The continuous-time optimal portfolio using a multivariate normal inverse Gaussian model

This paper develops the continuous-time portfolio model using a multivariate normal inverse Gaussian model. Though the weighted average of lognormal variables is no longer lognormal, it can be approximated by other distributions, such as a multivariate normal inverse Gaussian model. Our method belongs to the analytic approximation class. By comparing to Monte Carlo experiments, it illustrates the computational efficiency and accuracy of our approach.