2012 Jun 18 at 13:00
DC 2314
Shu Tong (Stephen) Tse, PhD candidate, David R. Cheriton School of Comp. Sci., Univ. Waterloo
We formulate optimal trade execution as a stochastic control problem and obtain the solution by numerically solving a nonlinear Hamilton-Jacobi-Bellman (HJB) Partial Differential Equation (PDE). The HJB PDE is expensive to solve due to high dimensionality and flatness of local objective. By carefully analyzing the HJB PDE, we have developed an efficient numerical method by combining the techniques of similarity reduction, non-standard interpolation, and careful grid construction.