This thesis develops a methodology to determine an optimum composition of a portfolio of real assets. It involves selecting real assets which will be included in the portfolio and taking into consideration all the constraints which apply. The possibility of exercising real options is taken into consideration. The determination of the optimum composition is done by maximizing a performance measure called Omega. Omega measure is defined as the relation between the expected average gain (Expected Chance) vs. the expected average loss (Expected Shortfall) of the returns or Net Present Values (NPV) distribution. This measure requires the decision maker to define previously the minimum desirable level of return or NPV, depending on the context it is being used, which is the border between the gains and losses areas in the distribution. Omega takes into account of all the moments of the distribution of the future returns or NPV, implying it does not restrict to the Mean-Variance world. It is a well known empirical fact that many financial variables don´t follow normal distributions or most investors don´t have quadratic utility functions, which causes the classical model of portfolio composition proposed by Markowitz (1952) inappropriate. The Omega measure can deal satisfactorily in all cases not having normal distributions or even in cases which have normal distributions. In this thesis, the numerical method of Monte Carlo Simulation is used to determine the NPV distribution and calculate Omega measure.