Is a Dissipative System Always a Gradient System or a Gradient Like System?

Rafael Rangel 1 · 2025

Read Is a Dissipative System Always a Gradient System or a Gradient Like System? by Rafael Rangel 1 (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-2509-14409.txt.

Introduction

Rafael Rangel, writing alone in 2025, asks whether each system that loses energy can be read as a gradient system, one that flows downhill on a potential. Some writers have claimed the two classes are one and the same. Plain changes of coordinates seem unable to settle the claim.

Rangel links a differential form to any such dynamics and splits it in two parts. One part comes from a potential, and the rest is a leftover part. When the leftover part is zero, the form counts as closed, which marks the gradient case. When it is not zero, no plain coordinate change can make it closed. A wider kind of coordinate change is then brought in to aim at a gradient form. The puzzle turns into a question about which changes are allowed.

Catalog record

CREATORRafael Rangel 1
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-14409

Browse the full library · About the Tower of Babel

Full text

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

Is a Dissipative System Always a Gradient System or a Gradient Like System? Rafael Rangel 1 Abstract We find that to a given dissipative system a p=1 differential form can be associated which has a general decomposition into a potential term and a non-potential residual part. If the residual part is absent the form is closed and the system is gradient system or gradient like. If it is non-closed, in the differential form approach, it remains non-closed under a variable change of coordinates, i.e., the system is not a gradient one or a gradient like in any coordinate system. On the other hand, there are claims that a potential should always exists, i.e., the class of dissipative systems and the the class of gradient systems should coincide. We fix this conundrum by introducing a generalized change of coordinates that aims a transformation to a gradient system or a gradient like system. The condition of being closed in the new coordinates of a certain, through the generalized change of coordinates defined differential form, results in a nonlinear differential equation together with a consistency condition. We give examples of physical systems where an analytical solution for the transformation can be found, and hitherto, the potential, but even when the potential is not accessible analytically, we find that it always exist, and therefore we give in principle an affirmative answer to the defining question of this work. Our findings removes loopholes in the question if a potential may exist but it is not known. pacs 74.81.Fa,05.10.Gg I Introduction. Dynamical systems represented by a vector \vec{x} in \in R^{N} , and whose evolution is given by stochastic equation(SDE): \dot{\vec{x}}=\frac{d\vec{x}}{dt}=\vec{f}(\vec{x})+\vec{g}(\vec{x})\zeta(\vec{x},t) (1) constitutes a universal matter of study BerntOksendal . Here each of the functions f(\vec{x}) and g(\vec{x}) has sufficient well behavior such that ODE to have solutions, and the function \zeta(t) represent an stochastic process (noise) CramerLeadbetter taken to be a Markov process. This equation without the noise term, has been intensively studied for the nature of the dynamics in particular, when the system is dissipative GuckenheimerHolmes , StephenWiggins HCartanI . Another issue is concerned with the existence of Lyapunov functions and its role in the classification of the dynamics MSanMiguelRToral . If there exist a scalar function V(\vec{x})=V(x_{i}) , i=1,..N , which is usually called the potential, such that it is bounded from below, and stationary on the limits set of the dynamical systems GuckenheimerHolmes and such that: \dot{\vec{x_{i}}}=f_{i}(\vec{x})=-\frac{\partial V(\vec{x})}{\partial x_{i}}\Rightarrow\frac{dV(\vec{x})}{dt}=-\sum_{i=1}^{N}\frac{\partial V(\vec{x})}{\partial x_{i}}\frac{\vec{x}_{i}}{dt}=\sum_{i=1}^{N}\left(\frac{\partial V(\vec{x})}{\partial{x_{i}}}\right)^{2}\leq 0, (2) the potential is called a Lyapunov function for the system GuckenheimerHolmes . One can consider a more general situation given in which in the governing dynamics a residual vector function {\vec{\it{v}}} appears, in addition to a positive definite matrix \mathcal{S}(\vec{x}) : \displaystyle\dot{\vec{x_{i}}}=\vec{f}_{i}(\vec{x}):=-\sum_{j=1}^{N}\mathcal{S}_{ij}\frac{\partial\it{V}(\vec{x})}{\partial x_{j}}+\vec{\it{v}}_{i}(\vec{x}), \displaystyle{\vec{\it{v}}}\cdot\nabla\vec{V}=0 (3) i.e., the vector functions {\vec{\it{v}}} and \vec{\nabla}V(\vec{x}) are orthogonal in the configuration space U \subset R^{N} (orthogonality condition). In this case the potential \it{V}(\vec{x}) still behaves as a Laypunov function, i.e., \frac{dV(\vec{x})}{dt}\leq 0 . The orthogonality condition can be written, by directing rewriting it in the form: \displaystyle\left(\vec{f}(\vec{x})+\mathcal{S}\vec{\nabla}V(\vec{x})\right)^{T}\cdot\vec{\nabla}V(\vec{x})=\vec{f}^{T}(\vec{x})\vec{\nabla}V(\vec{x})+(\vec{\nabla}V(\vec{x}))^{T}\mathcal{S}\vec{\nabla}V(\vec{x})=0 (4) Previous equation, giving the orthogonality condition and the function \vec{f}(\vec{x}) , can be though as an equation for determining V(\vec{x}) DezaRRWioHS . This is Graham and Tel GrahamTel approach, where the existence of a globally defined potential function with respect to the positive define matrix \mathcal{S} is defined. The residual part {\vec{\it{v}}} , does not participate in the relaxation dynamics. When considering Equ.