In this paper
AbstractIntroductionThe Noether symmetries of time dependent conformal LagrangianThe Damped Duffing Oscillator (DDO)Homothetic Algebra and Potential ConditionsCase I-A: Branch 1: (DHO, )Branch 2: (Bateman invariant, DHO)Branch 3: (new invariant, DDO)Case I-B: ()Case II: Case II-A: Gradient KV (translations)Case II-B: Non-gradient KV (rotations)Verification of the InvariantsVerification of Branch 1Verification of Branch 2 (Bateman invariant)Verification of Branch 3 (new invariant)Numerical verification of the new invariantConclusionProof of Theorem 1The First Prolongation and Noether ConditionCubic Terms: The Integrability Condition for The Constraint EquationSubcase II-B: is not a gradient KVReferencesWorksPhysics and Astronomy
Geometric classification of Noether invariants of the Time-Dependent damped Duffing Oscillator
Abstract
We present a systematic geometric classification of the Noether point symmetries of time-dependent conformal Lagrangians of the form , whose Euler–Lagrange equations describe dynamical systems with time-dependent linear damping. The classification is carried out by coupling the temporal parameters of the conformal factor with the homothetic algebra of the kinetic metric through a set of compatibility and integrability conditions. The central result is a main theorem which determines the Noether generators, the associated gauge functions and the resulting invariants. The theorem splits naturally into two cases according to whether the quantity is non-zero or zero, with further subcases depending on whether the homothetic vectors are gradient or non-gradient. The theorem imposes no restrictions on , or , and is therefore fully general and covariant. We apply the theorem to the time-dependent damped Duffing oscillator (DDO) in Euclidean space with and potential V = \frac{1}{2}a\,r^{2} + \frac{% λ}{4}r^{4}. For the damped harmonic oscillator () the theorem reproduces the classical Bateman invariant as a special case. For the anharmonic case we obtain a new exact time-dependent quadratic invariant, valid when the linear stiffness satisfies . Each invariant is verified explicitly by direct differentiation along the equations of motion. The new invariant provides a rare example of a global conservation law for a fully nonlinear dissipative system, and can be used for the integration and quantization of the DDO.
Introduction
The identification of exact invariants and conservation laws remains a fundamental pillar in the analysis of non-linear dynamical systems. While conservative frameworks enjoy well-established geometric properties, dissipative systems present severe structural challenges due to the explicit symmetry breaking introduced by friction terms. Historically, the classical dissipative system, the damped harmonic oscillator (DHO), has been modeled through auxiliary variable formulations which determines the time Ermakov–Lewis invariant. This has been used in various ways e.g. for the solution of the dynamical equations or as a Hamiltonian in the quantization of the DHO [1, 2, 3, 4, 5, 6, 7]. However, in more complex systems the determination of invariants with this method is not available and one has to work either by approximation methods, simplifications or an ansatz. One such example is the anharmonic oscillator, or Duffing oscillator (DO) a holonomic system with potential . The dynamical equations for the undamped and the unforced case of the DO have been integrated in terms of the Jacobi elliptic function cn [8, 9]. For the damped and the forced case this is not possible and one looks for approximate solutions using various methods [10].
In the present work we are not looking at the dynamical equations but we are looking for the determination of invariants of dynamical systems with time dependent linear damping. Obviously Noether point symmetries offer an elegant covariant methodology to address this point.
As a consequence of the above we consider a general conformal Lagrangian of the form:
where is the kinetic metric defined by the kinetic energy (not the metric of the space where motion takes place) and is a general potential. We then discuss the geometric classification of Noether point symmetries of this type of Lagrangian. The reason for choosing this form of Lagrangian is that the Euler Lagrange equations are:
These are the dynamical equations for a system with time dependent linear damping term. In particular these equations generate the time dependent damped DO (DDO) if one considers the Euclidean metric and the potential . Earlier applications of this approach on the damped harmonic oscillator (DHO) and other damped systems can be found in [11].
Lagrangian (1) contains elements of two different natures
a. Geometric elements, which are the conformal function and the metric
b. Dynamical element, which is the potential specifying the dynamical equations.
