Constrained Symplectic and Contact Hamiltonian Systems: A Review

Callum Bell, David Sloan · 2026

Read Constrained Symplectic and Contact Hamiltonian Systems: A Review by Callum Bell, David Sloan (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-2604-27940.txt.

Introduction

Bell and Sloan review how constraints shape Hamiltonian mechanics when a theory's description has degeneracies. They set out the geometry behind pre-symplectic and pre-contact spaces, which suit conserved and leaking systems in turn, and build the rules that find where steady motion can exist. Each rule set ends with a worked example.

Catalog record

CREATORCallum Bell, David Sloan
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54723
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.27940

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Constrained Symplectic and Contact Hamiltonian Systems: A Review Callum Bell, David Sloan Abstract Singular theories, characterised by the presence of degeneracies in their Lagrangian or Hamiltonian descriptions, require the systematic implementation of constraints in order to obtain well-defined dynamics. While the symplectic framework provides the standard geometrical setting for conservative mechanical systems, those theories which exhibit dissipative effects are most appropriately discussed within the context of contact geometry. In this review, we present the geometrical structure underlying pre-symplectic and pre-contact manifolds, and develop the corresponding constraint algorithms that determine the admissible subset of phase space upon which consistent Hamiltonian evolution exists. We then close the discussion of each of the constraint algorithms with an example. I Introduction Singular theories of both particles and fields are amongst the most frequently encountered and physically interesting class of models in the mathematical description of natural phenomena. In the present context, we take ‘singular theory’ to be synonymous with any model whose Hamiltonian (or Lagrangian) description must, in some way, be constrained due to the presence of degeneracies. Of course, there exist numerous kinds of singular theories, with the exact nature of the constraints being heavily context-dependent; however, of greatest interest to us will be those models which exhibit gauge invariance. Typically, such theories are addressed from the perspective of principal and/or vector bundles, in which a gauge transformation simply refers to a change of local trivialising section of the appropriate bundle. A more detailed exposition of this treatment may be found in nakahara2018geometry , for example. In general, we shall concern ourselves with singular theories of two distinct types: those described by symplectic Hamiltonians, and those with contact Hamiltonians. Symplectic geometry is by far the prevailing framework within which analytical mechanics is formalised da2008lectures . Such manifolds are characterised as being of even dimension, and possessing a closed, non-degenerate differential 2-form \omega . A theorem due to Liouville guarantees that volumes of subsets of any symplectic manifold are preserved under the flow of Hamiltonian vector fields arnol1990symplectic . As such, symplectic geometry provides a particularly elegant approach to the description of conservative mechanical systems. In section ( II ), we review a number of elementary geometrical concepts, which allow us to describe pre-symplectic Hamiltonian systems. Section ( III ) is then dedicated to a detailed treatment of the process of systematically deducing the subset of a system’s phase space upon which the Hamiltonian equations of motion admit well-defined solutions. This constraint algorithm, first developed in Nester1 , uses geometrical principles, rather than the more familiar algebraic methods of the Dirac-Bergmann procedure dirac2013lectures . For the reader acquainted with this latter method of constraint, section ( IV ) provides a discussion of the similarities and differences between Dirac’s construction and the geometrical ideas used throughout. The second broad category of singular systems of interest are those whose phase spaces are pre-contact manifolds bravetti2017contact . Contact geometry is far less prevalent within the physics literature. This is principally due to the fact that, unlike symplectic manifolds, the contact counterparts are odd-dimensional, with a phase space measure which is not conserved, but instead may undergo focusing/defocusing effects. By virtue of this property, contact Hamiltonian systems are frequently employed to describe dissipative phenomena. In section ( VI ), we develop the formalism of