| After 25 years development, The Grey System has permeated through politics, economy and culture and so on. It obtains great progress space in many domains. On account of its wide applying bound, the high simulating precision and the simple computing process, The Grey System is welcomed by people. GM(1,1) Model is the most popular dynamic grey model during all the grey models because that it can simulate and forecast the business' growing trend and transformation law, although its structure is simple, The model is used by most people, but that is not to say that the model is all-purpose, In fact, people find that this model can't be used to simulate some data well, what is worse, the simulated results is fuzzy entirely. By analyzing and researching the model, we have the conclusion as follows:(1) The original data can't fulfill modeling condition commendably.(2) The precision of the approximate method is low in the model's solving process.Based on these two aspects, the thesis comprises four pieces of works just as follows:(1) By summarizing other methods, the thesis puts forward the composite method which can enhance the lubricity.(2) The main error source is the low computing precision of the digital integral by analyzing the parameter solving in conditional methods. This thesis puts forward a method which we called 3-order Newton Cotes formula whose precision is higher than trapezoid formula used in the traditional methods.(3) We analyze the GM(1,1) Model's white differential equation deeply, then we know that its form is consistent with 1-order differential equation whose variable can be separated.Based on this fact,this thesis puts forward a new method-separating variable method.(4) Abundant data is used to validate the methods raised in the thesis, and then we draw the conclusion that the composite method can improve the original data's lubricity effectively. The two methods raised in this thesis both have superior precision which possess practical applying value.
|
PDF Full Text Request