A Moser-Type Construction for the Liouville Equation
A. Borzì, M. Caponigro, A. Vicari · 2026
Read A Moser-Type Construction for the Liouville Equation by A. Borzì, M. Caponigro, A. Vicari (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-2603-21850.txt.
Introduction
Borzì and co-authors propose a Moser-type construction for the kinetic Liouville equation, built on interpolation that follows the equation's characteristics. The Liouville equation tracks how a distribution of particles evolves under given forces. For the special case of Hamiltonian accelerations, their construction simplifies. The problem reduces to a family of weighted elliptic problems in the velocity variable alone. The authors also derive the kinetic compatibility condition that the construction must satisfy. The result is a new route to solutions with the right structure.
Catalog record
| CREATOR | A. Borzì, M. Caponigro, A. Vicari |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21349 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.21850
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Moser-Type Construction for the Liouville Equation A. Borzì, M. Caponigro, A. Vicari Abstract We present a novel extension of Moser’s volume form lemma to the kinetic Liouville equation. In particular, we show that two smooth, positive phase-space densities f f and g g can be connected in unit time by the Liouville equation if and only if a natural compatibility condition on velocity marginals is satisfied. Under this condition, an explicit family of force fields is constructed via a weighted elliptic problem in the velocity variable. Results of numerical experiments are presented to validate the theoretical framework. 1 Introduction In the seminal work [ 10 ] , J. Moser shows that on a compact manifold without boundary, any pair of smooth and positive volume forms of equal total mass can be connected by a diffeomorphism. Later developments of this result, in particular [ 7 ] , provide a constructive approach based on interpolating the densities and solving an elliptic Neumann problem for a time-dependent vector field whose flow realizes the desired transport. More precisely, let \Omega\subset\mathbb{R}^{n} be a bounded connected domain with smooth, i.e. C^{\infty} , boundary. Let f,g\in C^{\infty}(\overline{\Omega}) satisfy f>0 , g>0 on \overline{\Omega} and be such that \int_{\Omega}f(x)\,dx=\int_{\Omega}g(x)\,dx. Then, Moser’s volume forms lemma [ 10 ] states the existence of a diffeomorphism \Gamma:\overline{\Omega}\to\overline{\Omega} pointwise fixing the boundary, such that \Gamma_{\#}(f\,dx)=g\,dx , namely f(x)=g(\Gamma(x))\det D\Gamma(x),\quad\mbox{ for every }x\in\Omega. A possible strategy to construct such a diffeomorphism, following the approach of Dacorogna and Moser [ 7 ] (see also [ 11 , Box 4.3] ), is to connect the initial and final densities by a smooth path, for instance,the linear interpolation \rho(\cdot,t)=(1-t)f+tg,\qquad t\in[0,1]. The goal is then to find a time-dependent velocity field b(x,t) generating this path as a solution of the continuity equation \partial_{t}\rho(x,t)+\nabla\cdot\big(b(x,t)\rho(x,t)\big)=0,\qquad\rho(\cdot,0)=f. (1) Hence, denoting by \Gamma_{t} the flow associated with the vector field b , the solution of the continuity equation ( 1 ) can be written as \rho(\cdot,t)dx=(\Gamma_{t})_{\#}(f\,dx) . In particular, the 1 -time flow \Gamma_{1} pushes forward the initial density onto the final one (\Gamma_{1})_{\#}(f\,dx)=g\,dx, providing the diffeomorphism solving the prescribed transport problem. In the present work, we apply this construction to the case of phase-space densities of Hamiltonian systems. In this setting, the continuity equation governing the evolution of densities takes the form of \partial_{t}\rho(x,v,t)+v\cdot\nabla_{x}\rho(x,v,t)+\nabla_{v}\cdot\big(a(x,v,t)\,\rho(x,v,t)\big)=0, (2) where (x,v)\in{\mathbb{R}}^{n}\times{\mathbb{R}}^{n} . More precisely, the problem is formulated as follows. Let \Omega=X\times V , with X={\mathbb{R}}^{n} or {\mathbb{T}}^{n} ( n -dimensional torus) and V\subset{\mathbb{R}}^{n} . Let f,g\in C^{\infty}(\overline{\Omega}) be strictly positive and such that \int_{\Omega}f(x,v)\,dx\,dv=\int_{\Omega}g(x,v)\,dx\,dv. Consider the Liouville equation ( 2 ) with initial condition \rho(0)=f . Find a:\Omega\times(0,1)\to{\mathbb{R}}^{n} such that the associated solution \rho(t) satisfy \rho(1)=g . Equation ( 2 ), called Liouville Equation by L. Boltzmann in his lectures [ 4 ] , plays a fundamental role in kinetic theory and in the statistical description of non-interacting many-particle systems, where \rho denotes the normalized material density function. At the microscopic level, the particles’ dynamics is described by the differential system \begin{cases}\chi^{\prime}(t)&=\upsilon(t),\\ \upsilon^{\prime}(t)&=a(\chi(t),\upsilon(t),t),\end{cases} (3) where the spatial drift is fixed to be the velocity variable \upsilon , and the force field, which is the only degree of freedom, may be non-conservative. This kinetic formulation represents a significant obstruction in the direct application of Moser’s approach. Indeed, in our framework, the pair of densities (f,g) that can be be connected in finite time by a trajectory of the Liouville Equation ( 2 ) must satisfy an additional constraint on the velocity marginals. The problem has been studied from various perspectives in the context of the continuity equation and, more recently, within control theory [ 1 , 2 , 5 ] . To the best of our knowledge, the corresponding problem for the Liouville equation ( 2 ), defined on the phase space, is yet to be studied. In this paper we present a solution of this problem with a Moser-type construction adapted to the Liouville equation. We present a compatibility condition (see ( 6 ) below) and we construct an acceleration/force field a defined by an elliptic equation in the velocity variable, finally proving that the Liouville dynamics with drift (v,a) connects f to g in unit time. In this work we bridge the classical Moser approach with kinetic transport, opening the possibility of geometric transport methods in kinetic PDEs [ 3 , 6 ] . In the following section, we present our main result. In Section 3 , we present analytical and numerical validation of our framework. A section of conclusion completes this work. 2 Main result Throughout this section we consider X=\mathbb{T}^{n} , the flat n -torus, periodic in x , and V\subset\mathbb{R}^{n} bounded connected with C^{\infty} boundary \partial V and outward unit normal n_{v} . Let \Omega=X\times V . We consider the Liouville equation \partial_{t}\rho+v\cdot\nabla_{x}\rho+\nabla_{v}\!