基于Metropolis准则的BP模型改进及在太湖Chl-a预测中的应用

冯家成; 李勇; 李娜; 单雅洁; 钱佳宁

doi:10.13205/j.hjgc.202201024

基于Metropolis准则的BP模型改进及在太湖Chl-a预测中的应用

doi: 10.13205/j.hjgc.202201024

冯家成^1,2,
李勇^1,2, ,,
李娜^1,2,
单雅洁²,
钱佳宁²

1. 河海大学浅水湖泊综合治理与资源开发教育部重点实验室, 南京 210098;
2. 河海大学环境学院, 南京 210098

基金项目:

国家自然科学基金项目(51879081)

国家重点研发计划(2018YFC0407906)

详细信息

作者简介:
冯家成(1997-),男,硕士研究生,主要研究方向为河湖水环境模型。student_fengran@163.com

通讯作者:
李勇(1974-),男,教授,主要从事水体和土壤中污染物环境行为、规划评价及生态修复等方面的教学和研究工作。liyonghh@163.com

计量
- 文章访问数: 221
- HTML全文浏览量: 19
- PDF下载量: 3
- 被引次数: 0
出版历程
- 收稿日期: 2021-03-25
- 网络出版日期: 2022-03-30
- 刊出日期: 2022-03-30

IMPROVEMENT OF BP MODEL BASED ON METROPOLIS CRITERION AND ITS APPLICATION IN CHLOROPHYLL-A PREDICTION FOR LAKE TAIHU

FENG Jiacheng^1,2,
LI Yong^{1,2
, ,},
LI Na^1,2,
SHAN Yajie²,
QIAN Jianing²

1. Key Laboratory of Shallow Lake Comprehensive Control and Resource Development, Ministry of Education, Hohai University, Nanjing 210098, China;
2. School of Environment, Hohai University, Nanjing 210098, China

摘要

摘要: 水体富营养化及藻华暴发已成为湖泊治理中的主要问题,利用历史监测数据,采用BP神经网络对水体中叶绿素a(Chl-a)浓度进行预测,已成为藻华预警的主要手段。但该方法存在迭代速度慢、易陷入局部极值等局限性,导致产生拟合结果不优或预测误差较大等问题。利用Metropolis接受准则的全局寻优能力,将其与BP神经网络相结合构建MBP模型,并以太湖水体中Chl-a月均浓度为预测对象进行验证模拟。结果表明,基于2009—2016年太湖月平均相关水环境数据训练出的MBP模型,相较于传统BP神经网络具有以下特点:1)权值在迭代过程的初始阶段能更快地收敛于较优值,模型拟合效果与预测精度也有所提升;2)针对不同的数据情况(数据噪声与样本数量),MBP模型的平均预测误差较传统BP神经网络明显降低,具有较强的鲁棒性与稳定性。构建的MBP模型进一步拓展了传统BP神经网络在叶绿素a浓度预测模型中的应用。
- BP神经网络 /
- Metropolis准则 /
- 模型适应性 /
- 叶绿素a浓度 /
- 太湖
Abstract: Eutrophication and algal blooms have become the main problems in lake management. Predicting chlorophyll-a concentration in water based on historical monitoring data using BP neural network is one of the main means for algal bloom early warning. However, traditional BP method has some limitations, such as low iteration speed and being easy to fall into local extremum, leading to a poor fitting result and larger prediction error. In this paper, a new model(MBP) was developed based on BP neural network by combining with the global optimization capability of Metropolis acceptance criterion, and then applied to predict the monthly average chlorophyll-a concentration of Lake Taihu.Resultsshowed that, comparing with traditional BP neural network, the improved MBP model had a relatively faster coverage velocity at the initial iteration stage, and showed lower fitting error and higher accuracy. Average prediction error of the MBP model was significantly lower than that of the traditional BP neural network. Additionally, the MBP model had a stronger robustness and stability for different data noise and smaller number of samples. This model further expanded the application of traditional BP neural network in predicting concentration of chlorophyll-a and provided a new idea for establishing an early warning system of algal bloom.
- BP neural network /
- Metropolis criterion /
- model adaptability /
- chlorophyll-a concentration /
- Lake Taihu

HTML全文

参考文献(39)