As a result the Noether symmetries of this type of Lagrangian relate the invariants of the dynamical system with the geometric elements [11, 12]. This result applies to other type of Lagrangians, for example for autonomous Lagrangians and Lagrangians with a time dependent separable potential of the form [13].
The purpose of the present work is to establish a rigorous main theorem (Theorem (1)) which classifies the Noether invariants of Lagrangians of the form (1). The classification is realized by means of a coupling between temporal parameters and the homothetic algebra of the kinetic metric through a series of compatibility and integrability conditions. It is found that one has to consider two main cases. Case I in which the quantity and case II in which . Each case has subcases depending on the vectors of the homothetic algebra being gradient or non-gradient.
We demonstrate the application of the Theorem to the case of the DDO defined by the potential:
The parameter controls the linear stiffness and the parameter controls the non-linearity of the DDO. Physically the equation of motion of the DDO can be seen as describing the oscillations of a mass attached to a nonlinear spring and a linear damper. The restoring force provided by the nonlinear spring is . Depending on the values one considers the hardening and the softening spring. Detailed information on the damped DO can be found in [8].
The reason for choosing this system is that it is non-linear and, in the forced case, exhibits chaotic behavior; if we obtain a genuine explicit Noether invariant for the unforced system, it can be used to ease the study of the behaviour of the system. Furthermore, because when the potential reduces to the potential of the damped harmonic oscillator (DHO), this study will produce as a byproduct the invariants of that classical non-linear system. Indeed as it will be shown one obtains new invariants for the DHO which include the classic Bateman invariant as a special case.
The structure of the paper is as follows. In section 2 we state the main Theorem. In section 3 we consider the DDO and determine the Noether quantities. The results concern two cases a. The DHO () which gives the Noether quantities of the DHO and b. A new non-trivial invariant of the DDO. In section 4 we verify the invariants both analytically and numerically. Finally in section 5 we draw our conclusions. The proof of the theorem is in the appendix.
The Noether symmetries of time dependent conformal Lagrangian
Theorem 1 Consider a dynamical system described by the time-dependent Lagrangian
with and in Case I we require . The Noether point symmetry generators are of the form
where:
The spatial component is
where is an element of the homothetic algebra of the metric, satisfying ( for a Killing vector, for a homothetic vector), and is the temporal scaling parameter.
The potential satisfies the spatial identity
where the constants are defined by
The temporal scaling parameter satisfies
We distinguish two major cases.
Case I: . In this case , and the temporal component of the generator is
The conformal factor and the temporal scaling satisfy the constraint
Two subcases arise.
Case I-A: . The vector is a gradient vector with generating function , . The constants are determined from
where is a constant. The Noether gauge function is
Case I-B: , . The vector may be gradient or non-gradient. The constant is determined from
the temporal component is
and the gauge function is
Case II: . In this case , and the quadratic constraint forces . Hence is a Killing vector (not necessarily gradient) and . The potential satisfies
Two subcases arise.
Case II-A: is a gradient KV, . The constant is determined from
and satisfies (10). The gauge function is given by (14).
Case II-B: is not a gradient KV. Then , , and the potential satisfies
The gauge function is
In all cases the Noether integral is
In Case II, since , the invariant reduces to , which is linear in the velocities.
The proof of the theorem is given in the appendix.
The Damped Duffing Oscillator (DDO)
We apply Theorem 1 to the DDO with the scaling ():
The associated Euler–Lagrange equations yield the Duffing oscillator:
where . For , this simplifies to the unforced Holmes–Duffing model, and for it reduces to the standard damped harmonic oscillator (DHO).
Homothetic Algebra and Potential Conditions
The homothetic algebra of comprises:
Gradient Killing vectors (translations): with generating function .
Non-gradient Killing vectors (rotations): .
Gradient Homothetic vector (dilatation): with and generating function .
Evaluating these vector fields on gives:
Gradient KVs (translations): .
Non-gradient KVs (rotations): .
Gradient HV (dilatation): .
Case I-A:
In Case I, with , the temporal scaling parameter satisfies
Solving this coupled system yields
The temporal component of the generator is
and the gauge function is
For the dilatation vector , matching coefficients in the potential identity
yields three valid parameter layouts.
