Topological complexity of motion planning in projective product spaces (Q1948307)

The topological complexity \(\mathrm {TC}(X)\) of a space \(X\) is a numerical homotopy invariant introduced by \textit{M. Farber} in [Discrete Comput. Geom. 29 No. 2, 211--221 (2003; Zbl 1038.68130)] in order to study the motion planning problem in Robotics from a topological perspective. Further developments of the subject have shown that topological complexity is not only related to robotics but also to difficult problems in algebraic topology such as the immersion problem for \(P^n,\) the real \(n\)-th projective space. In fact, the immersion problem for \(P^n\) is basically equivalent to the computation of its topological complexity [\textit{M. Farber} et al., Int. Math. Res. Not. 2003, No. 34, 1853--1870 (2003; Zbl 1030.68089)]. In the paper under review the authors study the topological complexity of projective product spaces. Such a family of manifolds (which gives a rather natural generalization of real projective spaces) is defined as follows: if \(\overline{n}\) stands for an \(r\)-tuple \((n_1,\dots,n_r)\) of positive integers with \(n_1\leq \dots\leq n_r\), then the \(\overline{n}\)-projective product space \(P_{\overline{n}}\) is the orbit space resulting from the diagonal action of \(\mathbb{Z}_2\) on the product of spheres \(S_{\overline{n}}=S^{n_1}\times \dots\times S^{n_r}.\) The authors prove that, in many nontrivial examples, the topological complexity of projective product spaces from at least two sphere factors is (much) lower than the dimension of the manifold, in contrast with the case of real projective spaces. The most important estimates in the paper are the following: Proposition 3.7. Let \(\phi (n_1)\) be the number of positive integers less than or equal to \(n_1\) which are congruent to 0, 1, 2, or 4 (mod 8). If \(\nu (n_{i+1})\geq \phi (n_1)\), then \[ \mathrm {zcl}_{\mathbb{Z}_2}(P^{n_1})+l(\overline{n})-1\leq \mathrm {TC}(P_{\overline{n}})\leq \mathrm {TC}(P^{n_1})+l(\overline{n})-1 \] where \(l(\overline{n})\) denotes the number of factors in \(S_{\overline{n}}\), and \(\mathrm {zcl}_{\mathbb{Z}_2}(P^{n_1})\) the zero divisors \(\mathbb{Z}_2\)-cuplength of \(P^{n_1}.\) Further, both inequalities above become equalities precisely for \(n_1\) a 2-power. After analyzing some results in an equivariant setting of topological complexity the following theorem is also proved: Theorem 3.8. If \(k\) denotes the number of spheres \(S^{n_i}\) with \(n_i\) even and \(i>1\), then \[ \mathrm {TC}(P_{\overline{n}})<\mathrm {TC}(P^{n_1}+1)(l(\overline{n})+k)) \] The authors also give a characterization of the Euclidean immersion dimension of projective product spaces by means of a generalized concept of axial map.