【摘要】 以泾河流域为例,通过对该流域2014-2016年各季节水质与水生态环境指标的提取与测定,构建了能够反映流域水环境整体状况的指标体系和分级标准;然后利用MATLAB语言建立了L-M优化改进的BP网络结构,并科学、准确地评价了泾河流域水环境承载力状况。研究表明:L-M算法不仅克服了经典BP网络训练速度慢、局部陷入极小的缺陷,而且可提高输出结果的有效性与准确性;泾河流域2014-2016年的水环境承载力分别为0.67、0.52、0.48,整体处于基本可承载水平,并表现出逐渐下降的趋势。更多还原

【Abstract】 Taking Jinghe River Basin for example, this paper constructs an index system and classification standard which can reflect the overall situation of water environment that through the extraction and determination of water quality and ecological environment indexes in each season from 2014 to 2016. Then, the L-M optimized and improved BP network structure are established by using MATLAB, the water environment carrying capacity of Jinghe River is evaluated scientifically and accurately. The results show that the L-M algorithm not only overcomes the defects of slow training speed and small local immersion in the classical BP network, but also improves the validity and accuracy of the output results. The water environment carrying capacity of Jinghe from 2014 to 2016 was 0.67, 0.52 and 0.48, respectively. The overall capacity is at the basic load-bearing level and shows a gradual downward trend.更多还原