On the expectation of the maximum for sums of independent random variables (Q1068448)

It is well-known that for a sequence \(X_ 1,X_ 2,..\). of i.i.d. random variables with mean zero \(E \sup_{n} ((S_ n- an)^+)^{\gamma}<\infty\) if and only if \(E(X^+_ 1)^{1+\gamma}<\infty\) for any \(a,\gamma >0\), with \(S_ n=X_ 1+X_ 2+...+X_ n.\) In this note we treat this problem without the assumption of identical distribution. We give an example of a sequence \(X_ 1,X_ 2,..\). with \(EX^ 2_ i=o(1)\) and \(E \sup_{n} (S_ n-n)^+=\infty\) and then derive a result which e.g. shows that for uniformly integrable random variables \(E \sup (S_ n-an)^+<\infty\) if \(E(X^+_ i)^{2+\delta}=O(1)\) for some \(\delta >0\). We then apply this result to a problem of optimal stopping in continuous time.