Because of the complexity caused by landslides factors,the conventional method is difficult to obtain highly accurate prediction results.The mind evolutionary algorithm has been proved to have strong global optimization ability.Based on the previous research work,this work adopted the evolutionary algorithm,the initial weights and thresholds of samples are optimized,which makes BP get an optimal weight in learning and forecasting,which accelerated the train-ing speed of the network.The results are as follows.The weights of the BP neural network after the optimization of the slope stability to the rate of 100%,while the random weights of the BP neural network against the rate of only 70%,the BP neural network against the rate of only 65%,the GABP neural network against the rate of 85%increased by 30%,35%and 15%.Therefore,the prediction accuracy of MEABP neural network is improved obviously,which is feasible and effective in the evaluation of slope stability in the future.
关键词
思维进化算法(MEA)/BP神经网络/边坡稳定性/预测/全局最优
Key words
mind evolutionary algorithm(MEA)/BP neural network/slope stability/prediction/global opti-mization
维普
万方数据