逼近

A new structure of neural network control system is introduced, its availability to accomplish feedbaek linearization and pole placement in the sense of stochastic approximation is proved, and the accuracy estimate is also discussed.

提出了一种新的神经网络控制系统的结构模型，证明了在随机逼近意义下实现反馈线性化和极点配置的有效性，讨论了逼近精度估计方法；（2）。

Based on Bellman stochastic nonlinear dynamic programming, we develop an optimal control algorithm for the discrete stochastic system with conditional Markov structure.

基于Bellman随机非线性动态规划法，提出了具有条件马尔科夫跳变结构的离散随机系统的最优控制方法，应用随机变结构系统的性质对最优控制算法进行了简化处理，并将后验概率密度函数用条件高斯函数来逼近，针对一类具有条件马尔科夫跳变结构的线性离散随机系统，给出了其逼近最优控制算法。

In this paper we investigate the researching state, the developing situation and the prospect for the summability , the strong summability and their relevant approximating oder , strong approximating oder of Cesdro mean - an important linear summation process of the Fourier - Laplace series of functions on sphere.

本文研究了球面上的函数用其Fourier—Laplace级数的一种重要的线性求和法-Cesaro平均来求和与强求和及其相应的逼近与强逼近的阶的研究状况、发展态势和前景展望等。

Taylor formula for the basic idea is to use polynomials to approximate a known function, and the coefficients of the polynomial by the given function to determine the derivative , and this approach gives an effective estimate.

泰勒公式的基本思想是用多项式来逼近一个已知函数，而这个多项式的系数由给定函数的各阶导数确定，并对这种逼近给出有效地估计。

Considering that the NURBS has become a standard in CAD/CAGD, we shall investigate the approximation to the solutions of the minimal surface equation with the tensor Bezier surface and B-spline surface.

首先我们讨论极小曲面的分片线性的逼近，由于NURBS已经成为形状描述的唯一工业标准，而且极小曲面本身至少应该是二阶光滑的，因此我们讨论了极小曲面方程的Bezier和B样条逼近。

Structure-Based Approximation Learning Theory: Under the description of manifold and topological architecture of neural field, we have proposed a learning frameworks of neural networks based on global architecture approximation theory borrowed from the concept of topology and manifold to solve the optimal problems of learning in non-nonlinear and non-Euclidean space, and discussed a few basic components of the learning theory, learning mechanism and learning algorithm in the term of structure-based approximation and global structure analysis.

基于结构的逼近学习理论：通过人工神经网络模型神经场的流形结构和拓扑结构描述，我们提出了在非线性和非欧氏空间中基于整体结构逼近的神经场学习理论框架；用对偶流形理论和拓扑概念作为工具，讨论了神经网络的学习理论、学习机理、学习算法和算法的具体实现。

According to the typical fuzzy concepts,such as eutrophication level of water environmental quality and the unideal and ideal water quality stations,a technique for order preference by similarity to ideal solution based on fuzzy set was presented and was used in the comprehensive assessment of water environmental quality.

基于传统的逼近于理想解的技术，根据水环境质量的富营养化程度和待评价的水环境质量检测点的不理想与理想这些典型的模糊概念，提出了基于模糊集合的逼近理想解法，并应用于水环境质量综合评价。

In either case ofg∈C2 or gnot∈C2, this paper proposes a convenient method for building a T-S fuzzy system and gives a loose sufficient condition of its approximation ability, when the fuzzy sets are uniformly distributed, continuous and normalized.

本文在逼近函数g∈C2及g不∈C2两种情况下给出了一种当模糊集满足均匀性、连续性、正规性条件下T-S模糊系统的构造方法以及该T-S模糊系统逼近能力的充分条件。

The TST was applied to multi-scaling vector to increase it's approximation order, and with PR condition, the dual of transformed multi-scaling vector was obtained. Then the special TST was taken for this dual, the dual of original multi-scaling vector with higher approximation order obtained, finally the multi-wavelet function with higher vanishing moment maybe constructed.

我们先对已知的尺度向量进行双尺度相似变换，以提高其逼近阶，再求出变换后的尺度向量的对偶，对这个对偶再进行特定的双尺度相似变换，就可得到原尺度向量的有高逼近阶的对偶，从而可能得到具有高阶消失矩的多小波函数。

This thesis focuses on studying the matrix equa-tion problem systematically, and proposed an abstract algorithm of solving the matrixequation with constraints, and established a strict convergence theory. Using this algo-rithm, we can solve the sets of matrix equation satisfying some constraint conditions,such as symmetric, antisymmetric, centrosymmetric, centroskew symmetric, re?exive,antire?exive, bisymmetric, symmetric and antipersymmetric, symmetric orthogonalsymmetric, symmetric orthogonal antisymmetric, Hermite generalized Hamilton ma-trix;So we can solve the problem with this algorithm, if the set of constrain matrixcan make a subspace in matrix space, and this algorithm also can solve the optimalapproximation and least squares problem. So this abstract algorithm has universal andimportant practical value.

本篇硕士论文系统地研究了此类问题，并找到了求解约束矩阵问题的抽象算法，并建立严格的收敛性理论，利用这一算法可求解约束条件为对称矩阵、反对称矩阵、中心对称矩阵、中心反对称矩阵、自反矩阵、反自反矩阵，对称正交对称矩阵、对称正交反对称矩阵、双中心矩阵、Hermite广义Hamilton矩阵等；可以说只要约束矩阵集合在矩阵空间中构成子空间，都可以考虑用此算法求解，而且这一算法还能把矩阵方程解及其最佳逼近，最小二乘解及其最佳逼近统一处理，因此本文算法有普适性和重要的实用价值。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。