: Applied Mathematics and Computation Lucas MacQuarrie, Pei Yu Abstract Simple conditions have been developed in [Zhang, Wahl and Yu, SIAM Rev. 2014; Yu and Wang, Math. Biosci. Eng. 2019], which are used to identify the existence of slow-fast relaxation oscillations that appear in differential systems, where the geometric singular perturbation theory is not applicable. In this study, we present seven models which fail to satisfy these conditions, yet still show recurrence phenomena, which motivate us to generalize these conditions to establish new simple hypothetical conditions for identifying slow-fast relaxation oscillations and providing arguments towards the development of a mathematical theory. 2010 MSC 34C07 , 34C23 , 34C25 , 34C26 † † graphicalabstract: † † highlights: Research highlight 1 Research highlight 2 1 Introduction Systems of differential equations are able to capture a wide variety of dynamics in their solutions, making them an ideal framework for modeling real world phenomena in biology [ 6 , 26 , 19 , 18 ] , climate [ 24 , 14 ] , and countless other areas of science [ 7 , 30 , 2 , 9 , 3 ] . In particular, systems of ordinary differential equations (ODEs), which exhibit solutions evolving over multiple timescales, are of interest as such behaviour can often be seen in models of natural phenomena. The study of such equations has inspired the development of new mathematical tools such as Geometric Singular Perturbation Theory (GSPT) [ 8 , 15 , 11 , 22 ] . Although these techniques allow us to study a broad range of slow-fast oscillations in ODEs, there still exist systems to which the GSPT cannot be applied. In the following, we briefly outline the basic idea of the GSPT, which requires a given system of ODEs to be in the form of \begin{array}[]{rl}\dfrac{dx}{dt}&=f(x,y,\varepsilon)\\[6.45831pt] \varepsilon\dfrac{dy}{dt}&=g(x,y,\varepsilon),\end{array} (1) where the nonlinear functions, f\!:\mathbb{R}^{n}\times\mathbb{R}^{m}\times\mathbb{R}\to\mathbb{R}^{n} and g\!:\mathbb{R}^{n}\times\mathbb{R}^{m}\times\mathbb{R}\to\mathbb{R}^{m} , are assumed to be differentiable, and \varepsilon is a small positive perturbation parameter. We see an explicit separation between the slow variable x and the fast variable y with appearance of \varepsilon . With a change of timescale \tau=\varepsilon t , the fast system ( 1 ) becomes the slow system, \begin{array}[]{rl}\dfrac{dx}{d\tau}&=\varepsilon f(x,y,\varepsilon)\\[6.45831pt] \dfrac{dy}{d\tau}&=g(x,y,\varepsilon).\end{array} (2) The systems ( 1 ) and ( 2 ) are equivalent for \varepsilon>0 , but setting \varepsilon=0 yields the reduced equation , \begin{array}[]{rl}\dfrac{dx}{dt}&=f(x,y,0)\\[6.45831pt] 0&=g(x,y,0),\end{array} (3) and the layer equation \begin{array}[]{rl}\dfrac{dx}{d\tau}&=0\\[6.45831pt] \dfrac{dy}{d\tau}&=g(x,y,0).\end{array} (4) Equation ( 3 ) is an algebraic-differential equation, describing a flow on a subset of the state space called the critical manifold \mathcal{M}_{0} defined by g(x,y,0)=0 . Equation ( 4 ) describes the dynamics on the whole state space, however in this case the critical manifold is now comprised of equilibrium points. In other words, ( 3 ) describes the dynamics on the critical manifold while ( 4 ) describes the dynamics around it. Fenichel’s theorem [ 8 ] tells us that if the critical manifold is both compact and normally hyperbolic, then a perturbed version of \mathcal{M}_{0} , \mathcal{M}_{\varepsilon} , exists in the system with 0<\varepsilon\ll 1 [ 8 ] . What’s more, the stable and unstable manifolds of \mathcal{M}_{0} persist under the perturbations as well. Thus, by studying the reduced equations, we can understand the solutions of the full system for small parameter values \varepsilon . The critical manifolds of GSPT can be used to explain the dynamics of solutions of singularly perturbed equations. Figure 1 shows two singularly perturbed systems with two typical critical manifolds. The slow-fast behaviour can be illustrated by the