Infinite-dimensional nonholonomic and vakonomic systems
Alexander G. Abanov, Boris Khesin · 2025
Read Infinite-dimensional nonholonomic and vakonomic systems by Alexander G. Abanov, Boris Khesin (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-2511-00629.txt.
Introduction
Alexander Abanov and Boris Khesin tour infinite-dimensional systems with nonholonomic constraints, the rolling-without-slipping rules that make dynamics subtle. In finite size, the two kinds of dynamics, nonholonomic and vakonomic, both arise as limits of ordinary systems with friction-like loss. A skate gliding on an inclined plane makes the split plain in a classical, concrete way.
The endless-dimensional menagerie includes subriemannian and Euler-Poincaré-Suslov systems on Lie groups, the Heisenberg chain, and the Camassa-Holm equation of shallow water fame. Also featured are a nonholonomic Moser theorem, fluid approximations, parity-breaking fluids, and transport-born solutions of Burgers-type equations. The tour ends with a car pulling endlessly many trailers, whose limit becomes a Chaplygin sleigh dragging a string, governed by an endless-dimensional distribution.
Catalog record
| CREATOR | Alexander G. Abanov, Boris Khesin |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-00629
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Infinite-dimensional nonholonomic and vakonomic systems Alexander G. Abanov, Boris Khesin Abstract In this paper, we present a collection of infinite-dimensional systems with nonholonomic constraints. In finite dimensions the two essentially different types of dynamics, nonholonomic or vakonomic ones, are known to be obtained by taking certain limits of holonomic systems with Rayleigh dissipation, as in [ Koz83 ] . After visualizing this phenomenon for the classical example of a skate on an inclined plane, we discuss its higher-dimensional analogue, the kinematics of a car with n n trailers, as well as its limit as n → ∞ n\to\infty . We show that its infinite-dimensional version is a snake-like motion of the Chaplygin sleigh with a string, and it is subordinated to an infinite-dimensional Goursat distribution. Other examples of nonholonomic and vakonomic systems include subriemannian and Euler–Poincaré–Suslov systems on infinite-dimensional Lie groups, the Heisenberg chain, the general Camassa–Holm equation, infinite-dimensional geometry of a nonholonomic Moser theorem, parity-breaking nonholonomic fluids, and potential solutions to Burgers-type equations arising in optimal mass transport. You know that in moments of stress You tend to get tenser, not less; But since stress is a tensor You needn’t feel denser; It’s tricky, I have to confess. H.K. Moffatt, FRS 1 Introduction The theory of infinite-dimensional Hamiltonian systems is by now well developed, with plenty of examples and several possible frameworks for infinite-dimensional Poisson brackets. The corresponding evolution PDEs include the Korteweg–de Vries, Nonlinear Schrödinger, Camassa–Holm, Kadomtsev–Petviashvili and many other equations, while examples of infinite-dimensional symplectic and Poisson structures include those of Marsden–Weinstein ones on the space of knots and membranes, Gelfand–Dickey brackets on pseudo-differential symbols, Lie-Poisson brackets on the duals of infinite-dimensional Lie algebras, etc. On the other hand, infinite-dimensional contact (or more generally, nonholonomic) systems are very rare, despite the fact that in finite dimensions contact geometry is regarded as the twin sister of symplectic geometry. General equations of nonholonomic dynamics in an infinite-dimensional setting were described in [ SBKZB17 , SZB20 ] . Here, we collect several examples of infinite-dimensional nonholonomic and vakonomic systems that, in our opinion, are suggestive for the future development of infinite-dimensional nonholonomic mechanics. The goal of this note is mostly expository. We start by describing two different types of nonholonomic dynamics, vakonomic or the nonholonomic one governed by the Lagrange-d’Alembert principle. They are known to be obtained by taking certain limits of holonomic systems with Rayleigh dissipation, see [ Koz83 ] . Vakonomic mechanics is closely related to subriemannian geometry and control theory, see [ Mon02 ] . In a nutshell, given a bracket-generating distribution on a manifold and a Lagrangian L of a physical system (more generally, any cost function), vakonomic approach describes trajectories as critical points (minimizers) of the functional \mathcal{L}=\int L\,dt on the set of admissible paths, i.e., paths subordinated to the given distribution [ Koz92 ] . Such trajectories, having the variational origin, may drastically differ from the dynamics given by the Lagrange–d’Alembert principle, defining the motion of nonholonomic systems and requiring that the constraint forces would do no work on virtual displacements consistent with the constraints [ Blo03 ] . Those two principles coincide for holonomic systems. In