Linking and coincidence invariants
Abstract
Given a link map f into a manifold of the form Q = N \times \Bbb R, when can it be deformed to an unlinked position (in some sense, e.g. where its components map to disjoint \Bbb R-levels) ? Using the language of normal bordism theory as well as the path space approach of Hatcher and Quinn we define obstructions \widetilde\omega_\epsilon (f), \epsilon = + or \epsilon = -, which often answer this question completely and which, in addition, turn out to distinguish a great number of different link homotopy classes. In certain cases they even allow a complete link homotopy classification. Our development parallels recent advances in Nielsen coincidence theory and leads also to the notion of Nielsen numbers of link maps. In the special case when N is a product of spheres sample calculations are carried out. They involve the homotopy theory of spheres and, in particular, James--Hopf--invariants.
Cite
@article{arxiv.math/0408046,
title = {Linking and coincidence invariants},
author = {Ulrich Koschorke},
journal= {arXiv preprint arXiv:math/0408046},
year = {2007}
}
Comments
16 pages