Linear Operators on Polynomials and $K$-Positivity Preserver -- A journey from analysis to algebra and back

1. The Problem: The Gap in Real Algebraic Geometry

The central problem addressed in this work is the study of linear operators $T: \mathbb{R}[x_1, \dots, x_n] \to \mathbb{R}[x_1, \dots, x_n]$ that preserve the cone of polynomials non-negative on a closed set $K \subseteq \mathbb{R}^n$.

While Real Algebraic Geometry has extensively studied the properties of the cone $\text{Pos}(K)$ (the set of polynomials $f$ such that $f(x) \geq 0$ for all $x \in K$), the behavior of linear operators acting upon these cones has remained largely unexamined. The core challenge is that while the definition of a $K$-positivity preserver ($T\text{Pos}(K) \subseteq \text{Pos}(K)$) appears purely algebraic, characterizing such operators requires sophisticated tools from functional analysis (moment theory, Choquet theory) and Lie theory (infinite-dimensional Lie groups).

2. Methodology: A Multi-Disciplinary Approach

The author employs a "journey from analysis to algebra and back" to bridge the gap between the algebraic structure of polynomials and the analytic nature of positivity. The methodology is structured into four distinct phases:

• Phase I: Moment Theory (Analysis): The author utilizes the duality between non-negative polynomials and moment functionals. By treating the existence of a representing measure as a criterion for positivity (via Haviland’s Theorem), the problem of operator positivity is transformed into a problem of characterizing moment sequences.
• Phase II: Topological Vector Spaces (Functional Analysis): To handle infinite sums of operators, the work establishes rigorous topological frameworks. It utilizes Fréchet spaces ($\mathbb{C}^{\mathbb{N}_0^n}$) and LF-spaces (the inductive limit of finite-dimensional polynomial spaces, $\mathbb{C}[x_1, \dots, x_n]$) to ensure convergence and continuity of operators and functionals.
• Phase III: Operator Representation (Algebra): The author develops canonical representations for linear operators, specifically focusing on diagonal operators and operators with constant coefficients. This allows for the use of differential calculus and power series.
• Phase IV: Lie Theory (Algebraic Dynamics): To study the "generators" of these operators (the $A$ such that $T = \exp(tA)$), the author employs the theory of Regular Fréchet Lie Groups. This allows the study of semi-groups of positivity preservers through their infinitesimal generators.

3. Key Contributions and Results

A. Characterization of $K$-Positivity Preservers

The most significant result is the characterization of general $K$-positivity preservers. For an operator $T = \sum q_\alpha \partial^\alpha$, $T$ preserves $\text{Pos}(K)$ if and only if, for every $y \in K$, the sequence $(\alpha! \cdot q_\alpha(y))_{\alpha \in \mathbb{N}_0^n}$ is a $(K-y)$-moment sequence. This links the coefficients of the differential operator directly to the geometry of the set $K$ via moment theory.

B. Generators of Positivity Preserving Semi-groups

The work provides a complete characterization of generators for the case of constant coefficients ($K = \mathbb{R}^n$). Using the Lévy–Khinchin formula, the author proves that an operator $A$ generates a positivity-preserving semi-group if and only if its coefficients correspond to the parameters of an infinitely divisible measure (drift, diffusion/Gaussian part, and a Lévy measure).

C. The Structure of the Set $\mathfrak{d}$

The author defines and characterizes the set $\mathfrak{d}$—the collection of all linear operators $A$ for which the exponential map $\exp(tA)$ is well-defined on the space of polynomials. A key finding is that $\mathfrak{d}$ is a cone but is not convex, meaning the sum of two operators that generate well-defined flows does not necessarily generate a well-defined flow.

D. Discovery of Eventually Positive Semi-groups

The notes provide novel examples of eventually positive semi-groups. These are operators that do not preserve positivity for small $t > 0$ but do preserve it for all $t \geq \tau$ (where $\tau$ is a threshold). The author provides explicit, low-dimensional examples (e.g., on $\mathbb{R}[x]_{\leq 2}$) where the transition to positivity can be calculated via the eigenvalues of Hankel matrices.

4. Significance

This work is significant because it provides a unified framework for a previously fragmented field. By integrating Real Algebraic Geometry with Infinite-Dimensional Lie Theory, it moves beyond the study of static polynomial cones to the study of the dynamical systems (semi-groups) that act upon them.

The results have implications for:

1. Stochastic Processes: The connection to infinitely divisible measures links these operators to the generators of Lévy processes.
2. Numerical Analysis: The characterization of operators via finite-dimensional invariant subspaces provides algorithmic pathways for computing operator exponentials.
3. Mathematical Physics: The study of generators and their evolution equations ($\partial_t p = Ap$) provides a rigorous foundation for studying positivity-preserving evolution in polynomial spaces.