3
$\begingroup$

Let $J$ denote the set of functions $f : [0, \infty) \to \mathbb{R}$ that are right-continuous and have left-hand limits (r.c.l.l.) and such that their points of discontinuity are jumps. Then $J$ is a vector space and contains the following subsets: all continuous functions $C$, all (r.c.l.l.) non-decreasing functions $M$ and thus distribution functions $D \subseteq M$ of probability measures on $[0, \infty)$, the linear hull $M - M$ of (r.c.l.l.) functions of locally bounded variation and many others.

For $f_n, f \in D$ define the convergence $f_n \to f :\Leftrightarrow f_n(x) \to f(x)$ for all continuity points $x$ of $f$. Then $\to$ satisfies:

  1. if $f_n \to f$ and $f_n \to g$ then $f = g$ (by right-continuity)
  2. if $f_n = f$ for all $f$ then $f_n \to f$
  3. if $f_n \to f$ then $f_{n_k} \to f$ for all subsequences $f_{n_k}$ of $f_n$

It follows that $\to$ is an $L$-convergence on $J$ and defines a sequential $T_1$ topology $\tau$ on $J$. Clearly, $\tau$ is weaker than the topology of (everywhere) pointwise convergence. Note also that the usual notion of "uniform convergence on a particular family $\mathcal{F}$ of subsets of $[0,\infty)$" (e.g. pointwise convergence, uniform convergence or compact convergence) is based on a specification of $\mathcal{F}$ in advance whereas for the convergence $\to$ we choose such a family depending on properties of its limit.

I have only met this notion of convergence on the subset $D$ of distribution functions in which case it is just the weak topology of probability measures. I haven't seen discussions of properties of the topologies induced by this notion of convergence on larger spaces like $J$, but it seems to "naturally" extend the convergence on $D$.

Questions:

  1. Does $\to$ has a special name or does anyone know for references regarding properties of this convergence / topology (e.g. descriptive properties or compatibility with vector space structure)?
  2. Similarly, one can also consider the (locally) uniform convergence at points of continuity of the limit function. Is there also information available on this convergence?
$\endgroup$

0

You must log in to answer this question.

Browse other questions tagged .