SSO-BP模型在水资源可再生能力评价中的应用

Application of SSO-BP model in evaluation of reproducible capability of regional water resources

为评价区域水资源可再生能力,提出了水资源可再生能力评价指标体系和分级标准,构建了基于BP神经网络的评价模型,并以云南省文山州水资源可再生能力评价为例进行实例研究。首先,遴选出单位面积水资源量等10个指标,构建水资源可再生能力评价指标体系和分级标准;其次,针对BP神经网络初始权值和阈值难以确定的不足,利用一种全新的仿生群体智能算法——群居蜘蛛优化(SSO)算法优化BP神经网络初始参数,提出了SSO-BP评价模型,并通过6个高维复杂函数对SSO算法进行验证,且与粒子群优化(PSO)算法进行对比;最后,利用SSO-BP模型对实例进行水资源可再生能力评价。结果表明:1 SSO算法具有较好的收敛精度和全局寻优能力,可有效提高BP神经网络模型的预测精度和泛化能力。2文山州各评价区域2014年水资源可再生能力处于最强与中等之间,符合区域现状。

To assess the reproducible capability of regional water resources, the evaluation index system for water resources re- producible capability and the classifying standards are put forward, and BP neural network -based evaluation model is built, which is used to evaluate the water resources reproducible capability of Wenshan prefecture, Yunnan Province. Firstly, 10 inde- xes including the water resources amount on unit area etc. are selected to construct the water resources reproducible capability e- valuation index system and classifying standards; secondly, in view of the difficult determination of BP neural network initial weights and thresholds, a new bionic swarm intelligence algorithms, gregarious spider optimization （SSO） algorithm is used to optimize the initial parameters of BP neural network, and the SSO - BP evaluation model is proposed and verified through 6 high dimension complex functions, and compared with particle swarm optimization （PSO） algorithm; at last, the reproducible ability of water resources for Wenshan Prefecture is evaluated with SSO - BP model. The results show that： the SSO algorithm has better convergence precision and global optimization capability, the optimization of the initial weights and thresholds of the BP neural network model by using SSO algorithm can effectively improve the prediction accuracy and generalization ability; The reproducible water resources capability of the evaluation regions in Wenshan prefecture in 2014 was between the strongest and the medium, which was in line with the regional present situation.