\cdot(a\,\rho)=0\quad\text{ in }\Omega\times(0,1),\qquad\rho(0)=f, (4) with zero–flux boundary conditions in v (\rho\,a)\cdot n_{v}=0\quad\text{ on }X\times\partial V\times(0,1), (5) and periodic boundary condition in x . Our main result is the following theorem. Theorem 2.1 . Let f,g\in C^{\infty}(\overline{\Omega}) be such that f>0 , g>0 on \overline{\Omega} . A necessary and sufficient condition for the existence of a field a\in C^{\infty}(\overline{\Omega}\times[0,1];{\mathbb{R}}^{n}) in ( 4 ) such that \rho(1)=g , is given by the following compatibility condition \int_{V}f(x,v)\,dv=\int_{V}g(x+v,v)\,dv,\qquad x\in X. (6) More precisely, if ( 6 ) holds, then a is given by a(x,v,t)=\nabla_{v}U_{t}(x-tv,v), (7) where U_{t} solves, for any x,t fixed, the following weighted elliptic equation in v : \nabla_{v}\!\cdot\!\big(\widetilde{\rho}_{t}(x,v)\,\nabla_{v}U_{t}(x,v)\big)=f(x,v)-g(x+v,v),\qquad\int_{V}\widetilde{\rho}_{t}\,U_{t}\,dv=0, (8) with \widetilde{\rho}_{t}(x,v)=(1-t)f(x,v)+t\,g(x+v,v). The solution of ( 4 ) with ( 7 ) satisfies \rho(0)=f , \rho(1)=g . Notice that the compatibility condition ( 6 ) implies the Moser’s condition for the phase space \Omega=X\times V , \int_{\Omega}f=\int_{\Omega}g. The proof splits into three steps: a comoving reduction, an elliptic Neumann problem in v , and a reconstruction step. In what follows we denote \Phi_{t}(x,v)=(x+tv,v) . Step 1. Comoving reduction Lemma 2.1 (Comoving transform) . For any smooth \rho and a , define \widetilde{\rho}_{t}(x,v):=\rho(\Phi_{t}(x,v),t),\qquad\widetilde{a}_{t}(x,v):=a(\Phi_{t}(x,v),t). Then \rho solves ( 4 )–( 5 ) if and only if \widetilde{\rho}_{t} solves \partial_{t}\widetilde{\rho}_{t}+\nabla_{v}\!\cdot\big(\widetilde{a}_{t}\,\widetilde{\rho}_{t}\big)=0\quad\text{in }\Omega\times(0,1),\qquad(\widetilde{\rho}_{t}\,\widetilde{a}_{t})\cdot n_{v}=0\ \text{ on }X\times\partial V\times(0,1), (9) with \widetilde{\rho}_{0}=f and \widetilde{\rho}_{1}=g\circ\Phi_{1} . Proof. By the chain rule, \partial_{t}\widetilde{\rho}_{t}=(\partial_{t}\rho)\circ\Phi_{t}+(D\rho)\circ\Phi_{t}\,[\partial_{t}\Phi_{t}]=(\partial_{t}\rho+v\cdot\nabla_{x}\rho)\circ\Phi_{t}. Hence, ( 4 ) is equivalent to \partial_{t}\widetilde{\rho}_{t}=-(\nabla_{v}\!\cdot(a\,\rho))\circ\Phi_{t}=-\nabla_{v}\!\cdot(\widetilde{a}_{t}\,\widetilde{\rho}_{t}), which is ( 9 ). The boundary condition is preserved since \Phi_{t} acts only on the spatial variable x , so the velocity v , and therefore \partial V , are unaffected. Notice that, in the chain rule above, the differentiation \widetilde{\rho}_{t}(x,v)=\rho(x+tv,v,t) is done with respect to t and not v , so the only additional term comes from the time dependence of the x -argument, namely v\cdot\nabla_{x}\rho . Furthermore, the map \Phi_{t} leaves the velocity variable v unchanged; consequently, (\nabla_{v}\cdot(a\rho))\circ\Phi_{t}=\nabla_{v}\cdot\big((a\circ\Phi_{t})(\rho\circ\Phi_{t})\big), or, in other words, \nabla_{v}\cdot commutes with pullback along \Phi_{t}. ∎ Step 2. Necessity of the compatibility condition Lemma 2.2 . If ( 9 ) admits a smooth solution \widetilde{\rho}_{t} with \widetilde{\rho}_{0}=f and \widetilde{\rho}_{1}=g\circ\Phi_{1} , then ( 6 ) holds. Proof. Integrate ( 9 ) in v and use the zero-flux condition ( 5 ) \frac{d}{dt}\int_{V}\widetilde{\rho}_{t}(x,v)\,dv=-\int_{V}\nabla_{v}\!\cdot(\widetilde{a}_{t}\,\widetilde{\rho}_{t})\,dv=-\int_{\partial V}(\widetilde{\rho}_{t}\widetilde{a}_{t})\cdot n_{v}\,dS_{v}=0. Hence, M(x):=\int_{V}\widetilde{\rho}_{t}(x,v)\,dv is independent of t . Evaluating at t=0 and t=1 gives \int_{V}f(x,v)\,dv=\int_{V}\widetilde{\rho}_{1}(x,v)\,dv=\int_{V}g(x+v,v)\,dv, as claimed. ∎ Step 3. Elliptic construction in the comoving frame Now, we assume ( 6 ) and construct \widetilde{a}_{t} so that (