Jim Primbs, Stanford University, Management Science and Engineering 
SLIDES

Receding Horizon Control Methods in Financial Engineering

Abstract: Receding horizon control (also known as Model Predictive Control) is a methodology in which on-line optimization problems are repeatedly solved in order to construct feedback control laws. Its development was primarily motivated by problems with constraints. Furthermore, advances in optimization methods and computing speed have greatly increased its domain of applicability. In this talk, we present newly developed semi-definite programming based receding horizon approaches for a class of stochastic systems. We then show how these methods can be used to effectively address a number of difficult financial engineering problems such as dynamic hedging of high dimensional options and index tracking under constraints.