Bayesian regularized BP neural network（BRBPNN） technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of t...Bayesian regularized BP neural network（BRBPNN） technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of the network were obtained. After the selection of input variables using stepwise/multiple linear regression method in SPSS i1.0 software, the BRBPNN model was established between chlorophyll-α and environmental parameters, biological parameters. The achieved optimal network structure was 3-11-1 with the correlation coefficients and the mean square errors for the training set and the test set as 0.999 and 0.000？8426, 0.981 and 0.0216 respectively. The sum of square weights between each input neuron and the hidden layer of optimal BRBPNN models of different structures indicated that the effect of individual input parameter on chlorophyll- α declined in the order of alga amount 〉 secchi disc depth（SD） 〉 electrical conductivity （EC）. Additionally, it also demonstrated that the contributions of these three factors were the maximal for the change of chlorophyll-α concentration, total phosphorus（TP） and total nitrogen（TN） were the minimal. All the results showed that BRBPNN model was capable of automated regularization parameter selection and thus it may ensure the excellent generation ability and robustness. Thus, this study laid the foundation for the application of BRBPNN model in the analysis of aquatic ecological data（chlorophyll-α prediction） and the explanation about the effective eutrophication treatment measures for Nanzui water area in Dongting Lake.展开更多
The modified sub regular solution model was used for a calculation of the activity coefficient of immiscible binary alloy systems. The parameters needed for the calculation are the interaction parameters, λ 1 a...The modified sub regular solution model was used for a calculation of the activity coefficient of immiscible binary alloy systems. The parameters needed for the calculation are the interaction parameters, λ 1 and λ 2, which are represented as a linear function of temperature, T . The molar excess Gibbs free energy, G m E, can be written in the form G m E= x A x B[( λ 11 + λ 12 T )+( λ 21 + λ 22 T ) x B ] The calculation is carried out numerically for three immiscible binary alloy systems, Al Pb, Cu Tl and In V. The agreement between the calculated and experimentally determined values of activity coefficient is excellent.展开更多
Many methods have been put forward to solve unconstrained optimization problems,among which conjugate gradient method(CG)is very important.With the increasing emergence of large⁃scale problems,the subspace technology ...Many methods have been put forward to solve unconstrained optimization problems,among which conjugate gradient method(CG)is very important.With the increasing emergence of large⁃scale problems,the subspace technology has become particularly important and widely used in the field of optimization.In this study,a new CG method was put forward,which combined subspace technology and a cubic regularization model.Besides,a special scaled norm in a cubic regularization model was analyzed.Under certain conditions,some significant characteristics of the search direction were given and the convergence of the algorithm was built.Numerical comparisons show that for the 145 test functions under the CUTEr library,the proposed method is better than two classical CG methods and two new subspaces conjugate gradient methods.展开更多
To confirm sub-regular solution model valid for predicting the activity of component in binary oxide systems, seven systems in the whole concentration and twelve systems presenting saturation concentration have been s...To confirm sub-regular solution model valid for predicting the activity of component in binary oxide systems, seven systems in the whole concentration and twelve systems presenting saturation concentration have been studied. The total average relative errors of component 1 and 2 are 3.2 % and 4.1% respectively by application of the sub-regular solution model into the systems within the whole concentration. However, the total average relative errors are 16 % and 1088 % in the systems presenting saturation concentration. The results show that sub-regular solu- tion model is not good for predicting the systems presenting saturation concentration, especially for the systems con- taining acidic or neutral oxide. The reason may be that the influence of the two types of oxide on the configuration is greater in binary oxide systems. These oxides can be present in the form of complex anion partly, Si-O, Al-O, Ti-O and so on, for example （SiO4）4-. That is contrary to sub-regular solution model which is supposed that the oxide systems consist of cation and O2-. But compared with regular solution model and quasi-regular solution model, sub- regular solution model is closer to the characteristics of actual solution and the calculated results are superior.展开更多
