A New Uniform Convergence for Partial Functions

Suppose X, Y are topological spaces. In this paper maps are not necessarily continuous. A map f from a non-empty subset of X to Y is called a partial map. Partial maps occur as inverse functions in elementary analysis, as solution of ordinary differential equations, as utility functions in mathematical economics, etc. In many applications, X and Y are metric spaces and there is a need to have a uniform convergence on a family of partial functions. Since partial maps do not have a common domain, the usual uniform convergence (u.c.) is not available. Noting that in many situations, all maps of a family under consideration, have a common range, we define a new uniform convergence (n.u.c.) that is complementary to the usual one. This n.u.c. does not preserve continuity but preserves (uniform) openness. Its usefulness stems from the fact that it can be used when u.c. cannot be defined. Moreover, in some situations where both u.c. and n.u.c. are available, the latter satisfies our intuition but not the former. We give applications to ODE's and throw some light on earlier literature.