Constituting optimal portfolio with stochastic programming

In this study, the portfolio optimization problems that had a enormus acceleration in the 1950?s was discussed. At first, it was deliberated that how to forecast the uncertainty and VAR method that was effectively applied in the literature for anticipating the future value of financial investment instruments was used in this application. The continuous random variables obtained from VAR method must be convert into discrete random variables to use in stochastic programing model. Therefore, a new approach was executed by improving the moment-matching method of Høyland and Wallace(2001) and it was obtained more effective and genuine result from this new approach in the acting of discretization. On the other hand, the investor?s attitude toward to risk was involed in the model by using Downside-Quadratic Utililty Function that was executed as the most powerful method by Dempster et al.(2003) with their application. The scenario tree was constructed by using all of the data obtained from new approach and the model was solved with stochastic dynamic/recourse programming that provide wide flexibility. Three-period dynamic model that was constituted supplies versatile and realistic investment process. Not only investor can incorporate his attitude toward to risk in the model, but also he can constitute flexible structure for changing state in investment process.