Limiting entropy of determinantal processes (Q2212605)

This paper studies limiting entropy of determinantal processes. Given an orthogonal projection \(P\), with rows and columns indexed with a finite set \(V\), there exists a unique determinantal probability measure \(\eta_P\) over subsets of \(V\). Denote by \(B^P\) a random subset of \(V\) with distribution \(\eta_P\). Let \(H(B^P)\) be the Shannon entropy of \(B^P\) defined by \(H(B^P)=\sum_{A\subset V}-\mathbb{P}(B^P=A)\log \mathbb{P}(B^P=A)\). If the sequence of \((G_n,P_n)\) is Benjamini-Schramm convergent and tight, where \(G_n\) is a graph on \(V_n\) if a sequence of orthogonal projection matrices \(\{P_n\}_{n\ge1}\) is considered. It is shown that the limit \(\lim_{n\rightarrow\infty}H(B^{P_n})/|V_n|\) exists.