the papers [ Koz83 , Koz92 ] , a general setting was described in which, by introducing a regularization via Rayleigh dissipation and taking different limits, one is led to either nonholonomic or vakonomic equations. We start by reviewing and illuminating with figures the classical example of a skate on an inclined plane considered in [ Koz83 ] . This is followed by infinite-dimensional examples that include subriemannian and Euler–Poincaré–Suslov systems on Lie groups, and in particular, the Heisenberg chain and the general Camassa–Holm equation. Next we describe in more details the infinite-dimensional geometry of a nonholonomic Moser theorem and parity-breaking nonholonomic fluids, as well as potential solutions to Burgers-type equations arising in optimal mass transport. After that we return to the skate example in the context of Goursat distributions and the n -trailer systems. The reader might find it interesting to compare the many-trailer systems for a unicycle or skate with those of a car, which turns out to be related to a dimensional shift for the corresponding configuration space. Finally, we present the kinematics of its limit as n\to\infty . We show that its infinite-dimensional version is a snake-like motion of the Chaplygin sleigh with a string and it is subordinated to an infinite-dimensional Goursat distribution. 2 Vakonomic and nonholonomic systems: the ideal skate problem Let us start with a finite-dimensional holonomic natural system, depending on parameters. By appropriately introducing the Rayleigh dissipation and taking various limits one can obtain vakonomic or nonholonomic systems, see [ Koz83 , Koz92 ] . 2.1 Various limits of natural systems Let M be a Riemannian manifold with a metric g and \tau a bracket-generating distribution on M . We assume the bracket-generating property of \tau so that there existed admissible paths (i.e., tangent to the distribution \tau ) connecting any two points in M . Consider a natural system \mathcal{L}(\gamma):=\int_{\gamma}L(q,\dot{q})\,dt for the Lagrangian L=K-U , where K=(1/2)g(\dot{q},\dot{q}) is the kinetic energy corresponding to the metric g and U=U(q(t)) , while the integral is taken over the paths \gamma=\{\gamma(t),\,t\in[0,1]\} with fixed endpoints. Vakonomic system describes extremals of the functional \mathcal{L} among admissible paths, i.e., among paths \gamma subordinated to the distribution \tau . It has a variational origin, hence the name. Solutions of a nonholonomic system satisfy the Lagrange–d’Alembert principle: \left(\frac{d}{dt}\frac{\partial L}{\partial\dot{q}}-\frac{\partial L}{\partial q}\right)\cdot\delta q=0 for every variation \delta q\in\tau . Now modify the Lagrangian L by introducing a small parameter \nu : L_{\nu}=L+(1/2\nu)g(\dot{q}^{\perp},\dot{q}^{\perp}) , where \dot{q}^{\perp} is the component of the velocity vector \dot{\gamma}(t) transversal to \tau . For \nu\not=0 this is a holonomic system on M with a nondegenerate metric. Trajectories of the corresponding system are extremals of the functional \mathcal{L}_{\nu}(\gamma) , i.e., solutions of the Euler-Lagrange equation corresponding to \delta\mathcal{L}_{\nu}/\delta\gamma=0 : \frac{d}{dt}\frac{\partial L_{\nu}}{\partial\dot{q}}-\frac{\partial L_{\nu}}{\partial q}=0\,. It is well known that nonholonomic systems can be realized as limits of systems subject to friction forces [ Car33 ] . Let us encode these forces via a Rayleigh dissipation function \frac{1}{\alpha}R(\dot{q}), so that the limit \alpha\to+0 corresponds to infinitely strong friction. The corresponding equation assumes the form \frac{d}{dt}\frac{\partial L_{\nu}}{\partial\dot{q}}-\frac{\partial L_{\nu}}{\partial q}=-\frac{1}{\alpha}\frac{\partial R(\dot{q})}{\partial\dot{q}}\,. The two types of systems — vakonomic and nonholonomic — can be obtained by taking different limits of the model with dissipation. The limit \nu\to+0 leads to vakonomic dynamics, while the limit \alpha\to+0 results in nonholonomic dynamics governed by the Lagrange–d’Alembert principle. Kozlov [ Koz83 ] showed that in the double limit \nu,\alpha\to+0 , with fixed ratio \mu=\nu/\alpha , one obtains a one-parameter family of systems whose equations interpolate between vakonomic ( \mu\to+0 ) and nonholonomic ( \mu\to+\infty ) regimes. 2.2 Example: the ideal skate problem As an illustration of the above, we consider a skate moving on an inclined plane. Let (x,y) denote the contact point of the skate and \theta its orientation angle. The configuration space is Q=\mathbb{R}^{2}\times S^{1} . The skate is subject to the nonholonomic constraint: \phi:=\dot{x}\sin\theta-\dot{y}\cos\theta=0. Following [ Koz83 , Koz92 ] , we consider an extended Lagrangian L_{\nu} of an unconstrained system with dissipation given by the Rayleigh dissipation function R_{\alpha} : \displaystyle L_{\nu} \displaystyle=\frac{1}{2}(\dot{x}^{2}+\dot{y}^{2}+\dot{\theta}^{2})-gx+\frac{1}{2\nu}\phi^{2},\hskip 20.00003ptR_{\alpha}=\frac{1}{2\alpha}\phi^{2}. Here g is a parameter