Matrices in \(\mathbb{M}_2 [\mathbb{F}_q [T]]\) with quadratic minimal polynomial (Q6150565)

The field of polynomial equations over a ring with matrix solutions or matrices that satisfy certain polynomials over a ring is a fascinating area of study within linear algebra, dating back to the time of Cayley and Hamilton. In 1933, \textit{C. G. Latimer} and \textit{C. C. MacDuffee} [Ann. Math. (2) 34, 313--316 (1933; Zbl 0006.29002)] established a one-to-one correspondence between matrix solutions and ideals. In this paper the authors developed an algorithm for constructing representatives of matrix solutions to \(p(X) = 0\) when \(R = F_{q}[T]\). Let \(p(X) \in R[X]\) be a monic, irreducible polynomial of degree \(n\) over a principal ideal domain \(R\). If \(C \in M_{n}[R]\) is a matrix solution to the equation \(p(X) = 0\), then the solution set of this equation is exactly \[\{S^{-1}CS : S \in \mathrm{GL}_{n}[R]\} \cap M_{n}[R],\] where \(R^{\prime}\) is the field of fractions of \(R\). We call two solutions \(A\) and \(B\) \emph{equivalent} if \(B = S^{-1}AS\) for some \(S \in \mathrm{GL}_{n}[R]\). Two ideals \(\mathfrak{a}\) and \(\mathfrak{b}\) of \(R[\beta]\) are \emph{equivalent} if there exist \(a, b \in R\) such that \(a\mathfrak{a} = b\mathfrak{b}\), where \(\beta\) is a root of \(p(X)\) in the algebraic closure of \(R^{\prime}\). Let \(p(X) = X^{2} -\Gamma X -\Delta \in F_{q}[T][X]\) be irreducible. Let \(\mathrm{deg}(\Gamma) = g\) and \(\mathrm{deg}(\Delta) = d\) and consider the \(2\times 2\) matrices over \(F_{q}[T]\) which satisfy the equation \[ X^{2} - \Gamma X - \Delta =0. \tag{1}\] A matrix solution \[ A= \left( \begin{array}{cc} b & -c \\ a & \Gamma -b \end{array} \right) \] to (1) is said to be \emph{reduced} if \(\mathrm{deg}(b) < \mathrm{deg}(a) < \max \{ \frac{1}{2}d, g\}\), and is said to be \emph{almost reduced} if \(\mathrm{deg}(b) < \mathrm{deg}(a) = \max \{ \frac{1}{2}d, g \}\). Define the mapping \(\phi\) to map the reduced matrix \[ A= \left( \begin{array}{cc} b & -c \\ a & \Gamma -b \end{array} \right) \] to the matrix \(S^{-1}AS\), where \[S =\left( \begin{array}{cc} x & -1 \\ 1 & 0 \end{array} \right)\] and \(x\) is the unique nonzero polynomial \(ax^{2} + (\Gamma - 2b)x + c\) has degree less than \(g\) for exactly two distinct values of \(x\). The main results of the article are as follows: (a) If \(\mathrm{deg}(\Delta)\) is odd and \(\mathrm{deg}(\Delta) > 2 \ \mathrm{deg}(\Gamma)\), then every matrix solution to (1) is equivalent to a unique reduced matrix; (b) Two reduced matrix solutions to (1) are equivalent if and only if \(B = \phi^{k}(A)\)for some integer \(k\). The above results can be extended to higher order matrices and over other types of rings.