Higher order derivatives and perturbation bounds for determinants (Q734922): Difference between revisions

Let \(\mathbb{M}_n\) denote the set of \(n\times n\) complex matrices. The authors define the \(k\)-th derivative of the determinant at a point \(A\in\mathbb{M}_n\) to be the map \[ D^k\det A:\mathbb{M}_n\times\mathbb{M}_n\times\cdots\times\mathbb{M}_n \rightarrow\mathbb{C} \] given by \[ D^k\det A(X_1,X_2,\dots,X_k)= \frac{\partial^k}{\partial t_1\cdots\partial t_k}\Bigg|_{t_1=\dots= t_k=0} \det(A+t_1X_1+\cdots+t_kX_k). \] For the case \(k=1\), a formula of Jacobi says that \(D\det A(X)=\text{tr}(\text{adj}(A)X)\). The purpose of this note is to give several analogues of this formula for the higher derivatives. The authors also evaluate the operator norm of \(D^k\det A\) in terms of the singular values of \(A\).