Branch 1: (DHO, )
For , the potential identity forces and
The constant . The generator components are
Substituting into the Noether integral (22) gives
This matches equation (32) of [11] for the damped harmonic oscillator (up to an overall sign, which is a convention).
Branch 2: (Bateman invariant, DHO)
For , the exponent in (27) vanishes, so is constant and . The potential is the DHO potential with arbitrary, and . The temporal component is
and the gauge function vanishes (, ). The Noether integral (22) gives
This is the classical Bateman invariant for the damped harmonic oscillator [1, 15].
Branch 3: (new invariant, DDO)
For , the potential identity forces
and the potential is
The temporal scaling parameter and the temporal component are
and the gauge function is
The generator is
Substituting into the Noether integral (22) gives the new exact invariant
This is the main result of the paper: a new exact time-dependent quadratic invariant for the anharmonic DDO, valid when the linear stiffness satisfies . The parameter has dimensions of inverse time, and the condition therefore relates a stiffness (dimensions of frequency squared) to a damping rate squared.
Case I-B: ()
In Case I-B, is constant, and the constraint (12) with gives
so . For the dilatation vector (), this gives , which is Branch 2 (Bateman). For the rotations (), this gives , which is actually Case II (since ). We therefore treat the rotations in § 3.4 below.
Case II:
In Case II, , , and . The potential satisfies
Case II-A: Gradient KV (translations)
For the translations (, ), the potential identity integrates to
With (generating function of the translation), this gives
which forces , , and . The potential is the DHO potential , and satisfies
The gauge function and the invariant follow from (14) and (22). Since these invariants are well known [11, 14], we do not reproduce them here.
Case II-B: Non-gradient KV (rotations)
For the rotations (, ), the potential identity gives . With constant, , and , the invariant (22) reduces to
This is the angular momentum scaled by the factor , holding for all central potentials .
Verification of the Invariants
We verify explicitly the invariants (34), (36), and (42).
Verification of Branch 1
Define
so that . We need to prove . Using (24) with and we compute:
With , the coefficient of simplifies to , and . Substituting we find:
Therefore .
Verification of Branch 2 (Bateman invariant)
Define
so that . We need to prove . Using (24) with :
Substituting,
Therefore .
Verification of Branch 3 (new invariant)
Define
so that . We need to prove . Using (24) with we calculate:
With , the coefficient of becomes , and . Replacing we find:
Therefore .
Numerical verification of the new invariant
To complement the analytical verification of § 4, we integrated the equations of motion (24) numerically for the anharmonic case with , , and , using a generic planar initial condition. Figure 1 shows the new invariant along the trajectory (left panel) and the relative deviation on a logarithmic scale (right panel). The invariant is conserved to machine precision (relative deviation ) throughout the integration, confirming that it is an exact conservation law and not a numerical artifact. For comparison, the right panel also shows the Bateman invariant for the damped harmonic oscillator () with the same and , which is likewise conserved to machine precision.
Figure 1. Left: the new invariant for the anharmonic DDO () and the Bateman invariant for the DHO (), both with and . Right: relative deviation on a logarithmic scale. Both invariants are conserved to machine precision, confirming the analytical results of § 3 and § 4.
Conclusion
In this work, we have provided, by means of Theorem 1, an algorithm for the classification of Noether point symmetries for time-dependent conformal Lagrangians in an -dimensional Riemannian space. There are three quantities that one has to supply to this algorithm:
The conformal factor ;
The metric ;
The potential function .
The conformal factor fixes the time-dependent part of the (linear) friction. The metric specifies the homothetic algebra, and the potential specifies the damped term in the dynamical equations. Theorem 1 does not impose any requirements or limitations on any of the three quantities, and is therefore quite general and covariant.
We demonstrated the application of the theorem in the following configuration:
Figure 2. Projection of the trajectory in the -plane for the anharmonic DDO with , , , and the initial conditions used in Figure 1.
The conformal factor ;
The metric is the Euclidean metric ;
The potential is , which defines the DO.
For this configuration Theorem 1 produced the following results:
All the invariants of the DHO, including the classical Bateman invariant
as a special case (Branch 2).
A new exact time-dependent quadratic invariant for the anharmonic DDO, valid when the linear stiffness satisfies :
This is the main result of the paper. Each invariant in the classification was verified explicitly by direct differentiation along the equations of motion, confirming the validity of Theorem 1.