The activity of FetO is very important in ironmaking and steelmaking process.In order to predict the activ-ity of Fe,O and optimize the operation conditions in ironmaking and steelmaking process,by application of regu...The activity of FetO is very important in ironmaking and steelmaking process.In order to predict the activ-ity of Fe,O and optimize the operation conditions in ironmaking and steelmaking process,by application of regular so lution model in molten slag systems,FeO-Fe2 O3-SIO2 ternary system,FeO-Fe2 O3-SiO2-CaO and FeO-Fe2 O3-SiO2-NiO quaternary systems have been studied by the chemical equilibrium between H2/H20 gas mixture and liquid slag con tained in solid iron.The values of interaction energy between cations concerning steelmaking slags have been deter-mined by application of ferric-ferrous iron equilibrium and iron-ferric iron equilibrium.And then the activity of Fe,O can be calculated.The results show that the relative error is 3.9%in FeO-Fe2O3-SiO2 system and 18%in FeO-Fe2O3-SiO2 CaO system.The prediction of activities of FetO in the systems are in good agreement with the measure-ments and the regular solution model is valid for predicting the activity of FetO in complex molten slags systems.The activity of Fe,O in FeO-Fe20a-NiO system have not been tested presently,and the calculated result can not be assessed.展开更多
To improve forecasting and sustained control level of underground pests, trapping quantity of underground pests （black cutworm,mole cricket and scar-ab） by lamps and their field dynamics in Hangzhou district from 20...To improve forecasting and sustained control level of underground pests, trapping quantity of underground pests （black cutworm,mole cricket and scar-ab） by lamps and their field dynamics in Hangzhou district from 2005 to 2011 were investigated in the paper. The results showed that different pests had obvious differences in population dynamic. The black cutworm （Agrotis ypsilon） had several damage peaks （late May, late June and late July） and the moth amount in early period was relatively high. The mole cricket （ Gryllotalpa africana） had two damage peaks （late May to early July, early September to mid and late October）. The scarab （Anomala corpulenta） had one damage peak （late May to late June）. There were periodic changes in total quantity of underground pests among years, and the peak period appeared in the year of 2005, 2007 to 2009 and 2011, respectively. On this basis, temperature, humidity, rainfall and light were used as forecas- ting factors, using the method of stepwise regression, 19 factors with significant correlation were screened out and prediction models for occurrence quantity and oc- currence period of the three pests were established. By using accuracy degree judge model for verification, the score values of prediction model for occurrence quan-tity and occurrence period of the three underground pests were more than 58 and 70, which indicated that the historical coincident rate and prediction accuracy of estabhshed prediction models were good.展开更多
Modified quasi-regular solution model has been available based on the sub-regular solution model and quasiregular solution model in this article. The three parameters of this model are set by the computer-aided analys...Modified quasi-regular solution model has been available based on the sub-regular solution model and quasiregular solution model in this article. The three parameters of this model are set by the computer-aided analysis based on the experimental data of activity at two different temperatures. Seven binary molten slag systems in the whole concentration are calculated by application of the model and the average relative error is within 10%. Ten binary molten slag systems presenting saturation concentration are also calculated, but the average relative error is bigger, especially for the systems containing acidic oxide and neutral oxide. The results show that the calculated results are superior to those calculated by application of regular solution model, sub-regular solution model and quasi-regular solution model.展开更多
Theoretical results related to properties of a regularized recursive algorithm for estimation of a high dimensional vector of parameters are presented and proved. The recursive character of the procedure is proposed t...Theoretical results related to properties of a regularized recursive algorithm for estimation of a high dimensional vector of parameters are presented and proved. The recursive character of the procedure is proposed to overcome the difficulties with high dimension of the observation vector in computation of a statistical regularized estimator. As to deal with high dimension of the vector of unknown parameters, the regularization is introduced by specifying a priori non-negative covariance structure for the vector of estimated parameters. Numerical example with Monte-Carlo simulation for a low-dimensional system as well as the state/parameter estimation in a very high dimensional oceanic model is presented to demonstrate the efficiency of the proposed approach.展开更多
