A minimax argument to a stronger version of the Jacobian conjecture

Wei LiuarXiv 2020 · DOI · arXiv

Abstract

The main result of this paper is to prove the strong real Jacobian conjecture under the symmetric assumption and reveals the link between it and the Jacobian conjecture. Precisely, we assume that $F: \mathbb{R}^n \to \mathbb{R}^n$ is of $C^1$ map, $n\geqslant 2$, if for some $\varepsilon >0$, $ 0\notin Spec(F)~~\mbox{and}~ Spec(F+F^T) \subseteq (-\infty,-\varepsilon)~\mbox{or} ~(\varepsilon,+\infty),$ where $Spec (F)$ denotes all eigenvalues of $JF$ and $Spec (F+F^T)$ denotes all eigenvalues of $JF+JF^T$, then we show that $F$ is injective. It is proved by using a minimax argument.

Introduction

The Jacobian Conjecture, first proposed by Ott-Heinrich Keller in 1939, remains one of the most significant open problems in algebraic geometry and polynomial analysis. It asks whether a polynomial map $F: \mathbb{C}^n \to \mathbb{C}^n$ with a constant non-zero Jacobian determinant must necessarily possess a polynomial inverse. A related problem, the Real Jacobian Conjecture (RJC), investigates whether a differentiable map $F: \mathbb{R}^n \to \mathbb{R}^n$ with a non-vanishing Jacobian determinant is injective (one-to-one).

While the complex version remains unsolved, the general Real Jacobian Conjecture was disproven in 1994 by Sergey Pinchuk, who constructed a non-injective polynomial map with a non-vanishing Jacobian determinant. This counterexample shifted the research focus toward identifying the specific additional conditions under which injectivity is guaranteed. This paper contributes to this effort by applying variational methods—specifically a minimax argument—to prove a strengthened version of the RJC under a symmetric assumption.

Theoretical Framework: The Symmetric Assumption

The core of this research lies in the analysis of the Jacobian matrix $JF$ and its symmetric part, $JF + JF^T$. The author explores the relationship between the eigenvalues of these matrices and the global injectivity of the map $F$.

The "symmetric assumption" requires that the eigenvalues of the symmetric part of the Jacobian are bounded away from zero. Specifically, for a $C^1$ map $F: \mathbb{R}^n \to \mathbb{R}^n$, we consider the spectrum $Spec(F+F^T)$. If these eigenvalues are contained within $(-\infty, -\varepsilon)$ or $(\varepsilon, +\infty)$ for some $\varepsilon > 0$, the map exhibits a form of uniform definiteness that precludes the "leakage" of eigenvalues toward zero—a phenomenon observed in Pinchuk’s counterexample.

Methodology: The Mountain Pass Theorem

To prove injectivity, the paper employs a minimax argument centered on the Mountain Pass Theorem, a staple of nonlinear analysis and the study of partial differential equations. The proof proceeds via contradiction:

  1. Assumption of Non-Injectivity: Suppose there exist two distinct points $a$ and $b$ such that $F(a) = F(b)$.
  2. Functional Construction: A functional $I(X)$ is constructed based on the map $F$. If $F$ is not injective, the geometry of this functional resembles a "valley" between two distinct peaks (the points $a$ and $b$).
  3. Minimax Principle: According to the Mountain Pass Theorem, if such a geometry exists, there must be a critical point where the derivative $I'(X) = 0$.
  4. Contradiction via Eigenvalues: The author demonstrates that the assumed bounds on the eigenvalues of $JF$ and $JF + JF^T$ make the existence of such a critical point impossible. The "landscape" defined by the functional cannot support a mountain pass or saddle point under these constraints, forcing the conclusion that $F$ must be injective.

Key Results

The paper establishes several pivotal results that bridge the gap between variational analysis and the Jacobian Conjecture:

  • Theorem 1.6: Establishes that for a $C^1$ map $F: \mathbb{R}^n \to \mathbb{R}^n$, if $0 \notin Spec(F)$ and the eigenvalues of $JF + JF^T$ are strictly bounded away from zero (either all less than $-\varepsilon$ or all greater than $\varepsilon$), then $F$ is injective.
  • Theorem 1.7: Provides a refinement for cases where eigenvalues might approach zero, offering specific bounds on the sums and products of eigenvalues that still guarantee injectivity.
  • Proposition 1.3: Demonstrates a formal link between the Real Jacobian Conjecture and the original Jacobian Conjecture. It shows that proving a conjecture by Chamberland and Meisters (regarding maps with eigenvalues bounded away from zero) is a sufficient condition to prove the general Jacobian Conjecture.

Conclusion

By introducing the Mountain Pass Theorem to the study of the Jacobian Conjecture, this paper provides a powerful new toolset for analyzing map injectivity. The results highlight that the symmetric part of the Jacobian matrix is a critical factor in determining global behavior. While the symmetric assumption is a specific condition that does not yet cover all polynomial maps, it successfully narrows the search space for a general proof and provides a clear mathematical path toward solving one of mathematics' most enduring puzzles.

Summary updated

Abstract

The main result of this paper is to prove the strong real Jacobian conjecture under the symmetric assumption and reveals the link between it and the Jacobian conjecture. Precisely, we assume that $F: \mathbb{R}^n \to \mathbb{R}^n$ is of $C^1$ map, $n\geqslant 2$, if for some $\varepsilon >0$, $ 0\notin Spec(F)~~\mbox{and}~ Spec(F+F^T) \subseteq (-\infty,-\varepsilon)~\mbox{or} ~(\varepsilon,+\infty),$ where $Spec (F)$ denotes all eigenvalues of $JF$ and $Spec (F+F^T)$ denotes all eigenvalues of $JF+JF^T$, then we show that $F$ is injective. It is proved by using a minimax argument.
BibTeX
@misc{liu2020minimax,
  author = {Wei Liu},
  title = {{A minimax argument to a stronger version of the Jacobian conjecture}},
  year = {2020},
  eprint = {2009.05464},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2009.05464},
}
Figure