算法的

Thirdly, the principal curve and polygonal line algorithm are introduced, and an implementation of polygonal line algorithm based on Quasi-Newton numerical optimization is illustrated. The simulation results show that when this algorithm isused in the detection of curvilinear defects on wafer map, it can converge fast, compute accurately and meet the demand of application of defects pattern analysis, and with the number of line segment increasing, the polygonal line will approximate to the Kegl principal curve.

研究了主曲线、多边形线算法，提出用多边形算法识别弧状缺陷想法，讨论了用拟牛顿法实现该算法的解决方案，针对硅片缺点模式分析中常出现的弧状缺陷，采用该算法进行检测分析，实验结果表明这种算法快速准确，当线段个数或最小距离平方期望值达到一定门限值后，多边形线就已经能够描述弧状缺点集的形状，随着线段个数增加，最小距离平方期望值还在不断下降，最终会收敛到Kegl主曲线。

It is difficult to insure globe convergence for regular BSS algorithms because they are easily trapped into local optimum, which can result in error of separation. This research associates genetic algorithm with traditional BSS and uses global optimize feature of genetic algorithm.

针对多数盲分离算法在多峰值情况下易陷入局部最优，难以保证全局收敛而导致分离误差的缺陷，文章将遗传算法与传统盲信号分离算法相结合，把全局收敛性能较好的遗传算法引入信号分离算法的寻优过程。

First, the traffic flow time series chaotic feature is extracted by chaos theory. pretreatment for traffic flow time series, and the wavelet neural networks model was build by this. Second, the chaotic mechanism and the chaotic probability is described. Based on chaotic learning algorithm, and the wavelet neural networks fast learning algorithm of traffic flow time series is designed based on chaotic algorithm. Last, a single-step and multi-step prediction of traffic flow chaotic time series is researched by BP neural networks, wavelet neural networks and wavelet neural networks based on chaotic algorithm. The results showed that the wavelet neural networks predictive performance is better than the BP networks and the wavelet neural networks by the simulation results and root-mean-square value.

首先，通过混沌理论提取了交通流量时间序列的混沌特征，并在此基础上建立了小波神经网络交通流量时间序列模型；接着，阐述了混沌学习算法的混沌机理、混沌产生的概率，设计了基于混沌算法的小波神经网络交通流量混沌时间序列快速学习算法；最后利用交通流量混沌时间序列对BP网络、非混沌算法的小波神经网络以及基于混沌算法的小波神经网络进行了单步预测和多步预测，并对预测结果的仿真图和真实值与预测值的方均根进行了比较，结果表明基于混沌学习算法的小波神经网络的预测性能明显优于应用BP网络和非混沌算法的小波神经网络。

We also introduce a control function, prove the global convergence, and prove the local superlinear convergence under some conditions. Then we propose another new algorithm by using an existed approximation functions, which has the same characters as the first algorithms does. Finally, the result of the numerical experiments indicates the efficiency of the first algorithm. The paper contains four parts.

本文还引入了新的控制函数，并证明了算法具有全局收敛性和在一定的条件下具有局部超线性收敛性；然后，在第一个算法的基础上，本文利用已有的光滑逼近函数提出了另外一个新的算法，通过适当参数选取，证明了新的算法具有与第一个算法同样良好的收敛性质；最后，通过数值计算说明了算法的高效性。

Premature convergence and weak local optimization are two key problems existing in the conventional genetic algorithm. To overcome the shortcomings, an improved genetic algorithm based on the particle swarm algorithm is proposed.

针对传统遗传算法中存在的易陷入局部最优解和后期收敛速度慢的问题，基于粒子群算法，对传统遗传算法作了改进，提出了一种基于粒子群算法的遗传算法。

Focusing on some disadvantages in standard BP algorithm, such as low convergence rate, easily falling into local minimum point and weak global search capability, Genetic algorithm is used to optimal the connection weight of BP Algorithm in this paper, and construct a GA-BP algorithm of evolutionary neural network, and the algorithm is applied to the control of BLDCM speed adjusting system.

针对标准BP算法存在全局搜索能力弱和易陷入局部极小点等缺点，本文将遗传算法与BP神经网络相结合，构造了一种新的进化神经网络GA-BP算法，并将该算法应用于无刷直流电机调速系统的控制，仿真结果表明，与传统的PI控制系统相比，该算法得出的电机控制曲线几乎无超调，与基于BP算法的速度控制系统相比较，具有收敛速度快、不易陷入局部极小的优点。

The method solves the limitation of converging to the local infinitesimal point in medical image segmentation,and adopts the initial algorithm to assure the initial searching scope of genetic algorithm which is better accommodable than standard genetic algorithm with fuzzy C-means clustering,speeding up the convergence of genetic algorithm.

该算法除了解决模糊C均值聚类算法在医学图像分割中容易陷入局部最优解的问题，而且采用的初值化算法比标准的遗传模糊C均值聚类算法能确定更合适的遗传算法的初始搜索范围，从而加速了遗传算法的收敛过程。

A Multi-Species Cooperative PSO algorithm is developed using the ideas of species dividing in the bionomical world. The algorithm is shown to have guaranteed convergence to the global optimum. Furthermore, the training strategies of RBF network structures and weights based on the MCPSO aigorithm is proposed. Simulation results show that the algorithm is effective in such fields as chaotic time-sequence forecasting, system identification, speech signal processing and etc.

在分析基本微粒群算法和协同微粒群算法特点的基础上，借鉴生态学中的种群划分思想，设计出一种多种群协同进化微粒群算法，并分析了该算法的收敛性能；在分析径向基函数神经网络结构的基础上，提出了基于多种群协同进化微粒群算法的径向基函数神经网络结构和权值优化设计策略，并在混沌序列预测、系统辨识、语音信号处理等应用仿真实验中验证了算法的有效性。

An explicit definition of local peak is shown, and the existence condition of local peak is given. It is proved that the GA is constringent at the neighbor of its local peak. The theoretical evidences of some improvements on GA are given, which points out the way forward for improving GAs performance.

关于遗传算法的收敛性研究的一些典型结果包括：文献 [2 ]分析了在种群规模无穷的情况下典型遗传算法的收敛性；文献 [3]证明了典型遗传算法不收敛，而如果对算法采取记录每一代中最佳个体的策略，则改进的算法收敛；文献 [4]提出了一种等价的遗传算法，并给出了收敛条件与收敛速度；文献 [5 ]研

Aiming at a typical complex electromechanical system—temper mill, the study takes its main driving control system as object of study, works out such a difficult problem that many parameters are coupled mutually, hard to be set and configurated in complex electromechanical system, and makes the performance goal of the control system be optimal in the adjustable ranges of parameters.

分析了遗传算法和模拟退火算法寻优性能和各自的不足与局限性，将两种算法巧妙地结合起来加以修正、取长补短构成新的退火遗传算法，提高了算法的整体寻优性能；并以典型的复杂机电系统—平整机为研究对象，以其主传动速度控制系统为切入点，应用退火遗传算法，解决了复杂机电系统中多参数、相互耦合、难于整定和组态的难题，使控制系统的性能指标在参数可调范围内达到最优。

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。