( 1 ) for the special case of a Gaussian random processes, i.e.: \displaystyle\dot{\vec{x}_{i}}=\vec{f}_{i}(\vec{x})+\sum_{j}^{N}\vec{g}_{ij}(\vec{x})\zeta_{j}(t);\hskip 28.45274pt<\zeta_{i}(t)\zeta_{j}(t)>=2\epsilon\delta(t-t^{\prime}), \displaystyle\vec{f}_{i}=-\sum_{j=1}^{N}\mathcal{S}_{ij}\frac{\partial\it{V}(\vec{x})}{\partial x_{j}}+\vec{\it{v}}_{i}\hskip 184.9429pt (5) The Ansatz for \vec{f}_{i} in previous equation fulfills the orthogonality condition. Now, the question for finding the potential (see WioHSDezaRR for a review), when considering the associated Fokker-Planck to Equ.( 5 ), and other assumptions, namely, the validity of the fluctuation dissipation relation for the case when \mathcal{S} is a constant matrix, when \mathcal{S}=gg^{T} , and the validly of the orthogonal relation and the divergenz free of \vec{v} , then there exist a stationary distribution P_{stat} for the associated Fokker-Plank-equation given by in the small noise limit (lim \epsilon\rightarrow 0 ): P_{stat}=Nexp\left(-\frac{V(\vec{x})}{\epsilon}\right) (6) from which one could obtain the potential if P_{stat} could be independently obtained, as the limit: V(\vec{x})=lim_{\epsilon\rightarrow 0}(-\epsilon\ln{P_{stat}}) (7) also known as Graham potential. Graham and Tel GrahamTel have pioneered the question for the general conditions for a potential to exists for dissipative dynamical systems. They have shown that both ways of reasoning are equivalent, i.e., the potential obtained from Equ.( 4 ) and the one from Equ.( 7 ). They also study the concrete question of sufficient conditions for the existence of a potential, given a function f(\vec{x}) . The method bases on solving the question of the integrability of a formal associated of a Hamilton systems to the Equ.( 8 ). However, we demonstrate in the appendix, that their scheme is just the condition for certain diff-form to be closed WioHSDezaRR , or equivalently, to posses an integrating factor. Also, we show below that for every dissipative systems there is a decomposition quantitatively similar to equation Eqn.( 3 ). Remarkably, the question if a given dissipative systems is a gradient like systems also find relevance in mathematics. There are claims RalphChillEvaFasargova that with high probability one will not be able to avert that question. However, the authors Chill and Fasarngova do not shrug off the importance of the question, on the contrary, they express hope by saying that they seem to see gradient systems every where and reaffirm the importance of the question by saying that they hail a gradient system whenever they see one. But their main inquires are how to show that a dissipative systems is a gradient systems and how to find the potential when one feels it exist?. In this work we analyse this question RalphChill , specifically for systems like: \beta_{c_{j}}\ddot{x}_{j}+\dot{x}_{j}+g_{j}(x_{1},\ldots,x_{N+1})+\zeta_{j}(t)=0;\hskip 14.22636ptj=1,\ldots,N+1 (8) These are in fact a sufficient general systems which contain interesting limits. In this system, the x_{j} variables defines a vector in \vec{x}\in R^{N+1} , and the functions \vec{g}=g_{j}(x_{1},\ldots,x_{N+1}) defines the vector function \vec{g}\in R^{N+1} , these functions are assumed to be at least twice continuous differentiable, such that existence and uniqueness of classical solutions the ODE system is guarantee HirschSmaleDevaney . In addition, the last term represents a white noise term, ever present in physical systems, therefore we are dealing with stochastic differential equations. The parameters \beta_{cj} is a measure of an inertia like concept in physics. In cases where the noise term can be neglected, justified for example at low enough temperatures, one can analyze the system by the flow properties of its associated first order system: \displaystyle\vec{\beta}_{c}\dot{\vec{x}}=\vec{\bar{x}} \displaystyle\dot{\vec{\bar{x}}}=-\frac{1}{\beta_{c}}\vec{\bar{x}}-\vec{g}(\vec{x}) (9) where \vec{\beta}_{c}=\beta_{1},....,\beta_{N+1} , \vec{x}=(x_{1},..,x_{N+1}) , \vec{\bar{x}}=(x_{N+1},..,x_{2(N+1)}) are vectors \in R^{N+1} GuckenheimerHolmes StephenWiggins . If there exist a function , such that \vec{\nabla}V(\vec{x})=\vec{g}(\vec{x}) , we say the system is gradient like, but note that It is not a gradient system. Here we are in the case of a the dynamical system which is phase space contracting like the systems treated in GrahamTel and write for it for short: \dot{x_{i}}=X_{i} . For negligible inertia, one has the over-damped limit: \dot{\vec{x}}=-\vec{g}(\vec{x}) . On the other hand, if the dissipative term (second term in Equ.( 8 )), or second term in the

Read the complete text