Homotopy invariants for tangent vector fields on closed sets (Q2493834)

The article deals with a new topological degree theory for single and convex valued tangent vector fields on locally compact closed subsets \(M\) of a Banach space \(E\). More exactly, it is assumed that \(M \subset E\) is an \({\mathcal L}\)-retract, that is a closed set \(M\) such that there exists an open neighborhood \(\Omega\) of \(M\), a retraction \(r: \Omega \to M\), and a constant \(L\) satisfying the inequality \(\| r(x) - x\| \leq Ld(x,M), \;x \in \Omega\). (In particular, closed convex sets, sets being bi-Lipschitz homeomorphic to closed convex ones, epi-Lipschitz sets, proximate retracts, compact sets having the neighborhood Lipschitzian extension property, strictly regular sets, Lipschitz neighborhood retracts are \({\mathcal L}\)-retracts; \({\mathcal L}\)-retracts are ANRs). Let \(U\) be a nonempty open subset of a locally compact \({\mathcal L}\)-retract \(M\) and let \(\Phi: U \to E\) be an upper semicontinuous map, single-valued or with compact convex values, and satisfying the properties: (A1) the set \(\Phi^{-1}(0) = \{x \in U\mid 0 \in \Phi(x)\}\) is compact; (A2) \(\Phi\) is weakly tangent to \(M\), i.e. \(\Phi(x) \cap T_M(x) \neq \emptyset\) for all \(x \in U\); here \(T_M(x)\) is the Clarke tangent cone defined by \[ T_M(x) = \Biggl\{u \in E\;\biggl|\;\limsup_{y \overset{M}{\to} x, \;{h} 0^+} \frac{d(y + hu,M}{h} = 0\Biggr\}. \] In this case the authors define the degree \(\deg (\Phi,U)\) of \(\Phi\) on the boundary of \(U\) having standard properties of the degree: (a) Existence Property (if \(\deg(\Phi,u) \neq 0\), then \(\Phi^{-1}(0) \cap U \neq \emptyset)\); (b) Additivity Property (if \(U_1, U_2 \subset U\) are open in \(M\) and \(\Phi^{-1}(0) \cap U \subset (U_1 \cup U_2) \setminus (U_1 \cap U_2)\), then \(\deg (\Phi,U) = \deg (\Phi,U_1) + \deg (\Phi,U_2))\); (c) Normalization Property (if \(M\) is compact, then \(\deg (\Phi,M) = \chi(M)\), \(\chi(M)\) is the Euler characteristic of \(M\)); (d) Homotopy Invariance Property (if \(\Phi_0, \Phi_1: U \to E\) are homotopic in an appropriate sense, then \(\deg(\Phi_1,U) = \deg(\Phi_2,U)\)). This degree allows the authors to study vector fields on sets which are neither smooth nor convex. Further, the authors construct the degree \(*\text{-}\deg(\Phi,U)\) for vector fields \(\Phi: U \to E^*\) having properties analogous to (A1) and (A2); instead of equilibria (zeroes) of fields here they consider co-equilibria (a point \(x\) is a co-equilibrium if \(\Phi(x) \cap N_M(x) \neq \emptyset\), \(N_M(x) = \{p \in E^*\mid \langle p,u \rangle \leq 0 \;\forall u \in T_M(x)\}\)). New theories are applied to study continuation and bifurcation of equilibria and co-equilibria of tangent vector fields in different closed sets, including parameterized families of ones. The authors cite more than 40 references, however they omit references onto the basic manuscript by \textit{J. Leray} [Bull. Soc. Math. Fr. 87, 221--233 (1959; Zbl 0093.36702)] and numerous investigations of Voronež mathematicians devoted to the relative degree of mappings leaving invariant a closed convex sets (without the assumption about the local compactness of \(M\)).