Discrete Morse theory and Hopf bundles (Q627400)

Discrete Morse theory provides a very useful tool for studying the homotopy type of CW-complexes: the consideration of an acyclic matching on the poset of nonempty cells of a regular CW-complex \(X\) leads to the existence of a CW-complex \(Y\) homotopy equivalent to \(X\) and whose number of cells of dimension \(i\) is equal to the number of critical cells (i.e., cells not appearing in the matching) of dimension \(i\). Of course this does not mean that \(X\) has the homotopy type of a wedge of spheres, even in the particular case where the homology of \(Y\) is given by the enumeration of critical cells because there are no critical cells in two consecutive dimensions. In this paper, the author illustrates this last fact; after having described a cellular structure on the complex projective plane \(\mathbb{C}P^2\), he exhibits a matching on the poset of nonempty cells of this CW-structure which shows a homotopy equivalence with a CW-complex with one cell in each of the dimensions 0, 2 and 4; but \(\mathbb{C}P^2\) is not homotopy equivalent to a wedge \(S^2\vee S^4\) (the homotopy type of \(\mathbb{C}P^2=\mathbb{C}P^1\cup_hD^4\) depends on the homotopy type of the Hopf map \(h:\partial D^4=S^3\to S^2=\mathbb{C}P^1\) and this gives the link with Hopf bundles and other examples of the same type). However, by considering certain paths in posets (called generalized alternating paths), the author introduces the notion of feasibility domain of an element of a poset from which he finds conditions on acyclic matchings that ensure the homotopy equivalence with a wedge of spheres enumerated by critical cells. He also shows with an example modeled on \(S^1 \vee S^2\) that these conditions are not necessary.