Browse by: Author name - Classification - Keywords - Nature

2 matches found
V: 12, 127-137, LNM 191 (1971)
The approximate Laplacian'' method of computing the increasing process associated with a supermartingale does not always converge in the strong sense: solves a problem open for many years
After a suitable translation, Dubins' theorem can be stated as follows: if $X$ is a positive submartingale with $X_t\in L^2$ for all $t$, then $X^2-[X,X]$ is a submartingale