Approximation of Levy processes

Jan Rosinski

University of Tennessee, Knoxville

Refreshments: 3:30 - 4:00 p.m. Friday, September 24, at 327 Yost
Talk: 4:00 - 5:00 p.m. Friday, September 24, at 327 Yost.

Levy processes are building blocks of various stochastic models, including certain models with heavy tailed distributions. Computer simulation as well as theoretical analysis of such models presents a problem when the L\'evy processes has infinitely many jumps. The usual approximation by discrete random walks does not converge in the uniform metric because one is always missing large jumps under deterministic time steps.

In this talk we will discuss the method of the uniform approximation of Levy processes by means of series expansions. Specifically, we will discuss series expansions without compensation which are based on Poissonian truncation of i.i.d. sequences. By an appropriate choice of the truncation function (and possibly an optional randomization) one obtains series expansions without compensation for many L\'evy processes, including all processes with infinite variation of positive and negative jumps.

We will also mention a different idea of approximation, and its limitations, of the small jump part of L\'evy processes by Brownian motion.

Questions? Nidhan Choudhuri