Invariant Euler-Lagrange Equations

• Invariant Euler-Lagrange equations are variational equations derived from G-invariant action functionals that encode both geometric structure and conservation laws.
• They are computed using local moving frame techniques and global algebraic methods like Lie–Tresse theory, ensuring coordinate-free formulations.
• Applications span from classical elastica and field theory to non-holonomic systems, providing robust tools for analyzing invariant variational problems.

Invariant Euler-Lagrange equations are the variational field equations derived from action functionals that are invariant under specified symmetry groups. Such equations possess manifest invariance under the action of Lie groups or pseudogroups, diffeomorphisms, or more general Lie-theoretic transformations, and they encode both the geometric content and conservation laws corresponding to these symmetries. The theory serves central roles in differential geometry, continuum mechanics, mathematical physics, the calculus of variations, algebraic invariant theory, and geometric discretization.

1. Formalism of Invariant Lagrangians and Euler-Lagrange Equations

For a manifold $M$ acted upon by a symmetry group $G \subset \operatorname{Diff}(M)$, a Lagrangian functional is said to be $G$-invariant if its value is unchanged under the prolonged group action on the relevant jet space or bundle of dependent variables. Given a variational problem for unparametrized curves or submanifolds, the functional is typically constructed as

$S[\gamma] = \int_\gamma L \cdot \omega$

where $L$ is a rational function of a finite set of absolute (differential) invariants $I^i$, their invariant derivatives $D^j(I^i)$, and $\omega$ is an invariant horizontal 1-form satisfying $\omega(D) = 1$ for a canonical invariant derivation $D$ (Kruglikov et al., 11 Jan 2026). In field theories, an invariant action is defined over sections of bundles or multivector fields, with Lagrangian densities depending only on invariant geometric data.

The derivation of invariant Euler-Lagrange equations proceeds via the variational bicomplex, replacing standard jet variables, contact forms, and horizontal differentials with their invariant counterparts under $G$. Integration by parts, adjoints of invariant operators, and decomposition of the variational differential yield a system

$$E_\mathrm{inv}(L): \quad \mathcal{A}^* E(L) - \mathcal{B}^* H(L) = 0$$

where $\mathcal{A}, \mathcal{B}$ are invariant operator matrices determined by the geometry and differential structure of the invariants, and $E(L), H(L)$ are invariant Eulerian and “Hamiltonian” expressions computed via invariant differentiation of $L$ (Kruglikov et al., 11 Jan 2026). This construction ensures that the resulting system is manifestly invariant under $G$—both locally and globally—and avoids coordinate dependencies and chart patching.

2. Methodologies: Global and Local Invariant Calculus

Invariant Euler-Lagrange equations can be derived either via local moving-frame methods (symbolic normalization of coordinates and explicit computation of the prolongation of the symmetry action), or with global, algebraic approaches utilizing rational differential invariants and invariant derivations.

• Local methods (e.g., Fels–Olver): Require explicit local cross-sections and can produce invariants involving radicals or transcendental functions. Effective in low dimension, but algebraically cumbersome in high-dimensional settings or for non-free group actions.
• Global algebraic approaches (Lie–Tresse theory): Use generators of the field of rational (absolute) differential invariants and canonical invariant coframes. The resulting formulas depend solely on rational data, allow effective computation of invariant Euler-Lagrange systems in any dimension, and identify geometric singularities via vanishing of relative invariants such as the Wronskian (Kruglikov et al., 11 Jan 2026).

The global invariant construction also provides canonical separation of group orbits, recurrence relations for invariant derivatives, and efficient computation of conservation laws or singular strata.

3. Representative Examples

Curves in Classical Geometries

• Euclidean Plane and Space: The natural invariants are curvature ($\kappa$) and torsion ($\tau$), with invariant arclength 1-forms. The invariant Euler–Lagrange equation for the elastica problem ($L = \kappa^2$) in the plane is

$(D_s^2 + \kappa^2) E(L) + \kappa H(L) = 0$

where $D_s$ is the arclength derivative (Kruglikov et al., 11 Jan 2026).

• Minkowski Spacetime (SE(3,1)): Invariants include curvature, torsion, and a fourth invariant (Wronskian-type). The invariant system generalizes elastica and conformal geodesics to Lorentzian signature, revealing lightlike extremals precisely at loci where the relative invariant degenerates.
• Projective and Conformal Geometries: Projective invariance leads to equations involving the projective curvature $\kappa_\mathrm{proj}$ and an invariant derivative $D_\sigma$. For conformal actions (e.g., on the circle or in 3D), the invariant Lagrangians and derivatives are constructed from local invariants like conformal torsion and curvature, with the corresponding invariant Euler–Lagrange systems involving higher-order differential operators whose structure is determined by the global algebraic approach.

Field Theory and Differential Geometry

• Pseudo-Kähler Manifolds: The generalized Lovelock functional constructed from a degree-$k$ characteristic form $c = \Phi(R^\nabla)$ and Kähler form $\Omega$ yields an action

$$\mathcal{L}[h] = \int_M \langle c, \Omega^k \rangle dV_h$$

whose Euler-Lagrange equations are purely algebraic, holomorphically invariant tensor equations of the form

$$F_c(h)_{pq} = -\frac{1}{(k-1)!} \iota_{\partial_p} \iota_{\partial_q} (c \wedge \Omega^{k-1}) = 0$$

This generalizes the Einstein–Gauss–Bonnet system, with invariance under holomorphic diffeomorphisms (Park, 2015).

