[Paper Review] Integrable homogeneous potentials of degree $-1$ in the plane with small eigenvalues

Thierry Combot|arXiv (Cornell University)|Oct 27, 2011

Advanced Differential Equations and Dynamical Systems31 references3 citations

TL;DR

This paper classifies meromorphically integrable homogeneous potentials of degree $-1$ in the plane under specific spectral constraints at Darboux points. Using Morales-Ramis theory and differential Galois theory, it proves that such potentials must be of the form $V = \frac{a}{r}$ or $V = \frac{a}{q_1} + \frac{b}{q_2}$, or a complex variant involving $q_1 + \epsilon i q_2$, and conjectures this list is complete for all eigenvalues in the non-degenerate case.

ABSTRACT

We give a complete classification of meromorphically integrable homogeneous potentials $V$ of degree $-1$ which are real analytic on $\mathbb{R}^2\setminus \{0\}$. In the more general case when $V$ is only meromorphic on an open set of an algebraic variety, we give a classification of all integrable potentials having a Darboux point $c$ with $V'(c)=-c,\; c_1^2+c_2^2 eq 0$ and $\hbox{Sp}( abla^2 V(c)) \subset\{-1,0,2\}$. We eventually present a conjecture for the other eigenvalues and the degenerate Darboux point case $V'(c)=0$.

Motivation & Objective

• To classify all meromorphically integrable homogeneous potentials of degree $-1$ in the plane under spectral constraints at Darboux points.
• To extend the applicability of Morales-Ramis-Simo theory to algebraic potentials beyond polynomial ones, including $n$-body type systems.
• To address the open problem of whether integrable potentials of degree $-1$ exist outside the known families, particularly for eigenvalues not in $\{-1, 0, 2\}$.
• To provide a complete classification under the assumption of non-degenerate Darboux points with restricted Hessian spectra.
• To conjecture that no further integrable potentials exist beyond the listed families, based on numerical evidence and structural analysis of higher variational equations.

Proposed method

• Applies the Morales-Ramis-Simo theorem to variational equations along straight-line orbits in homogeneous potentials, requiring the differential Galois group of the variational equation to have Abelian identity component.
• Uses the framework of algebraic potentials defined on a complex algebraic variety $\mathcal{S} = \{(q_1, q_2, r) \mid r^2 = q_1^2 + q_2^2\}$, allowing rational and algebraic potentials including $V = r^{-1}$.
• Imposes the condition $V'(c) = -c$ and $c_1^2 + c_2^2 \neq 0$ to identify Darboux points, which are critical for applying the integrability criterion.
• Analyzes the spectrum of the Hessian $\nabla^2 V(c)$ at Darboux points, restricting to $\operatorname{Sp}(\nabla^2 V(c)) \subset \{-1, 0, 2\}$ to ensure applicability of the theory.
• Employs recursive algebras $\mathcal{A}_0 = \mathbb{C}[t, (t^2 - 1)^{-1}]$, $\mathcal{A}_{i+1} = \int \mathcal{A}_i dt$ to model the Picard-Vessiot fields of higher variational equations and constrain integrability.
• Uses non-degeneracy conditions on higher variational equations to prove uniqueness and restrict possible integrable forms, especially in the $-1$ eigenvalue case.

Experimental results

Research questions

• RQ1 Which homogeneous potentials of degree $-1$ in the plane are meromorphically integrable under spectral constraints at Darboux points?
• RQ2 Can the classification of integrable potentials be extended beyond polynomial potentials to include algebraic potentials such as those arising in the $n$-body problem?
• RQ3 Are there integrable homogeneous potentials of degree $-1$ with Hessian eigenvalues outside $\{-1, 0, 2\}$ at non-degenerate Darboux points?
• RQ4 What is the role of higher-order variational equations in constraining the form of integrable potentials, especially when standard non-degeneracy fails?
• RQ5 Is the list of integrable potentials provided in Theorem 2 complete, or could other families exist with different eigenvalue spectra?

Key findings

• All real analytic homogeneous potentials of degree $-1$ on $\mathbb{R}^2 \setminus \{0\}$ that are meromorphically integrable must be of the form $V = \frac{a}{r}$ with $a \in \mathbb{R}$, as stated in Theorem 1.
• For holomorphic homogeneous potentials on $\mathcal{S}$ with a non-degenerate Darboux point satisfying $V'(c) = -c$ and $\operatorname{Sp}(\nabla^2 V(c)) \subset \{-1, 0, 2\}$, integrability implies $V$ is equivalent under rotation to one of four families: $\frac{a}{q_1} + \frac{b}{q_2}$, $\frac{a}{r}$, or $\frac{a q_1}{(q_1 + \epsilon i q_2)^2}$ with $a,b \in \mathbb{C}$, $\epsilon = \pm 1$, as per Theorem 2.
• The unexpected form $V = \frac{a q_1}{(q_1 + \epsilon i q_2)^2}$ was previously found by Hietarinta and is now shown to be the only new integrable family under the spectral condition.
• The paper proves that Theorem 2 implies Theorem 1 via a reduction argument based on the structure of the potential and its behavior at infinity.
• A non-degeneracy condition on higher variational equations is shown to imply uniqueness of the solution space, which is crucial for ruling out additional integrable forms.
• Based on extensive numerical computations, the author conjectures that no integrable homogeneous potentials of degree $-1$ exist outside the families listed in Theorem 2, even for other eigenvalues, suggesting the classification is complete.