On approximate commutativity of spaces of matrices

From MaRDI portal



Abstract: The maximal dimension of commutative subspaces of Mn(mathbbC) is known. So is the structure of such a subspace when the maximal dimension is achieved. We consider extensions of these results and ask the following natural questions: If V is a subspace of Mn(mathbbC) and k is an integer less than n, such that for every pair A and B of members of V, the rank of the commutator AB−BA is at most k, then how large can the dimension of V be? If this maximum is achieved, can we determine the structure of V? We answer the first question. We also propose a conjecture on the second question which implies, in particular, that such a subspace V has to be an algebra, just as in the known case of k=0. We prove the proposed structure of V if it is already assumed to be an algebra.


The authors extend some well-known result on the maximal dimension of commutative subspaces of \(M_n(\mathbb{C})\), the space of \(n \times n\) complex matrices. Let \(V\) be a subspace of \(M_n(\mathbb{C})\), and let \(k < n\) be fixed. The main question is: What can be said about \(V\) if \[ \mathrm{rank}(AB-BA) \leq k, \] for all \(A\) and \(B\) in \(V\)? They show that the maximal possible dimension of \(V\) is: \[ kn + \lfloor\frac{n-k}{2}\rfloor \cdot \lfloor\frac{n-k+1}{2}\rfloor + 1, \] which generalizes the known inequality for the maximal dimension of commutative subspaces of \(M_n(C)\) (where \(k = 0\)): \[ \lfloor\frac{n}{2}\rfloor \cdot \lfloor\frac{n+1}{2}\rfloor + 1\,. \] Furthermore, if this dimension is achieved, then \(V\) is an algebra. The authors present a conjecture on the structure of \(V\), which is here proved for the two limiting cases of \(k=1\) and \(k = n-1\).











This page was built for publication: On approximate commutativity of spaces of matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6178780)