PDF《第39 卷第 6 期》

第39 卷第

6 期2019 年6月环境科学学报Acta Scientiae Circumstantiae Vol.

39,No.6 Jun.,

2019 收稿日期:2018?11?26 修回日期:2019?02?13 录用日期:2019?02?13 基金项目: 云南省科技厅科技计划重点研发项目(No.2018BC001) 作者简介: 任婷玉(1994―),女,E?mail:rentingyu@ pku.edu.cn;

?责任作者,E?mail: rz5q2008@ gmail.com DOI:10.13671 / j.hjkxxb.2019.0077 任婷玉,梁中耀,刘永, 等.2019.基于贝叶斯优化的三维水动力?水质模型参数估值方法[J].环境科学学报,39(6):2024?2032 Ren T Y, Liang Z Y, Liu Y, et al. 2019.The parameters estimation method based on Bayesian optimization for complex water quality models[J].Acta Scientiae Circumstantiae,39(6):2024?2032 基于 贝叶斯优化的三维水动力?水质模型参数估值方法 任婷玉1 ,梁中耀1 ,刘永1 ,邹锐1,2,3,? 1. 北京大学环境科学与工程学院,水沙科学教育部重点实验室,北京

100871 2. 北京英特利为环境科技有限公司,锐思计算智能实验室(RCIL),北京

100085 3. 南京智水环境科技有限公司,南京

210012 摘要:随着水质目标管理要求的提升,基于复杂的三维水动力?水质模型的决策成为流域精准治理的必需.水质模型通常具有复杂的结构,包含 大量的方程和参数,而参数取值的准确性会影响模型对水体系统表征的可靠性,进而影响根据模型结果进行水环境管理的效果,因此,有必要 探究适用于复杂水质模型的高效参数估值方法.传统的自动参数估值方法应用于复杂的水质模型时会面临计算瓶颈,而贝叶斯优化适用于高 运算成本模型的优化问题.本研究提出基于贝叶斯优化的复杂水质模型参数估值方法,主要包括:①重要影响参数识别;

②重要参数敏感性排 序与筛选;

③采用贝叶斯优化对筛选出的参数进行估值;

④方法的适用性评估.同时,将该方法应用于云南异龙湖的三维水动力?水质模型的参 数估值中,发现进行参数估值后模型 lg(NSE)均大于 0.65,表明模型达到了满意的级别.研究表明,当贝叶斯优化算法的采集函数为 EI 时,仅需 要141 次迭代 lg(NSE)即可达到 0.766,该方法对复杂水质模型的参数估值具有一定的借鉴意义. 关键词:水质模型;

参数估值;

贝叶斯优化;

高斯过程;

采集函数 文章编号:0253?2468(2019)06?2024?09 中图分类号:X32 文献标识码:A The parameters estimation method based on Bayesian optimization for complex water quality models REN Tingyu1 , LIANG Zhongyao1 , LIU Yong1 , ZOU Rui1,2,3,? 1. College of Environmental Sciences and Engineering, the Key Laboratory of Water and Sediment Sciences, Ministry of Education, Peking University, Beijing

100871 2. Beijing Inteliway Environmental Sci. &

Tech. Ltd., Rays Computational Intelligence Lab(RCIL), Beijing

100085 3. Nanjing Innowater Co. Ltd., Nanjing

210012 Abstract: Reliable decision?making based on complex three?dimensional hydrodynamic and water quality modeling becomes essential under the circumstances of increasingly demand for water management requirement. However, due to the complicated modeling structure, enormous parameters and governing equations, it is extremely difficult if not impossible to obtain reasonable parameters to represent the underlying mechanisms in lake systems, which is the prerequisite of robust decision?making support. It is hence critical to explore highly effective parameter estimation techniques for complex water quality models. Traditional automatic parameter estimation techniques are usually computationally intensive, while Bayesian optimization algorithm has been shown to be able to tackle optimization problems for computational expensive models in a timely manner. In this study, we proposed a Bayesian optimization?based parameter estimation strategy, which includes ① critical parameters identification;