Probabilistic convergence structures (Q750931): Difference between revisions

In a topological space, a net either converges or not: the probability of its doing so is 1 or 0 respectively. This paper deals with ideas of probable convergence: p(\(\alpha\),x) is the probability that a net \(\alpha\) converges to a point x. The author extends the various formulations of topology to model this situation by (a) closure operators \(\theta\) (.,\(\lambda\)) on X, for \(\lambda\) in the unit interval I, (b) neighbourhood filters \({\mathcal N}_{\lambda}(x)\), for \(x\in X\) and \(\lambda\in I,\) (c) a probabilistic convergence function p: net\({}_ X\times X\to I\) (where \(net_ X\) is the set of all nets on X), and (d) diagonal convergence classes \({\mathcal C}_{\lambda}\) on X, for \(\lambda\in I\). (One can characterize the convergence nets in a pre- topology axiomatically, and capture `the fourth neighbourhood axiom' by a `diagonal condition'.) However, the author aims mainly to apply these ideas (1) to the probabilistic convergence structure of Menger space and extend the results of \textit{B. Schweizer} and \textit{A. Sklar} [Pac. J. Math. 10, 313-334 (1960; Zbl 0091.298)], and (2) to do the same for the probabilistic convergence, uniformity and Cauchy sequences in any E-space (the set of measurable functions from a probability space \(\Omega\) to a complete separable metric space Y, together with a canonical Menger space structure [\textit{H. Sherwood}, J. Lond. Math. Soc. 44, 441-448 (1969; Zbl 0167.462)]).