Lusternik-Schnirelmann \(\pi_{1}\)-category of non-simply connected simple Lie groups (Q881461)

The Lusternik-Schnirelmann (L-S) category is a well-known, and much studied, numerical invariant of a space. There are several equivalent definitions of this invariant: the original is the so-called covering definition. For a space \(X\), the L-S category of \(X\), denoted \(\mathrm{cat}(X)\), is the least integer \(k\) such that there is a covering of \(X\) by \((k+1)\) (not-necessarily connected) open sets, each of which is contractible in \(X\). For instance, we have \(\mathrm{cat}(S^n) = 1\) for \(n \geq 1\). See [\textit{O. Cornea, G. Lupton, J. Oprea}, and \textit{D. Tanré}, Lusternik-Schnirelmann category. Providence, RI: American Mathematical Society (AMS) (2003; Zbl 1032.55001)] for a recent survey. L-S category is notoriously difficult to compute. For instance, the L-S category of many Lie groups and other familiar spaces remains unknown, despite much work on the subject. In view of this difficulty, many approximating invariants have been developed, some of which are more amenable to analysis and computation. The topic announced in the title of this paper, namely the L-S \(\pi_1\)-category is one such. For a space \(X\), the L-S \({\pi_1}\)-category, denoted \(\mathrm{cat}_{\pi_1}(X)\), is the least integer \(k\) such that there is a covering of \(X\) by \((k+1)\) (not-necessarily connected) open sets, in each of which every loop is contractible in \(X\). Evidently, we have the inequality \(\mathrm{cat}_{\pi_1}(X) \leq \mathrm{cat}(X)\). The implicit target of the present paper is the L-S category of Lie groups. In fact, the authors study a second approximating invariant, the {\({\pi_1}\)-cohomological dimension}, which is a lower bound for the L-S \({\pi_1}\)-category. It is direct from the definition that the \({\pi_1}\)-cohomological dimension of a space \(X\), denoted by \(d_{\pi_1}(X)\), is zero if \(X\) is simply connected. As an invariant of a cohomological nature, \(d_{\pi_1}(X)\) is amenable to analysis using Eilenberg-MacLane spaces, Bockstein sequences, and other pieces of machinery from homological algebra. The authors use such tools to compute \(d_{\pi_1}(X)\) for \(X\) any (non-simply connected) compact simple Lie group, except for almost all of the projective orthogonal groups. As a result of their computations of \({\pi_1}\)-cohomological dimension, the authors are also able to determine \(\mathrm{cat}_{\pi_1}(X)\) for various (non-simply connected) compact, simple Lie groups \(X\), including some selective infinite families. It should be emphasized that the computation of L-S category for (non-simply connected) compact, simple Lie groups is quite incomplete. All known results are summarized in [\textit{N. Iwase, M. Mimura}, and \textit{T. Nishimoto}, Topology Appl. 150, No. 1--3, 111--123 (2005; Zbl 1069.55001)]. Much of what is known is concentrated in low-end cases of infinite families. A recent, sporadic additional computation was given in [\textit{N. Iwase} and \textit{A. Kono}, Trans. Am. Math. Soc. 359, 1517--1526 (2007; Zbl 1112.55004)].