Minimum number of non-zero-entries in a stable matrix exhibiting Turing instability (Q2676231)

Let \(M_n\) be the set of real \(n \times n\) matrices. A matrix \(A \in M_n\) is said to be stable if each of its eigenvalues has a negative real part. A stable \(n \times n\) matrix \(A\) is said to exhibit Turing instability if there is a nonnegative \(n \times n\) diagonal matrix \(P\) such that \(A - tP\) is not stable for some positive \(t\). A matrix \(A \in M_n\) is called irreducible if there does not exist a permutation matrix \(Q\) such that \[QAQ^T = \begin{bmatrix} A_{11} & 0 \\ A_{12} & A_{22} \end{bmatrix},\] where \(A_{11} \in M_k\) and \(A_{22} \in M_{n-k}\) with \(1 < k < n\). The main result established in this paper is the following. Let \(S_n\) be the minimum number of nonzero entries that an \(n \times n\) irreducible matrix \(A\) must have in order for it to exhibit Turing instability. Then \(S_n \leq 2n + 1 - \lfloor \frac{n}{3} \rfloor\) if \(n \geq 3\). In particular, the equality holds when \(n = 3\) and \(S_3 = 6\). Furthermore, the authors determine all possible \(3 \times 3\) irreducible sign pattern matrices with \(6\) nonzero entries which can be realized by a matrix \(A\) that exhibits Turing instability.