A lower bound for higher topological complexity of real projective space (Q1748384)

This article is part of what is now called topological robotics. It derived from \textit{M. Farber}'s work on the problem of motion planning algorithms of a robot in its configuration space \(X\) [Discrete Comput. Geom. 29, No. 2, 211--221 (2003; Zbl 1038.68130)]. The main tool emanating from this work is the topological complexity \(TC (X)\) as well as its higher order analog \(TC_k (X)\) (\(k \geq 2\)) introduced later in [\textit{Y. B. Rudyak}, Topology Appl. 157, No. 5, 916--920 (2010; Zbl 1187.55001)]. This is a homotopy invariant which is one less than the minimal number of planning rules, required for any robot motion between \(k\) consecutive points specified in \(X\). Despite the easy nature of its definition, the determination of \(TC_k(X)\) is most difficult. It is approximated from below by the \(k^{th}\) zero-divisor cup-length \(zcl_k(X; \mathbb{K})\) over a field \(\mathbb{K}\). The author of this article is concerned with the determination of an explicit formula of such an approximation for \(X=\mathbb RP^n\), the real projective space. To this end, he uses the expression of \(zcl_k(\mathbb RP^n;\mathbb{Z}_2)\) established in [\textit{N. Cadavid-Aguilar} et al., Forum Math. 30, No. 2, 397--417 (2018; Zbl 1395.55005)] to prove the recursive formula for \(zcl_k(\mathbb RP^n;\mathbb{Z}_2)\) stated in Theorem 2.1 as follows: Let \(n = 2^e + d\) with \(0\leq d < 2^e\), and \(k \geq 2\). If \(h_k(n) = zcl_k(\mathbb RP^n;\mathbb{Z}_2) - (k - 1)n\); \((m\geq 0)\) and \( h_k(0) = 0\), then \[ h_k(n) = \min(h_k(d) + 2^e, (k - 1)(2^{e+1} - 1 - n)). \] This is in fact a variant of Theorem 1.6 central for the proof of the main Theorem 1.2 of the paper. A sharp result \(TC_k(\mathbb RP^n; \mathbb{Z}_2) = kn \) whenever \(zcl_k(\mathbb RP^n)=kn\), is then obtained as a consequence of Theorem 1.2 and its simply-stated partial formulation is given in Proposition 1.4 as follows: \[ \text{If }n\text{ is even, then }TC_k(\mathbb RP^n)=kn\text{ for }k \geq 2^{l+1} - 1, \] where \(l\) is the length of the longest string of consecutive \(1s\) in the binary expansion of \(n\).