Maximal inequalities and some applications (Q6158179)

Given an underlying stochastic process \((X_t)_{t\geq0}\), the authors provide a survey of maximal inequalities, that is, inequalities that feature the running maxima \(\sup_{s\leq t}X_s\) or \(\sup_{s\leq t}|X_s|\), including tail bounds, moment bounds and similar quantities. The survey begins with a well-paced introduction to general techniques and results, before devoting each of the remaining sections of the paper to particular choices of \((X_t)_{t\geq0}\). Each of these sections is self-contained, and begins with an introduction to the family of processes under consideration. The authors study maximal inequalities for martingales (including Doob's maximal inequalities, and Khintchine and Burkholder-Davis-Gundy inequalities), for Lévy processes, for process which behave locally like a Lévy process (of which several examples are given, including Feller processes and stable-like processes), for Markov processes, and for Gaussian processes. In all cases the discussion is restricted to continuous time and finite dimensions, for reasons of space. For similar reasons the authors do not cover the theory of large deviations. This allows the authors enough room to develop the topics which are the focus on this survey, providing a useful resource for those interested in the area.