Monte Carlo Methods in Stochastic Programming

Many stochastic programming problems involve complicated performance measures, such as probability, quantile and other risk measures, of random functions that generally have no closed-form expressions. Monte Carlo methods are often used to evaluate these performance measures and to optimize the stochastic programs.

My work in this area focuses on non-convex stochastic programs, such as chance-constrained programs, and developed sample-based (or data-based) convex approximation algorithms to solve them. I have also worked on robust simulation problems that may be formulated into stochastic programs.