Homotopical intersection theory. III: Multirelative intersection problems (Q2311435)

Fix a closed, connected, smooth manifold \(N\). Let \(Q_1, \dots, Q_j\) be pairwise disjoint submanifolds, and \(P\) another manifold. Suppose we are given a map \(f: P \to N\). The authors consider the question whether \(f\) is homotopic to a map from \(P\) to the space \(N\setminus\bigcup_{1 \le i \le j} Q_i\), assuming we are given coherently compatible maps from \(P\) to the spaces \[ N\setminus \bigcup_{\substack{1\leq i \leq j\\ i\ne i'}}Q_i \] where \(i'\in\{1, \dots, j\}\). The problem can be reformulated as follows: suppose we are given a map \[ g: P \to \mathop{\operatorname{holim}}_{S\subsetneq \{1,\dots, j\}} N \setminus \bigcup_{i\in S} Q_i. \] Does \(g\) lift to a map \(P\to N\setminus \bigcup_{1 \leq i \leq j} Q_i\)? The main technical result of the paper is a stable homotopy theoretic model for the total homotopy fiber of the cubical diagram \[ S\mapsto N \setminus \bigcup_{i\in S} Q_i. \] More precisely, the authors describe an infinite loop space depending on \(N\) and the \(Q_i\)'s, that is equipped with a natural map from the total homotopy fiber that induces an isomorphism on homotopy groups in a range depending on the dimensions of \(N\) and the \(Q_i\)'s. Using this model the authors define an obstruction to the lifting problem described above.