The lattice of functional Alexandroff topologies (Q2663163)

Let \(X\) be a set and \(f:X \longrightarrow X\) be a function; then \(\mathcal{P}(f):=\{O \subseteq X: f^{-1}(O)\subseteq O\}\) is an Alexandroff topology on \(X\). A space \((X, \mathcal{T})\) is said to be primal [\textit{O. Echi}, Topology Appl. 159, No. 9, 2357--2366 (2012; Zbl 1245.54033)], or functional Alexandroff [\textit{F. A. Z. Shirazi} and \textit{N. Golestani}, Hacet. J. Math. Stat. 40, No. 4, 515--522 (2011; Zbl 1237.54014)] if \(\mathcal{T}= \mathcal{P}(f)\), for some function \(f\). In general, an Alexanroff topology \(\mathcal{T}\) on a set \(X\) can be constructed using a quasi-order \(R\) on \(X\) defined by \((x,y)\in R\) if and only if \(y\in \overline{\{x\}}\), and conversely any quasi-order defines an Alexandroff topology. If instead of a binary relation \(R\) we use a function. We obtain an Alexandroff topology defined by a function. More precisely, if \(\leq\) is a quasi-order on \(X\), then the collection \(\mathcal{B}:=\{(\downarrow x)\mid xín X\}\) is a basis of an Alexandroff topology on \(X\), where \((\downarrow x)=\{y\in X: y\leq x\}\). Conversely, any Alexandroff topology on \(X\) arises in this way. Now, if we consider a function \(f: X\longrightarrow X\), the associated quasi-order is the transitive closure \(\leq_f\) of the binary relation \(f=\{(x,f(x)): x\in X\}\); that is to say \(x\leq_f y\) if and only if there exists a nonnegative integer \(n\) such that \(y=f^n(x)\). Alexandroff topologies defined by functions have been studied in 1997 by \textit{K. Belaïd} et al. [Lect. Notes Pure Appl. Math. 185, 79--88 (1997; Zbl 0877.06008), Example 1, Proposition 3], and later in details in the paper [\textit{K. Belaïd} et al., ``Quasi-spectral binary relations and ordered disjoint unions'', J. Math. Sci. (Calcutta) 11, No. 2, 139--157 (2000)]. It is worth noting that Theorem 3 of the paper under review, in which the authors show that ``the collection \(\mathrm{FA}(X)\) of all functionally Alexandroff topologies on a finite set \(X\), ordered by inclusion, is a lattice'' is incorrect. The supremum in \(TOP (X)\) of two primal topologies need not be primal. However, this does not impact the other results given in the paper.