针对短期风电功率预测,提出一类基于稀疏高斯过程(sparse gaussian processes,Sparse-GP)的概率预测方法。通过对数据集随机划分所形成的数据子集,给出基于数据点子集(subset of datapoints,SoD)近似、回归子集(subset of regressors,S...针对短期风电功率预测,提出一类基于稀疏高斯过程(sparse gaussian processes,Sparse-GP)的概率预测方法。通过对数据集随机划分所形成的数据子集,给出基于数据点子集(subset of datapoints,SoD)近似、回归子集(subset of regressors,SoR)近似、投影过程(projected process,PP)近似算法的3种Sparse-GP方法,该方法不仅能给出模型的均值预测,而且能获取模型的预测方差,这很好地解释了模型置信水平。不同的Sparse-GP方法在保持常规GP方法优点的同时,还能解决GP方法随着训练数据增加而产生的矩阵运算困难等难题,且计算效率高。将具有不同协方差函数形式的Sparse-GP方法应用于不同地区的短期风电功率单步与多步预测实例中,在同等条件下还与常规GP、SVM方法进行对比。实验结果表明,Sparse-GP方法可以给出较好的预测效果,且适用于较大规模数据集的训练。展开更多
This paper describes the development and application of a new mathematical technique to include stochasticity in environmental models. The techniques are studied to develop stochastic analysis model of the CBOD NBOD D...This paper describes the development and application of a new mathematical technique to include stochasticity in environmental models. The techniques are studied to develop stochastic analysis model of the CBOD NBOD DO equations for predicting the river water quality. The CBOD NBOD DO equations are expanded to include stochastic terms. Stochastic terms are separated from non fluctuating terms, and the resulting set of equations is solved simultaneously. Solutions are used to calculating the integral solutions and the distribution moments of the state variables. The moments are used in conjunction with the Fokker Planck equation to produce an analytical solution for the probability density functions of the dependent variables. Because the approach produces analytical solutions, it offers greater flexibility than a Monte Carlo approach in treating complex environmental situations. The model is used to calculate water quality for the Qingao river. The calculated results are in good agreement with the Monte Carlo model and field data.展开更多
We study the baryon-baryon interactions with strangeness S=-2 and corresponding momentum correlation functions in leading order covariant chiral effective field theory.The relevant low energy constants are determined ...We study the baryon-baryon interactions with strangeness S=-2 and corresponding momentum correlation functions in leading order covariant chiral effective field theory.The relevant low energy constants are determined by fitting to the latest HAL QCD simulations,taking into account all the coupled channels.Extrapolating the so-obtained strong interactions to the physical point and considering both quantum statistical effects and the Coulomb interaction,we calculate the ΛΛ and Ξ^(-)p correlation functions with a spherical Gaussian source and compare them with recent experimental data.We find a good agreement between our predictions and the experimental measurements by using the source radius determined in proton-proton correlations,which demonstrates the consistency between theory,experiment,and lattice QCD simulations.Moreover,we predict the Σ^(+)Σ^(+),Σ^(+)Λ,and Σ^(+)Σ^(-) interactions and corresponding momentum correlation functions.We further investigate the influence of the source shape and size of the hadron pair on the correlation functions studied and show that the current data are not very sensitive to the source shape.Future experimental measurements of the predicted momentum correlation functions will provide a non-trivial test of not only SU(3) flavor symmetry and its breaking but also the baryon-baryon interactions derived in covariant chiral effective field theory.展开更多
文摘针对短期风电功率预测,提出一类基于稀疏高斯过程(sparse gaussian processes,Sparse-GP)的概率预测方法。通过对数据集随机划分所形成的数据子集,给出基于数据点子集(subset of datapoints,SoD)近似、回归子集(subset of regressors,SoR)近似、投影过程(projected process,PP)近似算法的3种Sparse-GP方法,该方法不仅能给出模型的均值预测,而且能获取模型的预测方差,这很好地解释了模型置信水平。不同的Sparse-GP方法在保持常规GP方法优点的同时,还能解决GP方法随着训练数据增加而产生的矩阵运算困难等难题,且计算效率高。将具有不同协方差函数形式的Sparse-GP方法应用于不同地区的短期风电功率单步与多步预测实例中,在同等条件下还与常规GP、SVM方法进行对比。实验结果表明,Sparse-GP方法可以给出较好的预测效果,且适用于较大规模数据集的训练。
文摘This paper describes the development and application of a new mathematical technique to include stochasticity in environmental models. The techniques are studied to develop stochastic analysis model of the CBOD NBOD DO equations for predicting the river water quality. The CBOD NBOD DO equations are expanded to include stochastic terms. Stochastic terms are separated from non fluctuating terms, and the resulting set of equations is solved simultaneously. Solutions are used to calculating the integral solutions and the distribution moments of the state variables. The moments are used in conjunction with the Fokker Planck equation to produce an analytical solution for the probability density functions of the dependent variables. Because the approach produces analytical solutions, it offers greater flexibility than a Monte Carlo approach in treating complex environmental situations. The model is used to calculate water quality for the Qingao river. The calculated results are in good agreement with the Monte Carlo model and field data.
基金Supported by the National Natural Science Foundation of China (11735003, 11975041, 11961141004)。
文摘We study the baryon-baryon interactions with strangeness S=-2 and corresponding momentum correlation functions in leading order covariant chiral effective field theory.The relevant low energy constants are determined by fitting to the latest HAL QCD simulations,taking into account all the coupled channels.Extrapolating the so-obtained strong interactions to the physical point and considering both quantum statistical effects and the Coulomb interaction,we calculate the ΛΛ and Ξ^(-)p correlation functions with a spherical Gaussian source and compare them with recent experimental data.We find a good agreement between our predictions and the experimental measurements by using the source radius determined in proton-proton correlations,which demonstrates the consistency between theory,experiment,and lattice QCD simulations.Moreover,we predict the Σ^(+)Σ^(+),Σ^(+)Λ,and Σ^(+)Σ^(-) interactions and corresponding momentum correlation functions.We further investigate the influence of the source shape and size of the hadron pair on the correlation functions studied and show that the current data are not very sensitive to the source shape.Future experimental measurements of the predicted momentum correlation functions will provide a non-trivial test of not only SU(3) flavor symmetry and its breaking but also the baryon-baryon interactions derived in covariant chiral effective field theory.
