The LS category of the product of Lens spaces (Q894964)

Let \(\mathrm {cat } X\) denote the Lusternik-Schnirelmann category of a space \(X\). Let \(L_p^{2k-1}\) be any lens space of the form \(S^{2k-1}/(\mathbb Z/p)\). It is clear that \(k+1\leq \mathrm {cat} (L_p^{2k-1}\times L_q^{2k-1})\leq 2k\). The author proves the following: For all natural \(k\) and prime \(p, q\geq k, p\neq q\), and all lens spaces \(L_p^{2k-1}\) and \(L_k^{2k-1}\) one has \(\mathrm {cat} (L_p^{2k-1}\times L_q^{2k-1})=2k\). To motivate the author's research, note the following. In the paper [Topology 38, No. 1, 37--55 (1999; Zbl 0927.55005)] the reviewer asked if, for every degree 1 map \(f:M \to N\) of closed manifolds, the equality \(\mathrm {cat } M \geq \mathrm {cat } N\) holds. The author calls the affirmative claim \textit{Rudyak's conjecture} and attempts to disprove the above-mentioned inequality by proving the following: If there exist \(k\) and relatively prime \(p\) and \(q\) such that \(\mathrm {cat} (L_p^{2k-1}\times L_q^{2k-1})>2k+1\) then there exists a counterexample to the above-mentioned conjecture. However, as we see, currently we know no \(p,q,k\) with inequality \(\mathrm {cat} (L_p^{2k-1}\times L_q^{2k-1})>2k+1\), and so the conjecture survives. Also, the author proves the inequality \(\mathrm {cat } (M\# N) \leq\max\{\mathrm {cat }M, \mathrm {cat } N\}\) for all closed connected manifolds \(M,N\). Previously, \textit{R. Newton} [On Lusternik-Schnirelmann category of connected sums. Gainesville, FL: University of Florida (PhD Thesis) (2013)] proved this result in case \(\mathrm {cat } M, \mathrm {cat } N\geq 2\).