A characterization of convergence in \(L^ 1\). Application to quasimartingales (Q4028965)

[For the entire collection see Zbl 0754.00008.] The note contains characterizations of some types of convergence in the probability theory. They are as follows: 1. A characterization of the convergence in law. Let \(D\) be a dense set in \(\mathbb{R}\), and let \(X_ n\), \(X\) be integrable random variables. The following conditions are equivalent: (i) The sequence \(\{X_ n,\;n\geq 1\}\) of random variables converges in law to \(X\) and \(\{X_ n^ +,\;n\geq 1\}\) is uniformly integrable. (ii) \(\lim_{n\to\infty}E[(X_ n-t)^ +]=E[(X-t)^ +]\) for every \(t\in D\). 2. A characterization of the convergence in probability. Let \(X_ n\), \(X\) be random variables on a probability space \((\Omega,{\mathcal A},P)\). The following conditions are equivalent: (i) The sequence \(\{X_ n,\;n\geq 1\}\) of random variables converges in probability to \(X\). (ii) The sequence \(\{X_ n,\;n\geq 1\}\) of random variables converges in law for every probability measure \(Q\) of the form \(Q=f(X)\cdot P\) with \(f\in{\mathcal C}_ c(\mathbb{R},\mathbb{R})\). (iii) The sequence of random vectors \((X,X_ n)\) converges in law to the random vector \((X,X)\). 3. The characterization of the convergence in \(L^ 1\). Suppose that the integrable random variables \(X_ n\), \(X\) satisfy the condition \[ \lim_{n\to\infty}E[f(X)X_ n]=E[f(X)X],\quad f\in{\mathcal C}_ c(\mathbb{R},\mathbb{R}). \] Then \(X_ n\to X\), \(n\to\infty\), in \(L^ 1\) (strong) if and only if one of the following conditions is satisfied: (i) \(\{X_ n,\;n\geq 1\}\) is uniformly integrable and converges in law to \(X\). (ii) \(\lim_{n\to\infty}EX_ n=EX\) and there exists a real, strictly convex function \(g\) on \(\mathbb{R}\) such that \(\lim_{n\to\infty}E[g(X_ n)]=Eg(X)<\infty\). (iii) There exists a dense set \(D\) in \(\mathbb{R}\) such that \(\lim_{n\to\infty}E[| X_ n-t|]=E[| X-t|]\), \(t\in D\). Those results are used in proving \(L^ 1\)-convergence of uniformly integrable martingale.