本文引入几类向量f -互补问题，并在一定条件下给出了向量f -互补问题，广义向量变分不等式以及可行集的最小元问题之间的相互关系。
The relations of the vector F-complementarity problems, general vector variational inequalities, and least element problems of feasible set are given under certain conditions.
以二元件均衡器为链接模型，以非线性最小二乘法为最优化方式，解决了多节均衡器链接中初始参数值的选取问题。
Taking two-element equalizer as a link model, taking nonlinear least square method as optimum mode, selection problem of initial parameters in multiple node equalizer links is solved.
这种方法能够保证辨识出的参数是最佳的；而且不用求解对应的非线性最小二乘问题，只需求一元多项式的根，从而大大减少计算量。
This approach can guarantee that the identified parameters are optimal, by solving a one-dimensional polynomial equation instead of a nonlinear least square problem.
新算法选择很广一类的隐层神经元函数，可以直接求得全局最小点，不存在BP算法的局部极小、收敛速度慢等问题。
The algorithm can get global minimum easily with a wide variety of functions of hidden neurons, and no problems such as local minima and slow rate of convergence are suffered like BP algorithm.
讨论了四元数量子力学中带有谱约束的最小二乘解反问题，得到了此问题有解的充分条件。
The inverse least square problem with spectral constraints is also studied and a sufficient condition is given for its solvability.
研究了无穷边界的二阶椭圆问题，探讨了将其转化为一阶系统的最小二乘有限元逼近方法。
A least-squares mixed finite element methods for the first-order system formulation of second-order elliptic problems and a formulation for the infinite boundary conditions are presented.
该方法首先通过在加权最小二乘 支持向量机的基础上加入对数据偏斜的处理，解决了元 信息分类时关键词特征稀疏和样本高度不均衡问题；
Since the feature of the meta-information classification keywords is sparse and the distributing of sample is unbalanced, this thesis considered the factor of data skew based on LS-VSM.
该方法首先通过在加权最小二乘 支持向量机的基础上加入对数据偏斜的处理，解决了元 信息分类时关键词特征稀疏和样本高度不均衡问题；
Since the feature of the meta-information classification keywords is sparse and the distributing of sample is unbalanced, this thesis considered the factor of data skew based on LS-VSM.
应用推荐