The application of Tikhonov regularization method dealing with the ill-conditioned problems in the regional gravity field modeling by Poisson wavelets is studied. In particular, the choices of the regularization matri...The application of Tikhonov regularization method dealing with the ill-conditioned problems in the regional gravity field modeling by Poisson wavelets is studied. In particular, the choices of the regularization matrices as well as the approaches for estimating the regularization parameters are investigated in details. The numerical results show that the regularized solutions derived from the first-order regularization are better than the ones obtained from zero-order regularization. For cross validation, the optimal regularization parameters are estimated from L-curve, variance component estimation（VCE） and minimum standard deviation（MSTD） approach, respectively, and the results show that the derived regularization parameters from different methods are consistent with each other. Together with the firstorder Tikhonov regularization and VCE method, the optimal network of Poisson wavelets is derived, based on which the local gravimetric geoid is computed. The accuracy of the corresponding gravimetric geoid reaches 1.1 cm in Netherlands, which validates the reliability of using Tikhonov regularization method in tackling the ill-conditioned problem for regional gravity field modeling.展开更多
In order to decrease the sensitivity of the constant scale parameter, adaptively optimize the scale parameter in the iteration regularization model （IRM） and attain a desirable level of applicability for image denoi...In order to decrease the sensitivity of the constant scale parameter, adaptively optimize the scale parameter in the iteration regularization model （IRM） and attain a desirable level of applicability for image denoising, a novel IRM with the adaptive scale parameter is proposed. First, the classic regularization item is modified and the equation of the adaptive scale parameter is deduced. Then, the initial value of the varying scale parameter is obtained by the trend of the number of iterations and the scale parameter sequence vectors. Finally, the novel iterative regularization method is used for image denoising. Numerical experiments show that compared with the IRM with the constant scale parameter, the proposed method with the varying scale parameter can not only reduce the number of iterations when the scale parameter becomes smaller, but also efficiently remove noise when the scale parameter becomes bigger and well preserve the details of images.展开更多
文摘Bayesian regularized BP neural network（BRBPNN） technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of the network were obtained. After the selection of input variables using stepwise/multiple linear regression method in SPSS i1.0 software, the BRBPNN model was established between chlorophyll-α and environmental parameters, biological parameters. The achieved optimal network structure was 3-11-1 with the correlation coefficients and the mean square errors for the training set and the test set as 0.999 and 0.000？8426, 0.981 and 0.0216 respectively. The sum of square weights between each input neuron and the hidden layer of optimal BRBPNN models of different structures indicated that the effect of individual input parameter on chlorophyll- α declined in the order of alga amount 〉 secchi disc depth（SD） 〉 electrical conductivity （EC）. Additionally, it also demonstrated that the contributions of these three factors were the maximal for the change of chlorophyll-α concentration, total phosphorus（TP） and total nitrogen（TN） were the minimal. All the results showed that BRBPNN model was capable of automated regularization parameter selection and thus it may ensure the excellent generation ability and robustness. Thus, this study laid the foundation for the application of BRBPNN model in the analysis of aquatic ecological data（chlorophyll-α prediction） and the explanation about the effective eutrophication treatment measures for Nanzui water area in Dongting Lake.
文摘The modified sub regular solution model was used for a calculation of the activity coefficient of immiscible binary alloy systems. The parameters needed for the calculation are the interaction parameters, λ 1 and λ 2, which are represented as a linear function of temperature, T . The molar excess Gibbs free energy, G m E, can be written in the form G m E= x A x B[( λ 11 + λ 12 T )+( λ 21 + λ 22 T ) x B ] The calculation is carried out numerically for three immiscible binary alloy systems, Al Pb, Cu Tl and In V. The agreement between the calculated and experimentally determined values of activity coefficient is excellent.
基金Sponsored by the National Natural Science Foundation of China(Grant No.11901561).
