Completeness properties of the generalized compact-open topology on partial functions with closed domains (Q5928486)

Let \(X\) and \(Y\) be Hausdorff topological spaces and CL\((X)\) the family of all nonempty closed subsets of \(X\), and for any \(B\in\text{CL}(X)\) let \(C(B,Y)\) be the set of all continuous functions from \(B\) to \(Y\). A partial map is a pair \((B,f)\) such that \(B\in\text{CL}(X)\) and \(f\in C(B,Y)\), and \({\mathcal P}={\mathcal P}(X,Y)\) denotes the family of all partial maps. The topology \(\tau_c\) on \({\mathcal P}\) which is called generalized compact-open topology is the topology having subbase elements of the form \([U]= \{(B,f) \in {\mathcal P}:B \cap U\neq\emptyset\}\), where \(U\) is open in \(X\), \([K,I]= \{(B,f) \in {\mathcal P}:f (K\cap B)\subset I\}\), where \(K\) is compact in \(X\) and \(I\) is open (possibly empty) in \(Y\). The game \(\text{BM}_0(X)\) is a version of the Banach-Mazur game played on a topological space \((X,\tau)\) by two players \(\alpha\) and \(\beta\), as follows: \(\beta\) starts by picking some \(U_0\in \tau \smallsetminus \{\emptyset\}\) such that \(U_1\subset U_0\). In an even (respectively odd) step \(n\geq 1\), \(\beta\) (respectively \(\alpha)\) choose a \(U_n \in\tau \smallsetminus \{\emptyset\}\) with \(U_n\subset U_{n-1}\). Then \(\alpha\) wins if \(\bigcap_{n<\omega}U_n\) is a singleton for which \(\{U_n: n<\omega\}\) is a basic system of neighborhoods and otherwise \(\beta\) wins. In the case of Banach-Mazur game, \(\alpha\) wins if \(\bigcap_{n< \omega}U_n \neq\emptyset\) for above \(U_n'\), and otherwise \(\beta\) wins. The compact-open game \(\text{KO} (X)\) on \((X,\tau)\) is played as follows: \(\beta\) starts by picking a couple \((K_0, U_0)\in {\mathcal K}(X)\times \tau\) such that \(\overline U_0\) is compact, where \({\mathcal K}(X)\) denotes the set of all (possibly empty) compact subsets of \(X\). Then \(\alpha\) responds by some \(V_0\in \tau\) with compact closure that is disjoint to \(K_0\cup U_0\). In step \(n\geq 1\), \(\beta\) (respectively \(\alpha)\) choose a couple \((K_n, U_n) \in{\mathcal K}(X) \times\tau\) (respectively a set \(V_n\in\tau)\) such that \(\overline U_n\in {\mathcal K}(X)\) (respectively \(\overline V_n\in{\mathcal K}(X))\) and \(U_n\cap \bigcup_{i<n} (V_i\cup U_i\cup K_i)=\emptyset\) (respectively \(V_n\cap (\bigcup_{i<n}V_i\cup\bigcup_{i< n}(U_i\cup K_i)) =\emptyset\). Player \(\alpha\) wins if \(\{U_n:n\in \omega\}\cup\{V_n: n\in\omega\}\) is a locally finite family; otherwise \(\beta\) wins. Another game \(\text{KO}_0(X)\) is a modification of \(\text{KO}(X)\), where in \(\beta\)'s choice \(K_n=\emptyset\) for all \(n\). The authors discuss the relationship between game theoretical conditions and topological properties, especially Baireness, and may be picked up as some of the proved results: Proposition 1.3 (i) If \(X\) is a locally compact paracompact space, then \(\alpha\) has a winning strategy in \(\text{KO}(X)\). (ii) If \(X\) is an almost locally compact, nonlocally compact \(q\)-space, then \(\beta\) has a winning strategy in \(\text{KO}_0(X)\), where \(X\) is said to be almost locally compact if every nonempty open subset of \(X\) contains a compact set with nonempty interior. Proposition 1.4. (i) If \(X\) is a locally compact space, then \(\alpha\) has a winning strategy in \(\text{KO}(X)\) iff \(X\) is paracompact. (ii) If \(X\) is an almost locally compact \(q\)-space, then \(\alpha\) has a winning strategy in \(\text{KO}(X)\) iff \(X\) is paracompact and locally compact. Theorem 4.2. Let \(X\) be an almost locally compact space and assume that \(\alpha\) has a stationary winning strategy in \(\text{BM}_0(Y)\). Then (i) \(({\mathcal P}, \tau_c)\) is a Baire space if \(\beta\) has no winning strategy in \(\text{KO} (X)\). (ii) \(({\mathcal P},\tau_c)\) is weakly \(\alpha\)-favorable (i.e. \(\alpha\) has a winning strategy in \(\text{BM}(X))\), if \(\alpha\) has a winning strategy in \(\text{KO}(X)\). A (stationary) strategy in the game for one of the players is a function, which picks an object for the relevant player knowing all the previous moves of the opponent as well as of his own (respectively knowing only the previous move of the opponent). A (stationary) winning strategy \(\sigma\) for a player is a (stationary) strategy winning for the player every run of the game compatible with \(\sigma\).