Projective geodesic extensions by conformal modifications in nonholonomic mechanics
Malika Belrhazi, Tom Mestdag · 2025
Read Projective geodesic extensions by conformal modifications in nonholonomic mechanics by Malika Belrhazi, Tom Mestdag (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-2509-15863.txt.
Introduction
Malika Belrhazi and Tom Mestdag study when a constrained machine’s paths can be read as geodesics. Nonholonomic systems move under limits on their velocity, like a rolling wheel that cannot slide sideways. Their paths are not usually the shortest paths of any metric. A projective extension reparametrizes the paths so they can be seen as part of a metric’s geodesics.
The pair derive conditions that are both needed and enough for such extensions, when the energy is kept up to a conformal stretch. For systems with a symmetry group, they clarify how extensions relate to three near ideas: a simplicity trait, invariant measures, and Hamiltonization. Fresh examples mark the gaps between these notions.
Catalog record
| CREATOR | Malika Belrhazi, Tom Mestdag |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 131 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-15863
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Projective geodesic extensions by conformal modifications in nonholonomic mechanics Malika Belrhazi, Tom Mestdag Abstract Projective geodesic extensions are reparametrizations of the trajectories of a nonholonomic mechanical system (with only a kinetic energy Lagrangian), in such a way that they can be interpreted as part of the geodesics of a Riemannian metric. We derive necessary and sufficient conditions for the existence of these extensions, in the case where the constrained Lagrangian remains preserved up to a conformal transformation. When the nonholonomic system has a symmetry group (a Chaplygin system), we clarify the relation between projective geodesic extensions and closely related concepts, such as φ \varphi -simplicity, invariant measures and Hamiltonization. Throughout the paper, new and relevant examples illustrate the key differences between all these concepts. Keywords: Lagrangian systems, nonholonomic constraints, Riemannian metrics, Chaplygin system, geodesic extension, projective transformation, conformal change, invariant measure. Mathematics Subject Classification: 37J06, 37J60, 53Z05, 70G45, 70G65. 1 Introduction The idea that the governing equations of a physics problem should somehow be “optimal” or “minimal” with respect to the physical input is very prominent in many models of mathematical physics (and of science, in general). For example, in general relativity, the paths that objects follow when influenced only by gravity are geodesics. Roughly speaking, they represent the shortest possible paths the objects can follow in curved spacetime. In the absence of either a potential or external forces, the same can be said about mechanical systems: In that situation, the Euler-Lagrange equations are equivalent with the geodesic equations of the metric that defines the ‘purely kinetic’ Lagrangian. This property remains valid when the system is subjected to holonomic (position-depending) constraints: one may take the holonomic constraints implicitly into account by appropriately shrinking the configuration space. Apart from the inherent beauty of relying on optimal equations, a mathematical model of a physical problem that is based on geodesics has also many other advantages: For example, we may investigate the influence of curvature, exploit the benefits of symmetry, study its integrability, apply geometric numerial integrators, etc. In mechanics, the situation changes when the problem includes nonholonomic constraints. These are non-integrable velocity-dependent restrictions of the motion of the system, such as e.g. those caused by wheels (that roll without slipping), skates (that are prevented to move in a direction perpendicular to the blade), pursuit problems, etc. Here, we consider only nonholonomic constraints that can be given in the form of linear (autonomous) equations, and that therefore can be represented by a non-integrable distribution \mathcal{D} on the configuration manifold Q of the mechanical system. Even though most of our results may be extended by including a potential at the appropriate places, we have chosen to limit ourselves here to purely kinetic systems, for the clarity of our exposition. If L is the (kinetic energy) Lagrangian of the system, the equations of motion for such systems are the so-called Lagrange-d’Alembert equations of (L,\mathcal{D}) . They form a set of first- and second-order ordinary differential equations whose solutions we will call the nonholonomic trajectories of (L,\mathcal{D}) . The Lagrange-d’Alembert equations are not Euler-Lagrange equations, and are also not geodesic equations. However, it still remains useful to identify the solutions of these equations as (part of the possibly larger set of) geodesics of some metric on the configuration space. In the current paper, as in our paper [ 6 ] , we investigate methods by which we can determine such a metric, which we call a geodesic extension of the nonholonomic system. For a geodesic extension, the nonholonomic trajectories of (L=\frac{1}{2}g,\mathcal{D}) are among the geodesics of a newly constructed Riemannian metric \hat{g} . A special type of geodesic extension, is the one where \hat{g} is a so-called \mathcal{D} -preserving modification of the metric g . This means that \hat{g} preserves g on the constraint distribution \mathcal{D} . Under this assumption, we were able to derive in [ 6 ] two conditions (denoted by (A) and (B) in [ 6 ] ) whose solutions can be used to construct a geodesic extension \hat{g} . Even though this is already a very strong tool, there exists systems for which our method can not be applied (for example, when there is only one nonholonomic constraint, see the discussion in Section 9 of [ 6 ] ). In order to apply our theory to a broader class of nonholonomic systems, we will present in this paper an interesting generalization, by adding an extra twist. After some preliminaries in Section 2 , we introduce in Section 3 a new concept: projective geodesic extensions . Here, we search for a Riemannian metric \hat{g} which is such that its pre geodesics contain the nonholonomic trajectories, that is: geodesics of the metric \hat{g} , but possibly with a different parametrization than the one by time. In particular, we wish to accomplish this by means of a \mathcal{D} -conformal modification . This is, in essence, a \mathcal{D} -preserving modification, up to an extra conformal transformation by a conformal factor F . Under these assumptions, we derive in Lemma 1 two new conditions (now: (A^{\prime}) and (B^{\prime}) ), to obtain a valid projective geodesic extension. We illustrate our method with the examples of the generalized nonholonomic particle and the two-wheeled carriage. We will continue to use these examples throughout the whole paper, because they neatly demonstrate different aspects of our results. In the literature, one may already find some earlier results concerning (what we now call) projective geodesic extensions. For example, in [ 1 ] , the authors construct one such extension (in their Theorem 3.4), but only for the case of a Chaplygin system (a nonholonomic system with a symmetry Lie group) that is, in addition, \varphi -simple (in the wording of [ 18 ] ). One of the main goals of this paper is to show that this very specific case is not the only one of interest: we will find (both in theory and in examples) projective geodesic extensions for nonholonomic systems that do not adhere to any of these strong assumptions. If we set the geodesic extension problem in a broader context, it can be regarded as a special case of the kind of Hamiltonization problems that have been discussed in e.g. [ 8 , 5 ] . There, the main idea is to enlarge the mixed first- and second order system of (L,\mathcal{D}) to a (chosen) full system of second-order differential equations, and to write those as the Euler-Lagrange equations of a new Lagrangian \hat{L} . When compared to the setting of this paper, both the given Lagrangian L and the sought-for Lagrangian \hat{L} are here of purely-kinetic-type (i.e. coming from a Riemannian metric and with constant potential). Unfortunately, the terminology Hamiltonization is over-used in the literature. In other papers such as e.g. [ 2 , 3 , 9 , 15 ] (just to recall a few), Hamiltonization stands for the approach where one uses the symmetry group G of the (Chaplygin) nonholonomic system to reduce it from a dynamical system on Q to a system on the quotient manifold Q/G . This reduced system is then said to be Hamiltonizable if one can find a symplectic or Poisson structure on the quotient manifold for which it is Hamiltonian. To set