文摘Many methods have been put forward to solve unconstrained optimization problems,among which conjugate gradient method(CG)is very important.With the increasing emergence of large⁃scale problems,the subspace technology has become particularly important and widely used in the field of optimization.In this study,a new CG method was put forward,which combined subspace technology and a cubic regularization model.Besides,a special scaled norm in a cubic regularization model was analyzed.Under certain conditions,some significant characteristics of the search direction were given and the convergence of the algorithm was built.Numerical comparisons show that for the 145 test functions under the CUTEr library,the proposed method is better than two classical CG methods and two new subspaces conjugate gradient methods.
基金Item Sponsored by National Natural Science Foundation of China(50764006,50574045)Yunnan Basic Applied Research Foundation of China(2006E0021M)
文摘To confirm sub-regular solution model valid for predicting the activity of component in binary oxide systems, seven systems in the whole concentration and twelve systems presenting saturation concentration have been studied. The total average relative errors of component 1 and 2 are 3.2 % and 4.1% respectively by application of the sub-regular solution model into the systems within the whole concentration. However, the total average relative errors are 16 % and 1088 % in the systems presenting saturation concentration. The results show that sub-regular solu- tion model is not good for predicting the systems presenting saturation concentration, especially for the systems con- taining acidic or neutral oxide. The reason may be that the influence of the two types of oxide on the configuration is greater in binary oxide systems. These oxides can be present in the form of complex anion partly, Si-O, Al-O, Ti-O and so on, for example （SiO4）4-. That is contrary to sub-regular solution model which is supposed that the oxide systems consist of cation and O2-. But compared with regular solution model and quasi-regular solution model, sub- regular solution model is closer to the characteristics of actual solution and the calculated results are superior.
基金Sponsored by National Natural Science Foundation of China(50764006,50574045)Yunnan Basic Applied Research Foundation of China(2006E0021M)
文摘The activity of FetO is very important in ironmaking and steelmaking process.In order to predict the activ-ity of Fe,O and optimize the operation conditions in ironmaking and steelmaking process,by application of regular so lution model in molten slag systems,FeO-Fe2 O3-SIO2 ternary system,FeO-Fe2 O3-SiO2-CaO and FeO-Fe2 O3-SiO2-NiO quaternary systems have been studied by the chemical equilibrium between H2/H20 gas mixture and liquid slag con tained in solid iron.The values of interaction energy between cations concerning steelmaking slags have been deter-mined by application of ferric-ferrous iron equilibrium and iron-ferric iron equilibrium.And then the activity of Fe,O can be calculated.The results show that the relative error is 3.9%in FeO-Fe2O3-SiO2 system and 18%in FeO-Fe2O3-SiO2 CaO system.The prediction of activities of FetO in the systems are in good agreement with the measure-ments and the regular solution model is valid for predicting the activity of FetO in complex molten slags systems.The activity of Fe,O in FeO-Fe20a-NiO system have not been tested presently,and the calculated result can not be assessed.
基金Supported by Science and Technology Project of Hangzhou City (20110232B17)
文摘To improve forecasting and sustained control level of underground pests, trapping quantity of underground pests （black cutworm,mole cricket and scar-ab） by lamps and their field dynamics in Hangzhou district from 2005 to 2011 were investigated in the paper. The results showed that different pests had obvious differences in population dynamic. The black cutworm （Agrotis ypsilon） had several damage peaks （late May, late June and late July） and the moth amount in early period was relatively high. The mole cricket （ Gryllotalpa africana） had two damage peaks （late May to early July, early September to mid and late October）. The scarab （Anomala corpulenta） had one damage peak （late May to late June）. There were periodic changes in total quantity of underground pests among years, and the peak period appeared in the year of 2005, 2007 to 2009 and 2011, respectively. On this basis, temperature, humidity, rainfall and light were used as forecas- ting factors, using the method of stepwise regression, 19 factors with significant correlation were screened out and prediction models for occurrence quantity and oc- currence period of the three pests were established. By using accuracy degree judge model for verification, the score values of prediction model for occurrence quan-tity and occurrence period of the three underground pests were more than 58 and 70, which indicated that the historical coincident rate and prediction accuracy of estabhshed prediction models were good.
