A portfolio is a set of financial assets. The assets can be of different classes, e.g., stocks, currencies and bonds. In this thesis, the considered classes are stock, currency, zero coupon bond and fixed coupon bond. The distribution of the risk factor increments of each asset is modeled as the Generalized Hyperbolic distribution. Both the symmetric and asymmetric version of the Generalized Hyperbolic distribution will be discussed. In measuring portfolio risk, Value-at-Risk and Expected Shortfall are used, instead of the classical risk measure, volatility. Valueat-Risk will only be discussed in the case of symmetric Generalized Hyperbolic distributed risk factor increments since it is not coherent in the asymmetric case. Two versions of portfolio optimization will be discussed. The first one is the classic Markowitz version, while the second one involves the Return-on-Risk-Capital ratio. An important result of this thesis is that the portfolios produced using both symmetric and asymmetric models of Generalized Hyperbolic distribution perform better than JKSE portfolio, with performance measured by RORC. They are better at any type of risk measures used in this thesis. This concludes that the optimization method used in this thesis is more superior than the one used by JKSE.