A Scenario Generation Method with Heteroskedasticity and Moment Matching
We present a portfolio management framework composed of a new scenario generation algorithm and a stochastic programming (SP) model. The algorithm is built on heteroskedastic models and a moment matching approach to construct a scenario tree that is a calibrated representation of the randomness in risky asset returns. We also present a multistage SP model that maximizes the expected final wealth and controls the risk exposure through limiting conditional value-at-risk (CVaR) at each decision epoch over the scenario tree generated by the algorithm.
No Reference information available - sign in for access.
No Citation information available - sign in for access.
No Supplementary Data.
No Article Media
Document Type: Research Article
Affiliations: Industrial and Systems Engineering,Rutgers University, PiscatawayNew Jersey
Publication date: July 1, 2011