基金Item Sponsored by National Natural Science Foundation of China(50764006,50574045)Yunnan Basic Applied Research Foundation of China(2006E0021M)
文摘Modified quasi-regular solution model has been available based on the sub-regular solution model and quasiregular solution model in this article. The three parameters of this model are set by the computer-aided analysis based on the experimental data of activity at two different temperatures. Seven binary molten slag systems in the whole concentration are calculated by application of the model and the average relative error is within 10%. Ten binary molten slag systems presenting saturation concentration are also calculated, but the average relative error is bigger, especially for the systems containing acidic oxide and neutral oxide. The results show that the calculated results are superior to those calculated by application of regular solution model, sub-regular solution model and quasi-regular solution model.
文摘Theoretical results related to properties of a regularized recursive algorithm for estimation of a high dimensional vector of parameters are presented and proved. The recursive character of the procedure is proposed to overcome the difficulties with high dimension of the observation vector in computation of a statistical regularized estimator. As to deal with high dimension of the vector of unknown parameters, the regularization is introduced by specifying a priori non-negative covariance structure for the vector of estimated parameters. Numerical example with Monte-Carlo simulation for a low-dimensional system as well as the state/parameter estimation in a very high dimensional oceanic model is presented to demonstrate the efficiency of the proposed approach.
基金supported by the National Natural Science Foundation of China (Nos.41374023,41131067,41474019)the National 973 Project of China (No.2013CB733302)+2 种基金the China Postdoctoral Science Foundation (No.2016M602301)the Key Laboratory of Geospace Envi-ronment and Geodesy,Ministry of Education,Wuhan University (No.15-02-08)the State Scholarship Fund from Chinese Scholarship Council (No.201306270014)
文摘The application of Tikhonov regularization method dealing with the ill-conditioned problems in the regional gravity field modeling by Poisson wavelets is studied. In particular, the choices of the regularization matrices as well as the approaches for estimating the regularization parameters are investigated in details. The numerical results show that the regularized solutions derived from the first-order regularization are better than the ones obtained from zero-order regularization. For cross validation, the optimal regularization parameters are estimated from L-curve, variance component estimation（VCE） and minimum standard deviation（MSTD） approach, respectively, and the results show that the derived regularization parameters from different methods are consistent with each other. Together with the firstorder Tikhonov regularization and VCE method, the optimal network of Poisson wavelets is derived, based on which the local gravimetric geoid is computed. The accuracy of the corresponding gravimetric geoid reaches 1.1 cm in Netherlands, which validates the reliability of using Tikhonov regularization method in tackling the ill-conditioned problem for regional gravity field modeling.
基金The National Natural Science Foundation of China(No.60702069)the Research Project of Department of Education of Zhe-jiang Province (No.20060601)+1 种基金the Natural Science Foundation of Zhe-jiang Province (No.Y1080851)Shanghai International Cooperation onRegion of France (No.06SR07109)
文摘In order to decrease the sensitivity of the constant scale parameter, adaptively optimize the scale parameter in the iteration regularization model （IRM） and attain a desirable level of applicability for image denoising, a novel IRM with the adaptive scale parameter is proposed. First, the classic regularization item is modified and the equation of the adaptive scale parameter is deduced. Then, the initial value of the varying scale parameter is obtained by the trend of the number of iterations and the scale parameter sequence vectors. Finally, the novel iterative regularization method is used for image denoising. Numerical experiments show that compared with the IRM with the constant scale parameter, the proposed method with the varying scale parameter can not only reduce the number of iterations when the scale parameter becomes smaller, but also efficiently remove noise when the scale parameter becomes bigger and well preserve the details of images.
