On the Kronecker problem and related problems of linear algebra (Q2370779): Difference between revisions

Let \(K\) be a division ring with an automorphism \(\sigma\). In concrete form the semilinear Kronecker problem is as follows: classify the indecomposable pairs \((A,B)\) of rectangular matrices of the same size over \(K\) with respect to transformations of the form \((A,B)\mapsto(X^{-1}AY,X^{-1}BY^{\sigma})\) where \(X,Y\) are nonsingular square matrices. In 1890 Kronecker solved the problem in the case that \(K=\mathbb{C}\) and \(\sigma=1\) by finding a set of canonical forms. His classification was extended by \textit{D. Ž. Djoković} [Linear Algebra Appl. 20, 147--165 (1978; Zbl 0384.15007)] to the general case, and a more general problem (the pseudolinear bundle case) was considered by \textit{V. V. Sergejchuk} [Differ. Uravn. 25, 1721--1727 (1989; Zbl 0701.34018)]. The contragredient semilinear Kronecker problem is: classify the indecomposable pairs \((A,B)\) of rectangular matrices of the same size over \(K\) with respect to transformations of the form \((A,B)\mapsto(X^{-1}AY,Y^{-1}BX^{\sigma})\) for nonsingular square matrices \(X,Y\). A solution to the latter problem in the case \(\sigma=1\) was given by \textit{N. M. Dobrovol'skaja} and \textit{V. A. Ponomarev} [Usp. Mat. Nauk 20, 81--86 (1965; Zbl 0161.02801)]. In the first part of the present paper the author gives short proofs of these two classification theorems (including the pseudolinear bundle case) in a unified form. In the second part of the paper he considers the biquadratic matrix problem. Let \(G\) be a field of degree \(2\) over a field \(F\). The (homogeneous) biquadratic matrix problem is to classify the indecomposable rectangular matrices \(M\) over \(G\otimes_{F}G\) with respect to elementary row and column operations over \(G\) (operating on the left and right, respectively). The second part of the paper shows how this problem (in the case \(char(K)\neq2\)) reduces to the Kronecker problem and provides two sets of canonical forms for the biquadratic matrix problem. The final section shows how some of the author's earlier results on representations of equipped posets [see, for example, \textit{A. Zavadskij}, Linear Algebra Appl. 365, 389--465 (2003; Zbl 1067.16024)] can be generalized using the same ideas.