Font Size: a A A

The Research Of The Optimal Models For Portfolio Investment Bases On Genetic Algorithm

Posted on:2005-11-07Degree:MasterType:Thesis
Country:ChinaCandidate:K Y ChenFull Text:PDF
GTID:2156360122485419Subject:Systems analysis and integration
Abstract/Summary:PDF Full Text Request
Genetic Algorithm is advanced to calculate the model for portfolio investment in this paper, and the optimal portfolio meeting the specific investor is found. The model is improved during the research, and the Genetic Algorithm is also improved to fit the model in this paper.Because the selection of the stocks is a complex systems engineering, the plenty analysis must be done on the history data before selecting. In order to study whether the time is fit and whether the stock is fit, the stock that is primarily selected is forecast at first in this paper. After comparing, RBF model is adopted to forecast the price of the stock, and the stocks are gained which fit to portfolio.After gaining the stocks what we want, the expected return and variance are calculated, and the stock which we want to fit together are adjusted further more. Then a model for portfolio investment fitting the actual market more is structured, and the concrete portfolio are calculated.Because it is difficulty to settle the problem in this paper by tradition mathematics way, the Genetic Algorithm is used to solve it. And the improvement aiming at the bug of the Genetic Algorithm is done, which is the Adaptive Genetic Algorithm.Two methods are brought forward to obtain the optimal solution after gained the efficient solution: one is letting the Non-discrimination of specific investor and the efficient borderline have a common tangent at a point, and this point is the optimal solution; the other gains optimal solution basing on safety-first method.The optimal solutions we gain with the two methods are only the return rate and the risk rate of the optimal solution, which meet the specific investor; in addition, the concrete portfolio project must be settled, In this paper, model is amended to suit the actual situation at first, and the Constrained Optimization Genetic Algorithm is used to solve the problem.It is the optimal portfolio investment that meets the investor, and this portfolio realizes some balance between the maximal of the expected return and the minimal of the incertitude of the return variance (risk). And of course this balance point is a best point which meet the investormost.
Keywords/Search Tags:expectation return, risk, RBF nerve net, forecast, efficient borderline, Non-discrimination, Genetic Algorithm, Adaptive Genetic Algorithm, Constrained Optimization Genetic Algorithm, optimal solution
PDF Full Text Request
Related items