[Paper Review] On the structure of mu-classes

Carlos D’Andrea|ArXiv.org|Apr 15, 2002

Polynomial and algebraic computation4 references3 citations

TL;DR

This paper proves that every rational parametrization of degree $n$ and class $\mu < \lfloor n/2 \rfloor$ is a limit of parametrizations of the same degree and class $\mu+1$, confirming a conjecture in the literature. The authors construct explicit one-parameter deformations (approximating sequences) in $\mathbb{K}[\epsilon,t]^3$ that preserve degree and gcd while increasing the class by one, using syzygy theory and coordinate transformations to ensure the class increases precisely to $\mu+1$. This result fully describes the Zariski closure of the variety of parametrizations of fixed degree and class.

ABSTRACT

We prove that, if μ

Motivation & Objective

• To resolve a conjecture in algebraic geometry regarding the structure of $\mu$-classes in rational parametrizations.
• To establish that the Zariski closure of the variety $\mathcal{P}_n^\mu$ equals the union $\mathcal{P}_n^0 \cup \cdots \cup \mathcal{P}_n^\mu$ for all $\mu \leq \lfloor n/2 \rfloor$, which implies the irreducible components are nested.
• To construct explicit approximating sequences in $\mathbb{K}[\epsilon,t]^3$ that deform a given triple $(a,b,c)$ of class $\mu$ into a triple of class $\mu+1$ while preserving degree and gcd.
• To provide a complete geometric description of the moduli space of rational parametrizations of degree $n$ and class $\mu$ by proving the closure of $\mathcal{P}_n^\mu$ is the union of all lower-class varieties.

Proposed method

• Use of the syzygy module $\operatorname{Syz}(a,b,c)$ to characterize the class $\mu(a,b,c)$ as the minimal degree of a nontrivial syzygy.
• Leverage Theorem 2.1 to express $a,b,c$ in terms of two polynomial triples $(p_x,p_y,p_w)$ of degree $\mu$ and $(q_x,q_y,q_w)$ of degree $n-\mu$, which generate all syzygies.
• Apply a linear transformation $a \mapsto a + \lambda b$ to achieve genericity: ensure that for some $\lambda$, the polynomials $p_y - \lambda p_x$ and $p_w$ are coprime, which avoids unwanted common roots.
• Construct an approximating sequence $ (a_\epsilon, b_\epsilon, c_\epsilon) $ by shifting the parameter $t$ to $t - \alpha + \epsilon$ for a root $\alpha$ of $a$ that is not a common root of $p_y$ and $p_w$, ensuring the new syzygy has degree $\mu+1$.
• Prove that $\mu(a_\epsilon, b_\epsilon, c_\epsilon) = \mu + 1$ by contradiction: assume a syzygy of degree $\leq \mu$ exists, then use the unique decomposition from Theorem 2.1 to derive a contradiction via degree bounds and coprimality.
• Verify that the deformation preserves $\gcd(a_\epsilon, b_\epsilon, c_\epsilon) = 1$ and $\deg(a_\epsilon, b_\epsilon, c_\epsilon) = n$ for generic $\epsilon$, ensuring the family lies in $\mathcal{P}_n^{\mu+1}$.

Experimental results

Research questions

• RQ1 Is every rational parametrization of degree $n$ and class $\mu < \lfloor n/2 \rfloor$ a limit of parametrizations of class $\mu+1$?
• RQ2 Can the Zariski closure of the variety $\mathcal{P}_n^\mu$ be explicitly described as the union $\mathcal{P}_n^0 \cup \cdots \cup \mathcal{P}_n^\mu$?
• RQ3 Does there exist a constructive deformation (approximating sequence) in $\mathbb{K}[\epsilon,t]^3$ that increases the class by one while preserving degree and gcd?
• RQ4 Can the class of a rational parametrization be increased via a one-parameter family that converges to the original parametrization as $\epsilon \to 0$?
• RQ5 What conditions on the syzygy generators $p_x,p_y,p_w$ and $q_x,q_y,q_w$ ensure that the class increases under deformation?

Key findings

• The conjecture that $\overline{\mathcal{P}_n^\mu} = \mathcal{P}_n^0 \cup \cdots \cup \mathcal{P}_n^\mu$ holds for all $\mu \leq \lfloor n/2 \rfloor$, confirming the nested structure of the closure of $\mu$-classes.
• For every $(a,b,c) \in \mathcal{P}_n^\mu$ with $\mu < \lfloor n/2 \rfloor$, there exists an approximating sequence $ (a_\epsilon, b_\epsilon, c_\epsilon) \in \mathbb{K}[\epsilon,t]^3 $ such that $\mu(a_\epsilon, b_\epsilon, c_\epsilon) = \mu + 1$, $\gcd(a_\epsilon, b_\epsilon, c_\epsilon) = 1$, and $\deg(a_\epsilon, b_\epsilon, c_\epsilon) = n$, with $ (a_\epsilon, b_\epsilon, c_\epsilon)|_{\epsilon=0} = (a,b,c) $.
• The construction of the approximating sequence relies on choosing a root $\alpha$ of $a$ that is not a common root of $p_y$ and $p_w$, ensuring the deformation increases the class by exactly one.
• The proof shows that any hypothetical syzygy of degree $\leq \mu$ in the deformed system would force $p_y$ and $p_w$ to share a common root, contradicting the genericity condition, thus proving the class increases to $\mu+1$.
• The method is robust: for any $(a,b,c) \in \mathcal{P}_n^\mu$, a suitable $\lambda \in \mathbb{K}$ can be chosen so that $a + \lambda b$ has a root $\alpha$ not shared by $p_y - \lambda p_x$ and $p_w$, enabling the construction.
• An example is provided where a naive deformation fails to increase the class, but a corrected choice of $\lambda$ and $\alpha$ yields a valid approximating sequence, demonstrating the necessity of the genericity condition.