Application Of Improved Gray Genetic Algorithm In Price Anticipating

Application of Improved Gray Genetic Algorithm in Price AnticipatingIn China, the largest meat consumption is pork, the price of pork not only affects people's life, but also makes an impact on our future inflation trends.In the 2007, the price of pork quickly increased, let people can't bear it. Even the charge department caught off guard. Even the relevant departments have caught flat-footed, how to establish an early pork prices warning system, is an important topic of agricultural informatization.Gray system theory, which studies the little experiences and uncertainty data, The information is "less data uncertainty", the gray theoretical research's purpose is "to highlight information optimization, research real law." Gray theoretical prediction model is GM (1,1), multi-variable gray model GM (1, N), gray Verhulst model.Genetic Algorithm is a simulation of Darwinian biological evolution, natural selection and genetic mechanism of the process of biological evolution computing model, is a simulation of natural evolutionary processes through the optimal solution search methods, genetic algorithm consists of three basic operations:select, crossover and mutation.The subject of papers from the National Natural Science Foundation of major projects dealing with non-normative knowledge of the basic theory and core technology, the price of pork pigs prediction. Biological productivity will have a certain period, if we can expect farmers to keep their things in the listing of the price increase, we are able to make whether or not to sell, so that the interests of their own to maximize. The factors affect price fluctuations, as there are many related raw material prices, the government's macro-regulation, suppliers, but the self protection of short-term price has certain law, also can be predicted. When the expected price have been known, after month-end can provide effective strategies for specified. Affect the price of live pigs'prices, corn prices, soybean meal prices, if use GM (1,1) prediction model, the price of live pigs in the past through is not well reflect other factors on the impact on the price of live pigs, and often even if the predicted value of the residual dealt with because of the price and the actual existence of an unacceptable deviation. Gray prediction model GM (1, N) through the gray relational grade to find the main factors that affect price, data prediction. However, albino background value determined in accordance with principles of equilibrium, so that the factors that affect price again to the weakened. The use of genetic algorithm GM (1, N) model uses the adaptive method to search for a reasonable albino background values, using the differential form of the GM (1, N) for improvement, promotion, so that the price correlation with the pigs the most important factor to stand out so that prices forecast and actual values.Gray theory can in less data, uncertain state of short-term forecast, it is very consistent with the characteristics of a simple price, so you can use gray prediction model for price forecasting model from the multi-variable gray GM (1, N) of the definition of the form gray differential operation, using the relationship between roots and coefficients of the matrix will be of differential equation group, through the matrix to simplify the differential equation after the pre-and post differential operation, and to get through the line of integration of the form of two formula (?) According to the principle of least squares matrix, found a correlation coefficient of gray ash model calculation with the linkages between the various data and to use genetic algorithm to search the global optimal solution features to look for in a reasonable range so that the original value of p averaging, the value of trends affecting.By three kinds of models to simulate value and a comparison between the true value was found using the genetic algorithm improved multi-variable gray model GM (1, N, p) has markedly improved the prediction accuracy, it is not a single analog value, nor and then the average value of all kinds, using the fitness function can be set up to find a global optimal solution within a range, so the price you want to predict the enhanced role of the relevant factors, but also able to calculate this factor into each step, and the use of Matlab web server to establish a network server, apply it in the network..Factors affecting the price of key relationships, included, cost, supply and demand, national policies, the cost of pigs' prices are the price of pigs and feed prices, supply and demand is the slaughter, and market demand, national economic policies on the indirect effects of price formation, mainly for national implementation of certain economic policies will cause the cost of goods and factors such as changes in market supply and demand, giving rise to changes in prices, If the purchase price will be affected countries to raise feed grain prices, meaning that prices can not do long-term forecasts, can only play a certain period of time for reference. It can reflect the general price trend for some time. The improved algorithm can be applied not only in terms of price forecasts, but also able to use the production forecast, the number of forecasts, and projections to meet the multi-variable relationship.There are many prediction models, like regression forecasting, time series prediction, neural network prediction, Markov projections, a single prediction method is often unsatisfactory, if we can make several of them, according to different forecasting methods features integrated together to make a comprehensive forecasting system, to provide network supporting.