Fuzzy Porfolio Model And Parameter Estimation Of Generalized Linear Model Under Restriction Condition

This thesis is consisted of five parts totally.In the first part, creation and development are introduced firstly, and then the fuzzy number with some related theorems is introduced.In the second part, a few definitions of fuzzy random variables and equivalence of several definitions are introduced; properties of fuzzy random variables is introducedThirdly, the numerical characteristics (expectation and variance) of fuzzy random variables are discussed. In this part, a new way of definition of fuzzy random variance and examples are given.In the forth part, discuss applicaticn of fuzzy random variables. Fuzzy number is applied to portfolio; based on the supposition that the profit rate of the security is fuzzy number, a fuzzy optimization model to portfolio is given in the case that the risky security doesn't exist and short sales are allowed .In the light of fuzzy theory, fuzzy optimization model is changed into the programming model. Finally the steps of using GA to solve this model are given.In the fifth part, when errors in linear model are linear correlative and their covariances are not equal, pararreter estimation of linear model under linear restriction condition is given.