[1]	VINCON-LEITE B,CASENAVE C.Modelling eutrophication in lake ecosystems:a review[J].Science of the Total Environment,2019,651(Pt 2):2985-3001.
[2]	方娜,游清徽,刘玲玲,等.基于云模型的鄱阳湖秋季周边湿地水体富营养化评价[J].生态学报,2019,39(17):6314-6321.
[3]	黄小龙,郭艳敏,万斌,等.沉水植物恢复对城市富营养化湖泊生态环境影响[J].环境工程,2018,36(7):17-21.
[4]	苟婷,马千里,王振兴,等.龟石水库夏季富营养化状况与蓝藻水华暴发特征[J].环境科学,2017,38(10):4141-4150.
[5]	DIMBERG P H,HYTTEBORN J K,BRYHN A C.Predicting median monthly chlorophyll-a concentrations[J].Limnologica,2013,43(3):169-176.
[6]	王斌,马健,王银亚,等.天山天池夏季叶绿素a的分布及富营养化特征研究[J].环境科学,2015,36(7):2465-2471.
[7]	曾一川,王华,渠昊,等.滨湖河网叶绿素a时空分布特征及相关性分析[J].环境工程,2020,38(9):23-30 ,153.
[8]	LIANG Z Y,QIAN S S,WU S F,et al.Using bayesian change point model to enhance understanding of the shifting nutrients-phytoplankton relationship[J].Ecological Modelling,2019,393:120-126.
[9]	XIAO X,HE J Y,HUANG H M,et al.A novel single-parameter approach for forecasting algal blooms[J].Water Research,2017,108:222-231.
[10]	桑文璐,纪道斌,朱士江,等.基于WNN和SVM模型的香溪河Chl-a浓度预测[J].环境科学与技术,2018,41(增刊2):95-99.
[11]	白晓哲,张慧妍,王小艺,等.动态聚类最近邻法在湖库蓝藻水华预测中的应用[J].水土保持通报,2017,37(4):161-165.
[12]	孙菲,袁鹏,程建光,等.宜兴市殷村港叶绿素a与影响因子的多元分析[J].环境工程,2017,35(9):53-57.
[13]	WANG Y J,XIE Z C,LOU I C,et al.Algal bloom prediction by support vector machine and relevance vector machine with genetic algorithm optimization in freshwater reservoirs[J].Engineering Computations,2017,34(2):664-679.
[14]	JAYASEELAN B F,THADIKAMALA S,NAMBALI V V,et al.A novel approach to predict chlorophyll:a in coastal-marine ecosystems using multiple linear regression and principal component scores[J].Marine Pollution Bulletin,2020,152:110902-110909.
[15]	郑震.基于GLUE方法的湖库富营养化风险评估[J].灌溉排水学报,2020,39(6):132-137.
[16]	朱婕,李翠梅,薛天一.基于灰色关联分析-GA-BP模型的叶绿素a含量预测[J].水电能源科学,2020,38(10):25-28 ,147.
[17]	CHEN G,LIU F,MOHAMMAD S,et al.Chlorophyll-a concentration in water:a semi-analytical retrieval model research with a bp network[J].Frontiers in Artificial Intelligence and Applications,2016,281:151-155.
[18]	许阳春,张明峰,苏玉萍,等.基于BP人工神经网络平潭海域赤潮叶绿素a浓度模型演算研究[J].海洋科学,2020,44(3):34-41.
[19]	张雪,郑小慎.基于BP神经网络渤海湾表层叶绿素浓度反演方法探讨[J].海洋技术学报,2018,37(6):79-87.
[20]	裴洪平,罗妮娜,蒋勇.利用BP神经网络方法预测西湖叶绿素a的浓度[J].生态学报,2004,24(2):246-251.
[21]	YAN J Z,XU Z B,YU Y C,et al.Application of a hybrid optimized bp network model to estimate water quality parameters of beihai lake in beijing[J].Applied Sciences,2019,9(9):1863-1875.
[22]	高峰,冯民权,滕素芬.基于PSO优化BP神经网络的水质预测研究[J].安全与环境学报,2015,15(4):338-341.
[23]	LI W,LI G M,ZHANG R X,et al.Carbon reduction potential of resource-dependent regions based on simulated annealing programming algorithm[J].Sustainability,2017,9(7):1161-1177.
[24]	BIN S.Application of genetic algorithm and bp neural network in supply chain finance under information sharing[J].Journal of Computational and Applied Mathematics,2021,384:113170-113180.
[25]	ZHOU F Q,SU Z,CHAI X H,et al.Detection of foreign matter in transfusion solution based on gaussian background modeling and an optimized bp neural network[J].Sensors,2014,14(11):19945-19962.
[26]	陈智军,李洋莹.神经网络BP算法改进及其性能分析[J].软件导刊,2017,16(10):39-41.
[27]	王磊,王汝凉,曲洪峰,等.BP神经网络算法改进及应用[J].软件导刊,2016,15(5):38-40.
[28]	刘智斌,曾晓勤,刘惠义,等.基于BP神经网络的双层启发式强化学习方法[J].计算机研究与发展,2015,52(3):579-587.
[29]	王晶晶,王剑.一种BP神经网络改进算法研究[J].软件导刊,2015,14(3):52-53.
[30]	李海涛,袁森.基于遗传算法和BP神经网络的海洋工程材料腐蚀预测研究[J].海洋科学,2020,44(10):33-38.
[31]	GAO C Y,GAO Y L,LV S S.Improved simulated annealing algorithm for flexible job shop scheduling problems[C]//Control and Decision Conference,IEEE,2016.2223-2228.
[32]	傅文渊,凌朝东.布朗运动模拟退火算法[J].计算机学报,2014,37(6):1301-1308.
[33]	李美水,杨晓华.基于MCMC方法的SWMM模型参数不确定性分析[J].环境科学与技术,2019,42(4):25-30.
[34]	顾杰,王嘉,邓俊晖,等.基于ARIMA模型与BP神经网络算法的水质预测[J].净水技术,2020,39(6):73-82.
[35]	李恩玉,杨平先,孙兴波.基于激活函数四参可调的BP神经网络改进算法[J].微电子学与计算机,2008,25(11):89-93.
[36]	卢建中,程浩.改进GA优化BP神经网络的短时交通流预测[J].合肥工业大学学报(自然科学版),2015,38(1):127-131.
[37]	蒋定国,全秀峰,李飞,等.基于BP神经网络的水体叶绿素a浓度预测模型优化研究[J].南水北调与水利科技,2019,17(2):81-88.
[38]	李娜,李勇,冯家成,等.太湖水体Chl-a预测模型ARIMA的构建及应用优化[J].环境科学,2021,42(5):2223-2231.
[39]	徐逸,董轩妍,王俊杰.4种机器学习模型反演太湖叶绿素a浓度的比较[J].水生态学杂志,2019,40(4):48-57.