基于贝叶斯优化算法的新安江模型参数优化及应用
|
摘要
鉴于参数优化是保证水文模型预报效果的重要途径,以息县流域为例,采用贝叶斯优化算法对新安江模型参数进行优化,并与遗传算法进行比较。结果表明,在息县流域日径流预报问题上,贝叶斯优化算法相比遗传算法效率提升显著,精度略优于遗传算法,非常适用于解决新安江模型参数优化问题。
In view of the fact that parameter optimization is an important way to ensure the prediction effect of hydrological model,taking Xixian watershed as an example,Bayesian optimization algorithm was used to optimize the parameters of Xin'anjiang model.Compared with genetic algorithm,the results show that Bayesian optimization has a significant improvement in efficiency and slightly better accuracy of the daily runoff forecasting in Xixian watershed.It is very suitable for solving the parameter optimization problem of Xin'anjiang model.
In view of the fact that parameter optimization is an important way to ensure the prediction effect of hydrological model,taking Xixian watershed as an example,Bayesian optimization algorithm was used to optimize the parameters of Xin'anjiang model.Compared with genetic algorithm,the results show that Bayesian optimization has a significant improvement in efficiency and slightly better accuracy of the daily runoff forecasting in Xixian watershed.It is very suitable for solving the parameter optimization problem of Xin'anjiang model.
引文
[1]宋星原,舒全英,王海波,等.SCE-UA、遗传算法和单纯形优化算法的应用[J].武汉大学学报:工学版,2009,42(1):6-9.
[2]江敏,陈一民.贝叶斯优化算法的发展综述[J].计算机工程与设计,2010,31(14):3 254-3 259.
[3]崔佳旭,杨博.贝叶斯优化方法和应用综述[J].软件学报,2018,29(10):3 068-3 090.
[4] GZ.Probabilistic Machine Learning and Artificial Intelligence[J].Nature,2015,521:452-459.
[5]江敏,陈一民.贝叶斯优化算法的选择策略分析[J].计算机工程与设计,2011,32(1):266-269.
[6] SN,KA,KS,et al.Information-Theoretic Regret Bounds for Gaussian Process Optimization in the Bandit Setting[J].IEEE Transactions on Information Theory,2012,58(5):3 250-3 265.
[7]谢武明,李俊,周峰平,等.基于BP神经网络和遗传算法的污水处理厂电耗预测[J].水电能源科学,2018,36(8):202-204.
[8]刘苏宁,甘泓,魏国孝.粒子群算法在新安江模型参数率定中的应用[J].水利学报,2010,41(5):537-544.
[2]江敏,陈一民.贝叶斯优化算法的发展综述[J].计算机工程与设计,2010,31(14):3 254-3 259.
[3]崔佳旭,杨博.贝叶斯优化方法和应用综述[J].软件学报,2018,29(10):3 068-3 090.
[4] GZ.Probabilistic Machine Learning and Artificial Intelligence[J].Nature,2015,521:452-459.
[5]江敏,陈一民.贝叶斯优化算法的选择策略分析[J].计算机工程与设计,2011,32(1):266-269.
[6] SN,KA,KS,et al.Information-Theoretic Regret Bounds for Gaussian Process Optimization in the Bandit Setting[J].IEEE Transactions on Information Theory,2012,58(5):3 250-3 265.
[7]谢武明,李俊,周峰平,等.基于BP神经网络和遗传算法的污水处理厂电耗预测[J].水电能源科学,2018,36(8):202-204.
[8]刘苏宁,甘泓,魏国孝.粒子群算法在新安江模型参数率定中的应用[J].水利学报,2010,41(5):537-544.