Higher-dimensional converse for mixed Buchsbaum–Rim multiplicities

Determine whether, for a quasi-unmixed Noetherian local ring of dimension d≥2 with infinite residue field and finite-colength submodules E₁,E₂ of a free module F, equality of all mixed Buchsbaum–Rim multiplicities, e_BR(E₁^{[q−i]},E₂^{[i]})=c for every 0≤i≤q where q=d+rank(F)−1, implies equality of integral closures \overline{E₁}=\overline{E₂} in F.

Background

The paper investigates whether the numerical data consisting of all mixed Buchsbaum–Rim multiplicities determines the integral closures of two finite-colength submodules. It proves that equality of integral closures implies equality of the entire mixed-multiplicity vector, but gives a one-dimensional counterexample showing that the converse fails when d=1.

A positive converse is established when the integral closures of the modules are themselves the integral closures of modules of the form IF and JF for \mathfrak m-primary ideals I and J. The unrestricted case in dimensions at least two remains unresolved: the authors explicitly note that the one-dimensional counterexample does not settle the general question, and then formulate the higher-dimensional problem.

References

At present, the example over $k[[x,y]]/(xy)$ only disproves the implication \textup{(i)}$\Rightarrow$\textup{(ii)} when $d=1$. The example above does not settle Question~\ref{converse}, which asks whether the implication holds or fails in dimensions at least two.

\begin{question} Assume in addition that $R$ is quasi-unmixed, $d\geq2$, and the residue field is infinite. Let $E_1,E_2\subseteq F$ have finite colength. If

e_{\rm BR}(E_1{[q-i]},E_2{[i]})=c \qquad\text{for every }0\leq i\leq q, must one have

\overline{E_1}=\overline{E_2}?

\end{question}

— Counterexamples to Minkowski-type inequalities and the Rees converse for modules via mixed Buchsbaum-Rim Multiplicities  (2610.13888 - Perez, 9 Oct 2026) in Section 6, subsection “The converse problem in dimension at least two,” final Question