The Application Of Genetic Algorithm On The Chinese Security Market

For about ten years development of our country stock market have got great success. But we must know about that, there is a huge increase of investment risk in the stock market, for the inadequate of announce mechanism and watch and control mechanism. For the development of Chinese stock market and profit of investors, it is necessary to seek the security portfolios to reduce the risks on base of investment combination theory.There are four parts in this paper .The first part mainly suggest that the operator features and internal conflicts of Chinese stock market should be the final results of various investors behaviors determined by the benefits under a certain market system, which come from our special mechanism. So the market system should be the most important factor that affects the operation, development and improvement of Chinese security market. For the reason of the government intervention of the market made the supply of its own system invalid, so the market deviated from the government intention, have damaged the efficiency and vigor of the market . So the system innovation of Chinese security market is imperative.The model of to evade the risks is the focus of the second part. In this paper, the author utilizes the Markowitz mean—variance model and multi-factor model for portfolio investment decision. The third part mainly explains the essential theory of a modern novel optimization method - Standard Genetic Algorithms (SGA) .The relation betweenGenetic Algorithms (GA) and biological evolution is clarified briefly. Then the detailed design techniques and procedures of GA are discussed. At last, the primary character of GA, the state-of-art and the prospect of GA are reviewed.The forth part based on the character of Portfolio choice model, in this article studies coding, genetic operators and operators' parameter and designs a genetic algorithm which can solve portfolio model. The empirical analysis shows that computational result of this genetic algorithm is better than computational result of gradient algorithm when solving complex Portfolio model. And then as far as the transaction cost is concerned, a multi- factor model for portfolio investment decision is established based on arbitrage pricing theory under the condition of no short sale, and then its solution is studied with the help of genetic algorithm. The method is tested and the expected results are obtained.The last part mainly explains that nowadays genetic algorithms have been widely applied to various areas, there are a search bias and early-maturing problem in genetic algorithm s with global and parallel method. To solve this problem, there are many reform approaches in genetic algorithms. It can win through traditional genetic algorithm's shortcomings by applying accelerating genetic algorithm in solving combination forecasting problems. These shortcomings include poor adapt ability in search space (i.e. optimizing variable space), large measure quantity, premature convergence, no definitude instruct rule for set ting technique of control parameter, etc. The new approach does not need cauomcity in forecasting eiror information matrix, the objective function scale may extend widely, and the forecasting precision is high.