Prediction of Hydrocarbon Reservoir Parameter Using a GA-RBF Neural Network

Prediction of hydrocarbon reservoir characteristics using seismic at tributes is a very complicated problem with much nonlinear relation. The traditional BP neural network with a gradient decent approach may lead to local minima problem, resulting in the production of unstable and non-convergent solutions. To solve these problems and improve the precision, this paper introduces a GA-based optimized method of RBF neural network. A case studs shows that the GA-RBF algorithm not only works with high predicting precision comparable to real measured data in oil reservoir thickness, but also is superior to that of tradition BP neural network.