Stochastic ordering by \(g\)-expectations (Q2038280): Difference between revisions

Stochastic integral orders are used to compare random variables or stochastic processes and are defined via comparison of integrals of suitable classes of functions with respect to the random variable, respectively the process, e.g., monotone, convex, convex increasing functions. In the article it is argued that linearity of the conventional expectation operator is too restrictive for applications the authors are interested in. Instead, for the more complicated problems in case of ordering diffusion processes an ordering via so-called \(g\)-expectations is used. The main interest is in comparing the terminal states of two processes driven by forward stochastic differential equations by \(g\)-expectations of convex functions. The ``function'' \(g\) is a process adapted to the underlying Brownian motion and is called generator of a certain associated backward stochastic differential equation. Applications from financial mathematics are provided.