The application of multilayer dynamic forward net in predicting of oil field system

In order to eliminate limitations of conventional modeling and dynamic prediction methods, multilayer dynamic forward networks are considered as the models of oil field systems, the prediction models and technology of multilayer dynamic forward networks are studied. The deficiency of a recursive prediction error learning algorithm is analysed. An improvement scheme is given. So, the algorithm performance is improved. Thus the method of modeling and prediction for an oil field is renewed. During using the new scheme, excellent results have been obtained which proves that the new scheme is very effective