Application Of Hybrid Genetic Simulated Annealing Algorithm In Stock Investment Combination

From the day when the stock market formed, the benefits from appropriate investment have been attracting investors and people are not satisfied with the meager profits brought from saving money in bank. Thus, a large amount of money flows into the stock market. However, the cruelty of stock market should also be realized because once wrong judgments have been made the terrible outcomes will appear. Therefore, to find a mature stock investment method is what experts are exploring with efforts, which is quite hard to accomplish due to the comprehensive influencing factors and the complexity of inner and outer system of stock pricing. The application of traditional predicting and statistical tools have become unfit for the complex facts, so the production of a new analysis model in that field is the further object that is searched for.The traditional approach usually applies the genetic algorithm combined with the neural network in the investment forecasting domain of stock market, and this method has been developed to some extent and it mainly focuses on the small-scale prediction of a single stock or several shares. Nevertheless, the algorithm introduced in this paper is to explore stock investment method mainly from a different angle. Usually the more shares investors buy and the more money they invest, the more profits they will gain. Hence, the target algorithm is appropriate for companies and enterprises to gain profits in stock investment without risks. This algorithm depends on genetic algorithm, quadratic programming and the traditional mathematical model of annealing algorithm and has achieved good effects after corresponding improvement.1. The best investment combination is to be computed from a lot of stock set by using the genetic algorithm in order to avoid local convergence of genetic algorithm. We have obtained satisfactory results by adjusting crossover-mutation operator and fitness function, by improving the traditional calculation structures.2. For such a portfolio, we adopt quadratic programming with simulation of shenzhen 300 index to calculate the value of every share of stock and use annealing algorithm to control the annealing temperature and to adjust the conditions for the termination in order to fine-tune optimum fitting effects of this group of weights. 3. All the algorithm testing data adopted are from the real stock market. Through comparing the experiment results of joining annealing algorithm before and after and comparing with other similar algorithms, the feasibility of the presented algorithm has been proved.The whole realization process described in the paper is completed in a company, and this algorithm is also an important part of products of the company. After a great deal of improvement and optimization of the algorithm, the efficiency and accuracy of it have been achieved. Moreover, its advantages are quite obvious compared with similar products, which has justified the facts that a good feasibility of using hybrid genetic simulated annealing algorithm lies in the stock investment portfolios.