Scalar q-subresultants and Dickson matrices

From MaRDI portal
Publication:2285223



Abstract: Following the ideas of Ore and Li we study q-analogues of scalar subresultants and show how these results can be applied to determine the rank of an mathbbFq-linear transformation f of mathbbFqn. As an application we show how certain minors of the Dickson matrix D(f), associated with f, determine the rank of D(f) and hence the rank of f.


The concept resultant of two polynomials has applications in number theory and algebraic geometry. It is a basic tool for many computer algebra systems for polynomials over rational numbers. This single-authored article studies the extensions of some well-known results of Ore and Li. \begin{itemize} \item[1.] Classical Resultant: Let $f(x) = \sum ^{k}_{i=0} a_{i}x^{i}$ and $g(x) = \sum ^{l}_{i=0} b_{i}x^{i}$, with $a_{k}b_{l}\neq 0$, be two univariate polynomials with coefficients in the field $K$. In elimination theory, the classical resultant of $f$ and $g$ is \[ Res(f,g)=(-1)^{kl}b_{l}^{k}\prod^{l}_{i=1}f(\zeta_{i}), \] where $g(x) = b_{l} \prod^{l}_{i=1}(x - \zeta_{i})$ with $\zeta_{1}, \zeta_{2}, \dots, \zeta_{l} \in \bar{K}$ (where $\bar{K}$ denotes the algebraic closure of $K$). For $0 \leq m \leq \min\{k, l\}$ consider the following $(k + l - 2m) \times (k + l - 2m)$ matrix: \[ R_{m}(f,g)= \left( \begin{matrix} a_{k} & a_{k-1} & a_{k-2} & \cdots & a_{k-l+m+1} &\cdots & a_{2m-l+2} & a_{2m-l+1} \\ 0 & a_{k} & a_{k-1} & \cdots & a_{k-l+m+2} &\cdots & a_{2m-l+3} & a_{2m-l+2} \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ 0 & 0 & 0 & \cdots & a_{k} & \cdots & a_{m+1} & a_{m} \\ b_{l} & b_{l-1} & b_{l-2} & \cdots & b_{l-k+m+1} &\cdots & b_{2m-k+2} & b_{2m-k+1} \\ 0 & b_{l} & b_{l-1} & \cdots & b_{l-k+m+2} &\cdots & b_{2m-k+3} & b_{2m-k+2} \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ 0 & 0 & 0 & \cdots & b_{l} & \cdots & b_{m+1} & b_{m} \\ \end{matrix} \right) \] where coefficients out of range are considered to be $0$. The determinant of $R_{m}(f,g)$ is denoted by $|R_{m}(f,g)|$. \item[2.] Scalar $q$-subresultants: Let $f(x) = \sum ^{k}_{i=0} a_{i}x^{q^{i}}$ and $g(x) = \sum ^{l}_{i=0} b_{i}x^{q^{i}}$, with $a_{k}b_{l}\neq 0$, be two univariate $q$-polynomials with coefficients in the field $\mathbb{\bar{F}_q}$ where $q$ is a prime power. Put $q^{\mu} = \operatorname{deg} \gcd(f,g)$. For $0 \leq m \leq \min\{k, l\}$ consider the following $(k + l - 2m) \times (k + l - 2m)$ matrix $R_{m,q}(f,g)$: \[ \left( \begin{matrix} a_{k}^{q^{l-m-1}} & a_{k-1}^{q^{l-m-1}} & a_{k-2}^{q^{l-m-1}} & \cdots & a_{k-l+m+1}^{q^{l-m-1}} &\cdots & a_{2m-l+2}^{q^{l-m-1}} & a_{2m-l+1}^{q^{l-m-1}} \\ 0 & a_{k}^{q^{l-m-2}} & a_{k-1}^{q^{l-m-2}} & \cdots & a_{k-l+m+2}^{q^{l-m-2}} &\cdots & a_{2m-l+3}^{q^{l-m-2}} & a_{2m-l+2}^{q^{l-m-2}} \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ 0 & 0 & 0 & \cdots & a_{k} & \cdots & a_{m+1} & a_{m} \\ b_{l}^{q^{k-m-1}} & b_{l-1}^{q^{k-m-1}} & b_{l-2}^{q^{k-m-1}} & \cdots & b_{l-k+m+1}^{q^{k-m-1}} &\cdots & b_{2m-k+2}^{q^{k-m-1}} & b_{2m-k+1}^{q^{k-m-1}} \\ 0 & b_{l}^{q^{k-m-2}} & b_{l-1}^{q^{k-m-2}} & \cdots & b_{l-k+m+2}^{q^{k-m-2}} &\cdots & b_{2m-k+3}^{q^{k-m-2}} & b_{2m-k+2}^{q^{k-m-2}} \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot \\ 0 & 0 & 0 & \cdots & b_{l} & \cdots & b_{m+1} & b_{m} \\ \end{matrix} \right) \] \end{itemize} Result: The degree of $\gcd(f, g)$ is $t$ if and only if $|R_{0}(f, g)| = \dots = |R_{t-1}(f, g)| = 0$ and $|R_{t}(f, g)| \neq 0$. The author has proved the following main theorem of the article which is the $q$-analogue of the above result. Theorem: The $q$-degree of $\gcd(f, g)$ is $\mu$ if and only if $|R_{0,q}(f, g)| = \dots =|R-{\mu-1,q}(f, g)| = 0$ and $|R_{\mu,q}(f, g)| \neq 0$. The author has extended another result of Ore using the Dickson matrix. I must appreciate the author for his efforts to bring out a very high standard article in Algebra. I hope this article will attract more citations from other authors.











This page was built for publication: Scalar \(q\)-subresultants and Dickson matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2285223)