Fuzzy-neural Gray Forecasting Algorithm And It’s Application Research On TAC-SCM

The research of Price Forecasting Algorithm, an indispensable part of bid in themanagement of supply chain, plays an important role in simulating unknown price andproviding theoretical and practical support for corresponding orders. The essay analyses thehistorical price, stimulate the combination of fuzzy neural gray prediction algorithm, andpredicts its rule of development and offers data support to supply chains.During the period of information process, different prediction methods emphasizedifferently. Just by simply using a single prediction algorithm, some useful information wouldbe abandoned, which in the end may results in an incomplete prediction and the declination ofprediction accuracy. Therefore, multiple prediction algorithms are often combined to reach acomprehensive one. Combined forecasting algorithm, through comprehensive analysis onprediction model in prediction, can reduce the impact the loss of factors, which to a greaterdegree helps to analyses all kinds of information, so as to improve the accuracy of prediction.At present, the combined forecasting model is now used in all kinds of areas and achievesgood results.GM (1,1) model, because of its good analysis qualities like " small sample size " and "incomplete information " and its simple and practical advantages, plays an important part inthe gray prediction. As a result of its suitability for the model with a little information ofsupply chain history, gray algorithm and other predicted algorithms are combined to make thebest use of complementary characteristics of combination algorithm and various algorithms.The main innovation points and contributions of the essay are as following:In order to solve the problem of inaccuracy prediction when gray prediction in the longerdata prediction, I introduce a fuzzy neural network prediction algorithm, firstly make thehistorical data to achieve a gray prediction, and get the prediction error through a long-termtraining set by the study of neural network, which on the other hand retains the advantages ofgray forecasting in a short-term prediction, plays a neural network to predict the learningmechanism, and forms a combined forecasting algorithm to improve the forecasting precision.In the end, the combination of the fuzzy gray prediction algorithm is applied to thesolution of TAC-SCM contest price forecasting problems of supply chain. Many experiments prove the accuracy of fuzzy gray prediction algorithm in prediction.