Topological complexity and efficiency of motion planning algorithms

From MaRDI portal



Abstract: We introduce a variant of Farber's topological complexity, defined for smooth compact orientable Riemannian manifolds, which takes into account only motion planners with the lowest possible "average length" of the output paths. We prove that it never differs from topological complexity by more than 1, thus showing that the latter invariant addresses the problem of the existence of motion planners which are "efficient".


In this paper, the authors introduce a variant of Farber's topological complexity $TC(X)$ when $X$ is a smooth compact Riemannian manifold. To this end, they first notice that, for such spaces, $TC(X)$ is obtained by requiring that domains of continuity of any motion planner $\sigma :X\times X\rightarrow X^I$ (a section of the path fibration $\pi : X^I\rightarrow X\times X$) are mutually disjoint locally compact subsets recovering $X\times X$. This enables them to define the length of $\sigma $ by putting: \[ l(\sigma):=\int_{X\times X}l\circ \sigma. \] Here, the length $l(\alpha)$ is understood in the metric sense for paths $\alpha \in X^I$ which are merely continuous [\textit{P. Petersen}, Riemannian geometry. 2nd ed. New York, NY: Springer (2006; Zbl 1220.53002)]. A motion planner is then called \textit{efficient} if $l(\sigma)=\int_{X\times X}d$. The introduced variant, denoted $lTC(X)$ and called \textit{efficient topological complexity} is, by definition, the minimal integer $m\geq 0$ such that there exists an efficient motion planner $\sigma $ having exactly $m+1$ domains of continuity. \par The main result of the paper is stated as follows: \par Theorem 1: If $X$ is a smooth closed Riemannian manifold, then \[ TC(X)\leq lTC(X)\leq TC(X)+1. \] At the end, an example is presented showing that when $X$ has a non-empty boundary, Theorem 1 does not apply.











This page was built for publication: Topological complexity and efficiency of motion planning algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1725561)