The angular momentum scaled by the factor , valid for all central potentials (Case II-B).
The new invariant (b) is significant for two reasons. First, it provides a rare example of a global conservation law for a fully nonlinear dissipative system: for an arbitrary DDO the nonlinear coupling typically disrupts point symmetry generators entirely, and the equation can be integrated only through numerical or approximate means. When the linear stiffness satisfies , a dynamic resonance occurs in which the coordinate dilatation vector couples with a time-stretching transform , producing an exact invariant. Second, this invariant can be used for the integration of the dynamical equations of the DDO, or as a Hamiltonian in its quantization, in the same spirit in which the Bateman invariant is used for the DHO.
It is evident that Theorem 1 can be applied to a large number of important nonlinear systems and produce results that would otherwise be difficult to obtain. Future work will extend the application of the theorem to other more general configurations, either in the conformal factor, the metric, or the potential. One such application would be the Cornell potential [16, 17, 18, 19], which is the Coulomb potential plus a linear term. A further extension, currently in preparation, is the general case of a time-dependent separable potential , which opens additional avenues of application.
Proof of Theorem 1
We consider a dynamical system moving on a Riemannian manifold with metric under the action of the time-dependent conformal Lagrangian :
The equation of motion of the system is
where are the Christoffel symbols determined from .
The First Prolongation and Noether Condition
A vector field generates a Noether point symmetry of if it satisfies the Noether condition
where is the boundary gauge function. The first prolongation is given by
where
We compute:
We substitute the above into the Noether condition (62) and collect terms with the same power of .
Cubic Terms:
Symmetrizing the first term (relabel , , ) gives , so
from which follows
Quadratic Terms:
Utilizing , the quadratic velocity coefficients are collected as
Factoring out and using the definition of the Lie derivative of the metric tensor along the spatial generator field , we have
This establishes that
(a) must be a homothetic vector field satisfying ( for a Killing vector, for a homothetic vector).
(b) The temporal functions satisfy the constraint
Linear Terms:
This implies that if the vector is a gradient HV. Let be the generating function so that .
Zeroth-Order Terms:
Equivalently,
The Integrability Condition for
We consider the integrability condition . Taking the spatial derivative of equation (76) gives
Substituting the time derivative of (74), , we obtain
Dividing by we obtain
Since the first term depends only on the spatial coordinates, the coefficients of the second and third terms must reduce to invariant separation constants, which we denote as and :
This leads to the spatial identity governing the potential:
The Noether invariant is
The Constraint Equation
Expanding the derivative in the separation definition for (81) gives
To determine the temporal component , we expand equation (80):
We now have two independent expressions for : the relation from the quadratic velocity collection (73) and this separation identity. Equating them eliminates (recall that ):
hence
We consider cases.
Case I:
Substituting (87) back into the quadratic constraint (73) yields the closed, non-linear differential constraint restricting the conformal factor :
This is a necessary condition for Case I to be consistent. To proceed we consider two cases: and .
Case I-A:
From (74) we have that the vector field must be a gradient vector, allowing us to write . In this case (82) gives
Also equation (84) gives
Equation (87) gives , and (88) remains the same.
The Gauge Function and Noether First Integral. We integrate equation (74) spatially to determine the gauge function up to a temporal remainder :
Differentiating this form with respect to time gives
Substituting this back into the zeroth-order velocity equation (76) yields
Utilizing the separation constant definitions (80) and (81), this reduces to
Using the identity one finds
Therefore
The Noether invariant is given by (83):
Case I-B:
Setting constant in (84) gives . Then (82) becomes
From (74) it follows that . From (88) it follows (with )
The Noether function is computed from the relation (75), from which follows
This results in the two relations
It can be shown that the second relation is identical with (85). Therefore only the first relation survives, which gives
Concerning , from (87) follows
The Noether integral is
It is to be noted that the constraint (88) applies to both cases and .
Case II:
If , then from (87) follows
and from the constraint (73) . Therefore only the KVs survive and .
We now consider the spatial identity (82). With it reads
We distinguish two subcases according to whether is a gradient KV or not.
