The Using Of Blended BP Algorithm With Genetic Algorithm In Subway Construction's Sink Predicting

Currently, along with the approach of the Olympic game, city orbit transportationin Peking is developing fastly. So, The problem of the ground sink caused by subwayconstruction is more important than before. sink and decline result can be estimatedby subway construction's sink predicting in advance, which has important realisticmeaning to the safety of the subway engineering construction.Now,the methods about subway construction's sink predicting have such asexperience formula;model examination;numerical analysis and intelligentdecision-making. This paper main introduce the basal theory about the PECK formulamethod and finite element method of numerical analysis. Emphasizing to analyze theBP network's theory and the method of the improvement BP network. In themeantime, combine genetic evolution from genetic algorithms and back Propagationtraining method of gradient decreasing to form a new blended BP algorithm whichhave two kinds of methods of characteristics.This text do some example analysis about the tenth line subway engineering inPeking, make use of blended BP algorithm, carry on training the date which wasfrom the construction's measure date and the engineering geology parameter data.After this ,input the subsequent engineering geology parameter date and predict theprobably sink. finally contrast the predicting result with actual measure date, we cansee that the method have quickly, strongly function characteristics. the method cangive some reference to the subsequent engineering's construct plan. In addition, this isa new attempt to the sinking predicting method.