一维对流扩散方程逆过程LS-SVM解  

1-D convection-diffusion equation inverse process based on least square support vector machines method

作　　者：吴自库[1] 许海洋[1] 李福乐[1] 

机构地区：[1]青岛农业大学理学与信息科学学院,青岛266109

出　　处：《黑龙江大学自然科学学报》2016年第4期429-434,共6页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(61403233);山东省自然科学基金资助项目(ZR2009AL012);山东省教育厅科技计划项目(J09LA12)

摘　　要：将最小二乘支持向量机方法,应用于一维对流扩散方程逆过程研究。相对于其它已有的方法,该方法具有简便实用、稳定性好等优点。使用高斯核函数对方程的解进行整体逼近,利用样本点得到回归参数的值。数值结果表明,该方法具有较高的精度和稳定性,可以有效地求解其它反问题。Least square support vector machines method is used to study the inverse process of one-di- mensional convection-diffusion equation. The method has many advantages compared with the other meth- ods, such as simple, practical and good stability. The solution of one-dimensional convection-diffusion e- quation is globally approximated by use of Gauss kernel function. The values of regression parameters are estimated through sample points. This method has been successfully tested on practical examples and has yielded higher accuracy and stability solutions, so it can be used to other inverse problems.

关 键 词：一维对流扩散方程 逆过程 最小二乘支持向量机 近似解 

分 类 号：O175[理学—数学]