Subcase II-A: is a gradient KV
If , then (105) integrates to
where is a constant. The coefficient is determined from this relation, and satisfies (10). The gauge function is computed from (96) and the Noether integral from (97).
Subcase II-B: is not a gradient KV
If is not a gradient vector, the linear constraint
requires (since is exact while is not), so that
From (81) it then follows that
The temporal equation for is trivially satisfied, and the spatial identity (105) reduces to
where is a constant. The gauge function is obtained from (75), which in the present case becomes
from which follows
In all cases the Noether integral is given by (22). This completes the proof of Theorem 1.
References
- [1]H. Bateman, On dissipative systems and related variational principles Phys. Rev. 38, 815 (1931).DOI
- [2]E. Kanai, On the Quantization of the Dissipative Systems. Prog. Theor. Phys. 3, 440 (1950).
- [3]H. Lewis, Jr., Class of exact invariants for classical and quantum time-dependent harmonic oscillators J. Math. Phys. 9, 1976–1986 (1968).DOI
- [4]A. Margheri, R. Ortega, C. Rebelo et al., Dynamics of Kepler problem with linear drag, Celest. Mech. Dyn. Astron. 120, 19 (2014).DOI
- [5]A. Margheri, M. Misquero, A dissipative Kepler problem with a family of singular drags, Celest. Mech. Dyn. Astron. 132, 17 (2020).DOI
- [6]H. J. W. Müller-Kirsten, S. Bose, Solution of the wave equation for the logarithmic potential with application to particle spectroscopy J. Math. Phys. 20, 2471 (1979).
- [7]K. U. Kristiansen, Revisiting the Kepler problem with linear drag using the blowup method and normal form theory, Nonlinearity 37, 035014 (2024).arxiv.org/abs/2303.00283
- [8]The Duffing Equation: Nonlinear Oscillators and their Behaviour (Wiley 2011) Editor(s):Ivana Kovacic, Michael J. Brennan DOI:10.1002/9780470977859DOI
- [9]A. H. Salas, C. E. Hernández, J. E. M. Hernández, L. J. Julio, The Duffing Oscillator Equation and Its Applications in Physics, Mathematical Problems in Engineering, 2021, 9994967 (2021). DOI:10.1155/2021/9994967DOI
- [10]A. H. Salas and S. A. El-Tantawy, On the approximate solutions to a damped harmonic oscillator with higher-order nonlinearities and its application to plasma physics: semianalytical solution and moving boundary method The European Physical Journal Plus, 135, 833 (2020).DOI
- [11]M. Tsamparlis, Quantization of the Damped Harmonic Oscillator via the Noether Invariants Symmetry 18, 952 (2026).DOI
- [12]M. Tsamparlis, A. Paliathanasis, Lie and Noether symmetries of geodesic equations and collineations Gen. Rel. Grav. 42, 2957 (2010). DOI 10.1007/s10714-010-1054-9DOI
- [13]L. Karpathopoulos, A. Paliathanasis and M. Tsamparlis, J. Math. Phys. 58, 082901 (2017).
- [14]A. Choudhuri, S. Ghosh, B. Talukdar, Symmetries and conservation laws of the damped harmonic oscillator, Pramana J. Phys. 70, 657 (2008).DOI
- [15]E. H. Kerner, Note on the forced and damped oscillator in Quantum Mechanics Can. J. Phys. 36, 37 (1958).DOI
- [16]E. Eichten, K. Gottfried, T. Kinoshita, J. Kogut, K. Lane, and T. Yan, The spectrum of char-monium, Phys. Rev. Lett. 34, 369 (1975).
- [17]E. Eichten, K. Gottfried, T. Kinoshita, K. Lane, and T. Yan, Charmonium: The model, Phys. Rev. D 17, 3090 (1978).
- [18]E. Eichten, K. Gottfried, T. Kinoshita, K. Lane, and T. Yan, Charmonium: Comparison with experiment, Phys. Rev. D 21, 203 (1980).
- [19]M. Napsuciale, S. Rodríguez, A. Villanueva-Gutiérrez, Complete analytical solution to the Cornell potential and heavy quarkonium structure, Phys. Rev. D 111, 116022 (2025).DOI