• Intrinsic and Coordinate-Free Formulations: For fiber bundles and field theories, intrinsic derivations (Tulczyjew's formalism) yield the Euler–Lagrange equations as

$$\left[\,\lambda - (T^1_k \pi^k_Q)^*(dL)\,\right](\psi^{(1)}(x)) = 0$$

ensuring full coordinate invariance and applicability to both regular and singular Lagrangians (Salgado et al., 2019).

• Exterior Algebraic Derivation: Invariant Euler–Lagrange equations for multivector fields can be derived in coordinate-free form:

$$\frac{\partial L}{\partial \phi} = \partial \times \left( \frac{\partial L}{\partial (\otimes \phi) } \right)$$

or, when the Lagrangian depends on exterior/interior derivatives, as

$$\frac{\partial L}{\partial \phi} = (-1)^{s} \left[ d\!\rfloor\! \left( \frac{\partial L}{\partial (d \wedge \phi)} \right) - d \wedge \left( \frac{\partial L}{\partial (d\!\rfloor\!\phi)} \right)\right]$$

guaranteeing invariance under all smooth coordinate transformations (Colombaro et al., 2021).

4. Noether's Theorem and Conservation Laws in the Invariant Setting

A $G$-invariant Lagrangian guarantees that the associated Euler-Lagrange equations admit conserved quantities corresponding to variational symmetries (Noether's theorem). In the invariant framework, this correspondence is expressed via invariant operators and invariant coframes, with the conserved currents likewise built from invariant data.

For the incompressible Euler equations, Lagrangian right-invariance under the volume-preserving diffeomorphism group $\operatorname{SDiff}(\Omega)$ yields Kelvin’s circulation theorem as a Noether conservation law, and the momentum equation is equivalent to the geodesic equation on $\operatorname{SDiff}(\Omega)$ (Farazmand et al., 2018). In projective or conformal geometry, the invariants themselves often function as conserved quantities; for example, the Schwarzian derivative is a first integral for the fourth-order Euler-Lagrange equation associated with the projective-invariant Lagrangian $L = (u'')^2/(u')^2$ (Kryński, 2021).

The Lie algebra structure of conservation laws and the correspondence with variational symmetries is formalized geometrically: symmetries preserving both the contact distribution and the action yield first integrals via contraction with the Poincaré–Cartan form, closing under a natural bracket structure (Fiorani et al., 2014).

5. Computational Strategies and Discretization

Symmetry-preserving discretization schemes, essential for numerical stability and exact conservation, are constructed by invariantizing discrete Lagrangian functionals. The key tools are:

• Equivariant Moving Frames: Used to define invariant discrete jet variables and cross-sections, thereby producing invariantized discrete Lagrangian densities.
• Invariant Discrete Euler–Lagrange Equations: DEL operators inherit the invariance properties by construction, and a discrete Noether theorem yields exact (discrete) conservation laws associated to every continuous symmetry (Bihlo et al., 2021).

These methods outperform standard discretizations in long-time numerical fidelity, absence of drift in invariants, and stability near singularities.

6. Extensions: Measures, Generalized Functionals, and Non-Holonomic Constraints

• Invariant Euler–Lagrange Equations for Measures: In relaxed variational frameworks, such as the theory of closed (holonomic) measures, the condition of criticality with respect to admissible variations leads to invariant Euler–Lagrange PDEs for measures. Measures invariant under the Euler–Lagrange flow are characterized variationally: a measure is invariant if and only if it is “horizontally critical” for exact horizontal variations (Rios-Zertuche, 2018).
• Modified Formal Lagrangian Formulation: For non-variational systems, the augmented system with dummy variables enforces invariant block Euler–Lagrange equations encoding both the original and adjoint systems, yielding, under suitable identification, the original system and all its variational symmetries and conservation laws (Peng, 2020).
• Non-holonomic and Implicit Constraints: Intrinsic invariant formulations allow for the systematic treatment of non-holonomic systems (e.g., rolling Cosserat rods), Hamilton–Pontryagin principles, and singular Lagrangians, all yielding invariant Euler–Lagrange-type equations formulated intrinsically in terms of canonical 1-forms on appropriate fiber bundles (Salgado et al., 2019).

7. Impact and Foundational Aspects

The global algebraic differential invariant framework delivers a canonical, uniform method for the calculation of invariant Euler–Lagrange equations across all classical geometries and a wide array of physical systems (Kruglikov et al., 11 Jan 2026). Manifest invariance ensures compatibility with underlying physical or geometric symmetry principles, direct compatibility with conservation laws, and robust geometric discretizations. Limitations are primarily algebraic—existence of sufficient generating invariants and invariant coframes fail for infinite-dimensional pseudogroups, and singularities arise at relative-invariant loci corresponding to distinguished geometric structures.

The invariant formulation is foundational for the extension of classical and modern field theory, geometric mechanics, and symmetry analysis into the realms of singular, constrained, discrete, or measure-theoretic variational problems.