Research On Demand Forecast Of Supply Chain In Cloud Computing

With its flexible service, resource sharing, convenient network access features, a better development space and innovative management mode are brought for enterprise by Could Computing, which also provides a new direction and solution for the development of supply chain, and if provides a good integrated environment for supply chain loose coupling, feature of flexibility and building. The environment of supply chain in cloud computing changes a lot from the traditional supply chain. The error of demand forecast will be spread faster and be amplified because of the information sharing of supply chain enterprises in cloud computing platform. The accuracy of supply chain demand forecast in cloud computing is the core and key to subsequent collaborative operation of supply chain. Focusing on the issue of supply chain demand forecast in cloud computing, the rest of this thesis is organized as follow.First of all, a new thought which used the Multi-Agent System has been put forward to solve the problems of supply chain in cloud computing. According to the C-SCRA reference architecture model of supply chain management in cloud computing, modeling and analysis the method of Petri Net, combining the thought of Multi-Agent to make the demand forecast of supply chain, and demonstrate the feasibility of it.Secondly, according to the features of supply chain flexible topology changes in cloud computing environment that may cause the unstable of demand, Markov-grey demand forecast algorithm was proposed which is based on the weight of fitting precision. Using the grey prediction algorithm for the first demand forecasting, then weighted fitting precision by using the autocorrelation coefficient, get the range of grey prediction fitting accuracy using Markov prediction, to correct the first round of grey demand forecast value and get the final value. Finally, the experimental results demonstrate that the algorithm is reasonable and feasible.Thirdly, according the characteristic of the influence of many factors and mutual influence between them in cloud computing supply chain, a new demand forecast method of grey system and multi variable has been proposed based on BP neural network. Because of the characteristic of data sharing and marking of cloud computing platform, we can get the amount of data of the long-term cooperation supply chain to make the Multi variable grey forecast, then training the multi variable grey prediction error with the BP neural network to make up the shortage of grey forecasting model in long-term forecasting of the demand, so we can make a long-term demand forecast of the supply chain in cloud computing. Finally, verify the effectiveness of the proposed method through the examples.