Balanced distribution-energy inequalities and related entropy bounds (Q654875)

In the paper under review, the author considers positive trace class operators \(\rho \) and selfadjoint operators \(A\) on \(\text{L}^2(X;V)\), for some measured space \(X\) and some separable Hilbert space \(V\). In this general framework, the author etablishes a lower bound on some \(A\)-dependent Dirichlet-like energy of \(\rho\) in terms of \(\rho\) and the spectral resolution of \(A\). This abstract inequality implies (and sometimes improves) many known inequalities: the Berezin-Li-Yau inequality, the Lieb-Thirring inequality, the log-Sobolev inequalities. A similar situation appeared in a previous work [``Spectral density and Sobolev inequalities for pure and mixed states'', Geom. Funct. Anal. 20, No. 3, 817--844 (2010; Zbl 1218.58021)] of the author. Compared to this paper, the main novelty of the present one is the fact that the new inequality is ``balanced''. This means that the inequality enjoys some invariance. This new feature allows the author to remove the nonnegativity assumption on \(A\) used in [loc. cit.]. The introduction gives a list of inequalities implied by the main result. As an illustrative example, the case of the Laplacian (as \(A\)) in \(X=\mathbb{R}^n\) is treated in many places in the paper. Reviewer's remarks. Alas, the text contains several ambiguities. For example, in the main theorem (Theorem 1.2), it would come in handy to have an explicit definition of ``the integral has a finite positive part'' (explaining at the same time the meaning of \(\mathcal {E}_A(\rho )\) which is a priori ill-defined since \(A\rho\) is not trace class). Still, in Theorem 1.2, the author comments on ``the integral involved below'' (appearing in (6)), but there are two integrals in (6). Of course, after some effort, the reader can reconstruct in all such cases the correct meaning. Nonetheless, the referee could have pointed out many of these ambiguities and encouraged the author to remove them. In any case, this paper contains a deep result that is able to explain many interesting inequalities.