基于SA-PS0-EIman的万源市地灾危险性评价方法

打开文本图片集

中图分类号：TP183；P694 文献标识码：A 文章编号：2096-4706（2025）19-0141-04

Abstract： Inorder to improve the risk evaluationaccuracyof geological hazards，theElman Neural Network modelbased onSimulatedAnnealingalgorithmoptimizedwithParticleSwarmOptimization （SA-PSO-Elman）isestablishedforthenonlinear charactersticsofgeologicalhazardriskchange，andtraininganalysisisconductedbycombiningthegeologicalhzarddataof Wanyuan City.TheresultsshowthatcomparedwiththeoptimalvaluesofthecommonlyusedBP，ElmanandPSO-ElmanNeural Networkmodels，theMeanSquareEror（MSE）ofSA-PSO-Elmanisreducedby0.0364，theMeanAbsolutePercentageError （MAPE） isreduced by 3.45% ，and thenumberof its iterations isreducedby6 times，which indicates that the SA-PSO-Elman Neural Network model has higher evaluation accuracy and faster convergence speed.

KeyWords： geological hazard; risk evaluation; Simulated Annealing algorithm; Particle Swarm Optimization;Elman Jeural Network

0 引言

万源市地处四川省东北部，位于大巴山腹心地带，地形地貌复杂多样，境内山峦起伏，沟壑纵横。（剩余4979字）