基于IPSO-BP模型的滑坡位移预测

Landslide displacement prediction based on IPSO-BP model

滑坡是一种常见自然灾害。建立滑坡预测模型及时做出预警对避免或减少人员伤亡和财产损失有着重要的意义。以BP神经网络的不足和局限性作为出发点,引入非线性惯性权重改进粒子群优化BP网络模型。提出IPSO-BP模型,并将其应用于四川省一处滑坡预测中。结果表明,IPSO-BP模型比其他模型与实测数据拟合结果更接近,能够有效反映滑坡位移的趋势。

As a common natural disaster,it is important to establish a landslide prediction model to make early warning,for avoiding or reducing the casualties or property losses.Based on the deficiency and limitation of BP neural network,the nonlinear inertia weight improved particle swarm optimization BP network model is introduced.The IPSO-BP model is proposed in this paper and applied to a landslide prediction in Sichuan Province.The results show that the IPSO-BP model is more closer to the measured data than other models.The IPSO-BP model can effectively reflect the trend of landslide displacement.