The Study To The Measurement Of Risk On Securities&the Model

Researchers have made a great deal of researches on measuring methods for financial risk since applied variance and covariance to measure the risk of securities by Harry M Markowitz. a number of measuring methods for financial risk have been raised half century ago including semi-variance, absolute deviation, semi-absolute deviation, VaR and CVaR, designing effective optimization algorithms to solve the model with different risk index. But the measure method of risk have some lacunae. So research scholars domestic and overseas have done immense amounts of concrete study on it as to offset the lacunae. This paper research launches under this circumstance. The main results of this study are as follows:In order to better control the risk of securities, this thesis focuses on the absolute deviation as the measure of risk which added with variable factor and adjustable factor. Then, this dissertation applied a new genetic algorithm, which include cooling parameters in simulated annealing algorithm, to solve the improved portfolio investment model.It defines the conduction value at risk, analysis the meaning of it under Quantitative Economics. Then, a improved typical transaction cost function are added based on the original MCVaR model. An approach for this optimization model based on modified relative particle swarm optimization using dynamically changing inertia weight is proposed. Then a numerical example show the superiority of the modified model.