Path-connected inverse limits of set-valued functions on intervals (Q776385)

The paper presents a significant contribution to the theory of generalized inverse limits over intervals. The main goal of the paper is to find necessary and sufficient conditions for generalized inverse limits to be path-connected. For observing those questions, the authors introduce the path-component base (analogous to a component base introduced by \textit{S. Greenwood} et al. [Topology Appl. 221, 69--90 (2017; Zbl 1373.54043)]). With that, the authors again show the usefulness and power of this tool. In the theory of generalized inverse limits, a Mahavier product is one of the most frequently studied notions and used tools for detecting properties of generalized inverse limits. In the paper, the authors use this notion in the aspect of path-connectedness. A characterization is proven of path-connected Mahavier products of an inverse sequence of upper semicontinuous surjective functions. We should point out that for studying path-connectedness of Mahavier products, the authors use a (stronger) version of the mountain climbing problem (the so-called strong mountain climbing condition; Definition 4.2). Let us give some of the main theorems from the paper (all notions are given completely in the paper). Theorem 1.1. Suppose that \(\pmb{f}\) is an inverse sequence of upper semicontinuous surjective set-valued functions whose graphs are path-connected, and there exist positive integers \(m < n\), such that \(\pmb{f}\) has a path-connected base over \([m,n]\). Then neither \(\bigstar_{i \in [1,n]} \Gamma(f_i)\) nor \(\varprojlim \pmb{f}\) is path-connected. Theorem 1.2. If \(\pmb{f}\) is an inverse sequence of upper semicontinuous surjective set-valued functions on intervals and whose graphs are path-connected, and \(n >1\), then the following are equivalent: (1) \(\bigstar_{i \in [1,n]} \Gamma(f_i)\) is path-connected; (2) for all \(\pmb{a,b}\in \bigstar_{i \in [1,n]} \Gamma(f_i)\) there exists a \([1,n]\)-consistent sequence of paths \(\left\langle \gamma_i : [0,1] \to \Gamma(f_i) \ | \ i \in [1,n] \right\rangle\) such that for every \(i \in [1,n]\), \(\gamma_i\) is a path from \((a_{i-1},a_i)\) to \((b_{i-1},b_i)\); (3) for all \(\pmb{a,b}\in \bigstar_{i \in [1,n]} \Gamma(f_i)\) there is a sequence of paths \(\{\gamma_i : [0,1] \to \Gamma(f_i) \ | \ i \in [1,n] \}\) where \(\gamma_i\) is a path from \((a_{i-1},a_i)\) to \((b_{i-1},b_i)\) and if \({\mathbb J}_i := \pi_i \gamma_i ([0,1]) = \pi_i \gamma_{i+1}([0,1])\), then for every \(i \in [1,n]\), \(\pi_i \gamma_i, \pi_i \gamma_{i+1} :[0,1] \to {\mathbb J}_i\) satisfy the strong mountain climbing condition; (4) for all \(\pmb{a,b} \in \bigstar_{i \in [1,n]} \Gamma(f_i)\) there is a sequence of paths \(\{\gamma_i : [0,1] \to \Gamma(f_i) \ | \ i \in [1,n] \}\) where \(\gamma_i\) is a path from \((a_{i-1},a_i)\) to \((b_{i-1},b_i)\) and if \({\mathbb J}_i := \pi_i \gamma_i ([0,1]) = \pi_i \gamma_{i+1}([0,1])\), then for every \(i \in [1,n-1]\) the two functions \(\pi_i \gamma_i, \pi_i \gamma_{i+1} :[0,1] \to {\mathbb J}_i\) match.