Isotopic realization of continuous mappings (Q1404768)

Let \(X\) be a compact space and \(Q\) be a manifold without boundary. A continuous map \(f : X \to Q\) is said to be discretely realizable (D-realizable) if \(f\) is \(\varepsilon\)-approximated by an embedding, \(\forall \varepsilon >0\); moreover, \(f\) is said to be isotopically realizable (I-realizable) if there exists a pseudoisotopy \(H_t: Q \to Q,\) \(t \in I=[0,1]\), taking a certain embedding \(g\) to \(f\) [see \textit{E. V. Shchepin} and \textit{M. A. Shtan'ko}, Topology conference, Proc., Collect. Rep., Leningrad 1982, 135--142 (1983; Zbl 0592.54019)]. The isotopic realizability obviously implies the discrete one. The converse [which is generally false: see \textit{S. A. Melikhov}, Topology Appl. 120, No.1--2, 105--156 (2002; Zbl 1013.57016)] is the object of the so called isotopic realization problem (IR-problem), introduced by Shchepin in 1993. The present paper takes into account the following related problem: how can we know whether a given D-realizable mapping \(f: X \to Q\) is I-realizable? In particular, an obstruction to isotopic realization of a discretely realizable continuous mapping \(f\) of an \(n\)-polyhedron to an orientable PL \(m\)-manifold is constructed, and the completeness of this obstruction is established for \(m > \frac{3(n+1)}{2},\) in case of \(f\) being discretely realizable by skeleta. Then, for \(n>3\), a series of mappings \(\mathbb S^n \to \mathbb R^{2n}\) (with singular set consisting of a \(p\)-adic solenoid, \(p\geq3,\) and a point) is presented, for which the problem is solved in the negative. Furthermore, the authors prove that the problem is solved in the affirmative in the metastable range if stabilization with codimension one is allowed, as well as in the case of a mapping \(f: \mathbb S^n \to \mathbb R^m,\) under the condition that \(f\) is discretely realizable by skeleta and the configuration singular set \(\Sigma(f)=\{(x,y)\mid f(x)=f(y)\}\) is acyclic in dimension \(2n-m\) (according to Stenrod-Sitnikov homology).