基于BP神经网络的避洪安置区选址研究

Flood-avoiding Resettlement Area Site Selection Research Based on BP Neural Network

为了解决洪水灾害应急管理中避洪安置区的选址问题,在总结前人研究成果的基础上,从自然环境、可行性、生活条件、社会环境、实用性等5个方面出发,通过17个具体指标,构建了避洪安置区选址评价体系;在安置区的选址分析中引入BP神经网络,构建了基于BP神经网络的安置区选址模型,根据选址评价体系及相关经验确定网络模型各层次的神经元节点数量,并采用样本对模型进行训练分析,建立了稳定的网络结构,最终实例验证了该模型在避洪安置区选址问题中的有效性及可行性。

Natural disasters have frequent happened in our country,the flood has became most serious natural disaster in our country because of high frequency and great damage.In order to solve the flood-avoiding resettlement area site selection problem in the flood hazard emergency management and based on the summary of previous research achievements,establish the flood-avoiding resettlement area site selection assessment system which starts from the five aspects of natural environment,feasibility,living condition,social environment and practicability and by means of 17 specific indicators;introduce the BP neural network in the resettlement area site selection analysis,establish the resettlement area site selection model based on BP neural network,based on neuron node quantity of site selection assessment system and each level of network model confirmed according to relevant experiences and adopted certain samples to train the model,establish the stable network model and final examples verified the validity and feasibility of this model in the flood-avoiding resettlement area site selection problem.