局部网络

In order to overcome the problems of slow rate of convergence and falling easily into local minimum in BP algorithm, a new improved homotopy BP algorithm is put forward. It transfers the identity matrix of output mode in common use into the binary system.

为克服标准BP算法中存在的网络学习收敛速度慢、容易陷入局部极小等问题，引入同伦算法，提出了一种将同伦论与快速BP算法结合的改进算法，并采用二进制结构表示输出模式矩阵。

To address the defects of slow convergence and easy immergence in the partial minimum in the learning process of Back Propagation Neural Network CNN, a Levenberg-Marquardt algorithm was presented.

针对BP神经网络学习过程收鼓速度慢及易陷入局部极小值的缺陷，研究了levenberg-marquart算法。

Based on the segmented unstructured model of Nosiheptide fermentation process, the secondary variables are selected according to the implicit function existence theorem. The on-line identification of fermentation phases is accomplished by using an indicator variable which is gained by mathematical inference, and for each phase, a local soft sensor model is developed.

首先以分阶段的诺西肽发酵过程非结构模型为基础，根据隐函数存在定理进行辅助变量的合理选择；然后利用经数学推导得到的指示变量"伪比生长率"完成发酵阶段的在线辨识，并采用神经网络构建出对应于各阶段的局部软测量模型。

This paper discusses the nonlinear inversion algorithms instructively andmainlystudytwokinds ofnonlinearinversionalgorithms: SimulatedAnnealingConjugate Grads and Artificial Neural Network Genetic Algorithms, these two methods can decrease the iteration times and avoiddroppinginto the local extremum, we do the model calculation and confirm thevalidityofthealgorithm.

对非线性反演算法进行了有益的探讨，主要研究了两种非线性混合优化反演算法：模拟退火共轭梯度混合反演和神经网络遗传反演算法，两种混合算法都达到了减少迭代次数、避免陷入局部极值的目的，并进行了模型试算，验证了算法的有效性。

This paper applies GA-BP optimal algorithm to solve the problem of coordinate inversion in Robot's control.

但BP网络易形成局部极小值而得不到整体最优，缺乏全局搜索性。

First, this method calculates the iteration step size with the average value of connectivity ,then it optimizes the local cost functions.

该算法首先根据网络的平均连通度确定迭代步长，然后对每个未知节点的局部代价函数进行优化求解。

It is known that three-layer BP network can approximate any nonlinear function defined in L〓.

但神经网络的学习算法存在着局部极小点和收敛速度慢等缺点。

The global MLP neural network with 3 layers and the local linear polynomial map with first order are used to model the reverberation sequence.

本文分别用三层MLP神经网络和线性一阶映射多项式对混响进行全局非线性建模和局部线性建模及预测性能分析。

The neural networks correction method combining the local map with radial based function networks and the global map with linear transform was applied to calculate the deformation resistance of the strip and the friction coefficient in the rolling deforming region.

采用径向基函数的局部映射和全局线性映射相结合的神经网络校正模型求解带钢变形抗力和轧制变形区的摩擦因数；并采用轧制变形区离散化方法分析轧制变形区内张力、摩擦力及金属变形抗力等在带钢轧制方向上的分布规律，从而建立轧制力在线计算数学模型。

A A device that handles the switching of voice and data in a local area network.

的装置把手开关的语音和数据，在局部区域网络。

On the other hand, the more important thing is because the urban housing is a kind of heterogeneity products.

另一方面，更重要的是由于城市住房是一种异质性产品。

Climate histogram is the fall that collects place measure calm value, cent serves as cross axle for a few equal interval, the area that the frequency that the value appears according to place is accumulated and becomes will be determined inside each interval, discharge the graph that rise with post, also be called histogram.

气候直方图是将所收集的降水量测定值，分为几个相等的区间作为横轴，并将各区间内所测定值依所出现的次数累积而成的面积，用柱子排起来的图形，也叫做柱状图。