solutions following and leaving the critical manifold, and their difference can be seen from the shape of the critical manifold. Figure 1 (a) shows a critical U-shaped manifold, while Figure 1 (b) shows a critical S-shaped manifold. These shapes determine the possible behaviour of solutions as the solutions move in the state space. \begin{overpic}[Fig/addFig4.png] \end{overpic} \begin{overpic}[Fig/addFig5.png] \end{overpic} (a) U-shaped critical manifold (b) S-shaped critical manifold Figure 1: Simulations of two systems of ODEs displaying the typical U-shape and S-shape critical manifolds as found by the GPST in dotted orange. The two ODEs are (a) x^{\prime}=y-\frac{x^{2}}{2},\ y^{\prime}=-\varepsilon x(1+x^{2}),\ \varepsilon=0.005 , constructed based on [ 21 ] ; and (b) the Van der Pol oscillator: x^{\prime}=(x-\frac{x^{3}}{3}+y)/\varepsilon,\ y^{\prime}=-x,\ \varepsilon=0.01 [ 8 , 25 ] . These two systems are in the form of ( 1 ). The figures on the left side are the corresponding time histories, while the figures on the right side are the trajectories in phase space; the velocities of the solutions are indicated by solid color and the initial condition are marked with a red point. It is not necessary for a system of differential equations to be of the form ( 1 ) for the existence of slow-fast behavior. For example, the following 2-d dimensionless HIV model [ 36 , 37 , 35 ] exhibits slow-fast behaviour without being in the above form: \begin{array}[]{rl}\dfrac{dx}{dt}&=1-Dx-\Big(B+\dfrac{Ay}{y+C}xy\Big),\\[6.45831pt] \dfrac{dy}{dt}&=\Big(B+\dfrac{Ay}{y+C}\Big)xy-y,\end{array} (5) where x and y represent the healthy and infected cells, respectfully, and A,B,C,D are positive real parameters. \begin{overpic}[Fig/addFig1.png] \put(5.0,90.0){$y$} \put(100.0,6.0){$t$} \end{overpic} Figure 2: Simulated time history of model ( 5 ) for A=0.364 , B=0.06 , C=0.823 and D=0.057 with the initial condition (x(0),y(0))=(5,13) , showing the slow-fast motion which is possible in a system not in the form of the singular system ( 1 ). The authors of [ 36 , 37 ] showed that ( 5 ) exhibits a slow-fast oscillation called recurrence, as shown in Figure 2 , which is characterized by long periods of quiescence followed by a quick jerk away from equilibrium before quickly returning and then repeating. In the context of HIV, recurrence is the so-called viral-blips which manifest as a spike in viral population in the patient before returning to subcritical levels [ 29 , 5 ] . Using the simple language of bifurcation theory [ 10 , 33 ] , the following hypothesis with four conditions was proposed for the existence of the recurrence phenomenon in differential systems [ 36 , 37 ] : (i) there exist at least two equilibrium solutions; (ii) there exists a transcritical bifurcation at an intersection of the two equilibrium solutions; (iii) there is a Hopf bifurcation which occurs from one of the equilibrium solutions; (iv) large oscillations (or, more generally, global, persistent motions) can occur near the transcritical critical point. In [ 36 , 37 ] , the authors provide the explanation as follows: the slow-fast behavior occurs around a saddle point, that is, an equilibrium with one positive and one negative eigenvalue, near the transcritical bifurcation point. The saddle point connects to a stable and an unstable manifold with the corresponding eigenvalues \lambda_{1}<0 and 0<\lambda_{2}\ll 1 near the equilibrium. This indicates that the trajectories near the saddle point will quickly approach the saddle point via the fast-stable manifold and leave slowly via the slow-unstable manifold, thus yielding slow-fast motions. This can also be used to explain the situation if the transcritical bifurcation condition is changed to include a saddle-node bifurcation [ 34 ] . Later, the four conditions were modified when the recurrent behavior was considered in a 3-d oscillating networks of biologically relevant organic reactions [ 34 , 35 ] , which has a saddle-node bifurcation,