General Seminar
May 7, 2003, Wednesday, 15:40, Institute of Applied Mathematics, Room M205,
Ali Nüvit VEYSOĞLU
Toros Menkul Kıymetler A.Ş İSTANBUL
Portfolio Optimization In Stock Exchanges Modeled With Game Theory
In this study, an extensive-form game model of the stock exchange is introduced. In this model, the players will announce ask and bid prices together with their amounts, for each stock and with a given order. Therefore, the prices of the stocks will not be imposed to the market but will be determined according to both price and quantity strategies of the players, throughout a matching mechanism. It explains how artificial neural networks may be used to estimate the decision mechanisms of the players' stock valuations. By this way, explicit explanations are given to investor behaviors. Algorithms and computer programs for estimating the prices and risks of the stocks according to the results obtained, are introduced. Using two different risk definitions, a portfolio optimization model, designed for an analyst who does not have subjective stock valuations, where the majority of the players act according to their expectations, is introduced.