Cancellation of Morita and skew types (Q2055280)

The \textit{Zariski cancellation problem} asks whether an isomorphism \(A[x] \cong B[x]\) of \(\Bbbk\)-algebras, \(\Bbbk\) a field, implies \(A \cong B\). Though originally studied as a question in algebraic geometry and commutative algebra, there has been significant recent interest in this problems and its variants for noncommutative algebras, building off the work of \textit{J. Bell} and \textit{J. J. Zhang} [Sel. Math., New Ser. 23, No. 3, 1709--1737 (2017; Zbl 1380.16024)]. This paper explores two generalizations of the Zariski cancellation problem well-suited to the study of noncommutative algebras. The first, introduced in [\textit{D. M. Lu} et al., Can. J. Math. 72, No. 3, 708--731 (2020; Zbl 1454.16028)], is to ask whether a Morita equivalence between algebras \(A[x]\) and \(B[x]\) implies a Morita equivalence between \(A\) and \(B\). The authors of the present paper prove that an algebra with center \(\Bbbk\) is universally Morita cancellative. Furthermore, if the center \(Z\) of \(A\) is strongly retractable or \(Z/N(Z)\) is strongly retractable, where \(N(Z)\) is the nilradical of \(Z\), then \(A\) is both strongly cancellation and strongly Morita cancellative. The second generalization replaces polynomial extensions with Ore extensions. The \textit{skew cancellation problem} asks whether an isomorphism \(A[x;\sigma,\delta] \cong B[x';\sigma',\delta']\) implies an isomorphism \(A \cong B\). \textit{J. Bergen} [Commun. Algebra 46, No. 2, 705--707 (2018; Zbl 1410.16029)] considered a version of this problem for differential operator rings. Further progress on this problem appears in \textit{J. Bell} et al. [Beitr. Algebra Geom. 62, No. 2, 295--315 (2021; Zbl 1484.16024)]. The present paper invokes several invariants to study this problem, including the divisor subalgebra and the stratiform length. In particular, suppose \(A\) is a noetherian domain that is stratiform. The authors prove that \(A\) is strongly skew cancellative in the category of noetherian stratiform domains assuming the divisor subalgebra of \(A\) generated by \(1\) is \(A\) itself.