Financial Engineering Seminar
October 27, 2006 (Fri)
Time: 3:00 pm
Place: Weil Hall 307
There has been extensive research on utility maximization in continuous-time portfolio selection problems. However, we can show that simply maximizing the portfolio utility without considering the potential portfolio loss may result in large losses. We discuss the portfolio selection problem with dual objectives of maximizing the portfolio utility and minimizing the portfolio loss CVaR, and with state constraints. The efficient frontier of utility and CVaR is modelled and solved with a two-stage optimization problem: the first stage is a parametric non-differentiable stochastic control problem which can be solved with the dynamic programming principle and the viscosity solution technique, and the second stage is a convex minimization problem which can be solved with the sequential penalty function method.