Infinite games on finite sets (Q995392)

The paper extends existent results in the Model Game denoted by \(MG(X,m,l)\) where \(X=\mathbb{R}\) and \(0<l<m.\) The game consists of \(\omega\) moves played alternately by two players A(Adder) and R(Remover). A finite set \(X_{0}\subset X\) of cardinality \(m\) is presented first to the players. R then chooses a set \( R_{0}\subset X_{0}\) of cardinality \(l\), and A completes \(X_{0}\backslash R_{0}\) by choosing \(l\) points from \(X\) and adding them to \(X_{0}\backslash R_{0}\), thus producing \(X_{1}\subset X\) of cardinality \(m\) again. The two players define a sequence \(\langle X_{i}\mid i\in \omega \rangle \) with \( X_{i}\in [ X] ^{m}\), the collection of all subsets of \(X\) of cardinality \(m.\) In its \(k\)-th move, R chooses \(R_{k}\subset X_{k}\) of size \(l\) and A then defines \(X_{k+1}\in [ X] ^{m}\) so that \( X_{k}\backslash R_{k}\subset X_{k+1}.\) The resulting \(\omega\)-sequence of sets \(P=\langle X_{k},R_{k}\mid k\in \omega \rangle \) is called a play. The set of fixed points of \(P\) is: \(\text{fixed}(P) =\bigcup_{n\in \omega }\bigcap _{k\geq n}X_{k\backslash R_{k.T}}\). In other words, \(x\in \text{fixed}(P) \) iff for some \(n\in \omega ,x\) is in \(X_{n}\) and is never removed in any of the subsequent moves of R. The aim of R is to make \(\text{fixed}(P) \) as small as possible and that of A to make it as large as possible. It is interesting only to study the game under partial information in which one player, R, knows the stage number but does not recall the previous moves, i.e., R knows \(X_{n}\) and \(n\) but not the sequence of \(X_{i}\)'s nor the points removed so far. B. Model was mostly interested in \(X=\mathbb{R}.\) In this paper, the authors investigate the possibilities of \(X=\aleph _{n},\) revealing some interesting questions and putting the results about \(\mathbb{R}\) in context.