Abstract: In order to solve the problem of low prediction precision and long prediction time of steam flooding
development effect, we propose a novel prediction method, which is based on the combination of improved
artificial bee colony algorithm and RBF (Radial Basis Function) neural network. In the proposed method, we
apply the optimal solution of the population to modify the search equation of the employed bees and the onlooker
bees, perform the random perturbation of the population optimal solution and individual search solution with the
idea of differential evolution algorithm, and adopt hybrid encoding to optimize the parameters of RBF neural
networks. We use the Qi 40 block of Liaohe Oilfield as an example and make a trial calculation. The trial results
show that the method has better nonlinear fitting ability and higher prediction accuracy for steam flooding
development effect prediction.

Key words: steam flooding, artificial bee colony algorithm, prediction model, radial basis function (RBF) neural network, random perturbation