contact geometry, together with its implementation in the description of Hamiltonian systems. Section ( VII ) then contains the geometrical constraint algorithm for pre-contact systems, and closely mirrors the scheme presented in section ( III ). The principal motivation for studying these non-conservative theories resides in a particular kind of symmetry reduction, known as contact reduction sloan2018dynamical . We do not explicitly treat this procedure in the present work; however, the interested reader is invited to consult bravetti2023scaling ; sloan2021scale , and references therein. Somewhat informally, it is often found that classical symplectic systems admit a vector field, whose effect is to Lie-drag the symplectic form to itself, and to rescale the Hamiltonian by some non-zero factor. Physically, this vector field - known as a dynamical similarity - is the generator of a phase space rescaling, that has the particular property of mapping between mathematically distinct, but observationally indistinguishable solutions to the equations of motion. The premise of contact reduction is then to quotient the symplectic phase space by the one-dimensional orbits of the dynamical similarity. In doing so, the phase space dimension is reduced by one, and naturally inherits a contact structure. Note that this is structurally quite distinct from the more familiar reduction schemes of Marsden and Weinstein marsden1974reduction , in which the symmetry reduction reduces the phase space dimension by two, thereby preserving its symplectic nature. II Geometrical Preliminaries We begin with a number of fundamental ideas surrounding the symplectic description of Lagrangian and Hamiltonian systems. Suppose that Q is an n -dimensional smooth manifold, corresponding to the configuration space of just such a system. A Lagrangian function L:TQ\rightarrow\mathbb{R} is introduced on the tangent bundle TQ , with projection \pi:TQ\rightarrow Q . Fibre coordinates on TQ are denoted v^{i} , so that, given some q\in U\subset Q , induced bundle coordinates on \pi^{-1}(U) are given by (q^{i},v^{i}) . Every tangent bundle is endowed with a ( 1^{\textrm{st}} -order) almost-tangent structure prieto2014geometrical , provided by the vertical endomorphism J:T(TQ)\rightarrow T(TQ) which satisfies \star J^{2}=\textrm{id}_{TQ} \star \textrm{rank}\,J=n In order to construct this endomorphism, we briefly review the concept of the vertical lift of a tangent vector de2020review . Let X and u be vectors in the same tangent space T_{q}Q at q\in Q . The vertical lift of X to T(TQ) is written as (X)^{V}_{u}=\frac{d}{dt}\biggr|_{t=0}\left(u+tX\right) (1) In local bundle coordinates (q^{i},v^{i}) , we write X=X^{i}\,\frac{\partial}{\partial q^{i}}\biggr|_{q}\quad\quad\implies\quad\quad(X)^{V}_{u}=X^{i}\,\frac{\partial}{\partial v^{i}}\biggr|_{u} The vertical endomorphism J is then defined via its action on elements of T(TQ) . In particular, starting with a point q\in Q , we take a tangent vector u\in T_{q}Q , and consider some element Z\in T_{u}(TQ) defined at u . The action of J on the vector Z reads J(Z)=(D_{u}\pi(Z))^{V}_{u} (2) where D\pi:T(TQ)\rightarrow T(TQ) denotes the pushforward. This defines a tensor field of type (1,1) , and in local bundle coordinates, we have J=\frac{\partial}{\partial v^{i}}\otimes dq^{i} (3) In addition to the vertical endomorphism J , the tangent bundle TQ admits a second canonical geometrical structure: the Liouville vector field \Delta\in\mathfrak{X}^{\infty}(TQ) . This is defined to be the vector field whose flow generates dilatations on the fibres of TQ . More specifically, if \Delta_{u}=\frac{d}{dt}\biggr|_{t=0}\Phi_{t}(u) (4) so that \Phi:\mathbb{R}\times TQ\rightarrow TQ is the flow of the vector field \Delta , then we have \Phi_{t}(u)=e^{t}u (5) From this, it follows that \Delta may be written locally as \Delta=v^{i}\,\frac{\partial}{\partial v^{i}} (6) From these structures, we introduce a 1-form \theta_{L}\in\Omega^{1}(TQ) \theta_{L}:=\iota_{J}\,dL=dL\circ J\quad\quad\implies\quad\quad\theta_{L}=\frac{\partial L}{\partial v^{i}}\,dq^{i} (7) where we have used the local expression ( 3 ) for the vertical endomorphism J . The form \theta_{L} allows us to naturally define a closed 2-form on TQ as \omega_{L}:=-\,d\theta_{L} . The cases of greatest interest throughout will be those in which \omega_{L} is a pre-symplectic form: that is, when it is closed, but degenerate. This may be expressed as the requirement that the matrix of second derivatives W_{ij}:=\left(\frac{\partial^{2}L}{\partial v^{i}\partial v^{j}}\right) be of non-maximal rank, and thus singular. Finally, a pre-symplectic Lagrangian system

Read the complete text