A neuro-fuzzy GA-BP method of seismic reservoir fuzzy rules extraction

摘要

In this paper, we have prospered a new method to generate fuzzy rules using a genetic algorithm, back propagation, and fuzzy neural networks (FNN) algorithm for seismic reservoir fuzzy rules extraction. This method aims to combine the advantages of fuzzy systems (FS), artificial neural networks (ANN), and GA algorithms and to remedy their drawbacks. The hybrid algorithm can optimize not the number of rules but the membership functions of the antecedent and consequent by adopting multi-encoding of GA. Fuzzy IF/THEN rules were extracted from the optimized FNN.The extracted rules can help to reason the reservoir thickness and decide the optimal drill position in oil field exploration.