混合的问题

Two strategies, optimizing hill- climbing and large- scale gene mutation, are used to overcome the disadvantages of the primary hybrid algorithm, which easily converg to local optima.

用遗传与模拟退火相结合的混合算法对信道分配问题进行研究，并通过加入"寻优式爬山"与大规模基因突变两种优化方法对混合算法进行改进，克服了一般遗传算法收敛速度慢以及易于陷入局部最优解的缺点。

Many authors posed and discussed some boundary value problems,mainly the Dirichlet problem and mixed boundary value problem for second order elliptic equations with parabolic degeneracy.

许多作者提出和讨论了二阶退化椭圆型方程的一些边值问题，如Dirichlet边值问题和混合边值问题。

Three solution algorithms for classical transportation continuous/discrete/ mixed network design problems are designed based on particle swarm optimization approach, in which the continuous NDP is studied in detail, discrete and mixed NDP are only described the frameworks of the solution algorithms.2. Generally, road capacity constraints must be added to the lower level problem of classical discrete NDPs bilevel programming model in order to avoid traffic assignment on candidate roads.

本论文研究工作主要有如下几个方面：(1)基于粒子群优化思想设计了求解传统的城市道路交通网络设计问题的算法，其中详细研究了连续网络设计问题，并用两个网络例子进行了数值实验；并给出了离散网络设计问题和混合网络设计问题的求解算法思路和主要步骤。

In the second part, we studied blind source separation in ill conditioned cases, obtained the sufficient and necessary conditions on solvability, and presented the neural network model and corresponding adaptive algorithm. We also discussed the blind digital source separation, established the blind separation and extration models and algorithms. These algorithms can be used to deal with high noise and ill-conditioned cases.

在第二部分，我们研究了奇异混合情形下信号盲分离问题，给出了神经网络模型及相应的自适应算法；我们也研究了数字信号盲分离问题，得到了有关可解性的充要条件，提出了分别处理低噪声问题和高噪声问题的算法，这些算法对感知信号个数少于源信号个数等病态问题也适用。

From the distinguishing the issue of law and fact,the ChapterⅥ educes the mixedissue interposing between the issue of law and the issue of fact,that is,rule-applicationissue that applies the legal provisions to specific facts.

第五章基于对尚未引起我国学者足够重视的法律问题与事实问题区分的探讨，引出了介于法律问题与事实问题二者之间的混合问题，即将法律适用于特定事实的涵摄问题。

In this thesis, the barotropic primitive equation under the rectangular coordinate system and atmospheric circulation equation under p coordinate system are.

从正压原始方程开始，运用分层理论研究大气环流方程组定解问题的适定性，以及适定的定解问题(包括初值问题、边值问题和混合问题)在解析函数类空间准确解的求解方法。

Based on stratification theory, begin with the barotropic primitive equation, some theoretical problems of atmospheric equations are discussed, including the well posedness of their initial value problem, boundary value problem and mixed problem.

摘要从正压原始方程开始，运用分层理论研究大气环流方程组定解问题的适定性，以及适定的定解问题(包括初值问题、边值问题和混合问题)在解析函数类空间准确解的求解方法。

Based on the proposed ORPF models, the quadratic optimal models of minimizing the total load curtailment are established to identify power flow unsolvability for AC and AC-DC power systems, and can be further used to identify system voltage instability with the modal analysis method.

算例表明：当离散内点法不存在收敛问题时，启发式混合离散内点法能快速地获取优化解，反之，则要使用启发式混合随机优化算法来保证解的质量，但需要牺牲部分计算速度。

To eliminate the limit, a hybrid objective function weighted by antitangent function is constructed, which behaves like an objective function of norm for small data points and like an objective function of norm for large data points, restraining the influence of false data points on inversion procedure.

为了解决这一问题，利用反正切函数作为加权因子构造了反正切混合目标函数，并利用模型试验进行了验证。简要阐述了最优化目标函数的构造方法；给出了利用反正切函数构造混合目标函数的基本原理，并对目标函数的微分特征和概率特征进行了分析；利用模型论述了反正切混合目标函数中控制参数μ和σ的意义，μ控制反正切函数变化的速度，σ控制数据误差项被处理或加权的方式。

Then, the relation between vibration and sound, the acoustic holography method, the indeterminacy and identifiability of BSS, the BSS algorithm are discussed in detail. Based on above researches, the dissertation is divided in following four sections. The first section investigates the BSS algorithm suitable to acoustic feature separation. The algorithm of the joint approximate diagonalization of eigen-matrices is proposed. The spectra or the time-frequency distributions of source signals are the interesting features in ABD and they are separated from mixing signals by JADE algorithm. Then, the convolutive mixing model is transformed into a high-dimension instantaneous mixing model, and the deconvolution of source signals is achieved by the joint approximate block diagonalization of eigen-matrices. The proposed algorithm has a global minimum, and it is unsensitive to noise interference.

论文首先概述国内外声学诊断研究进展与盲分离研究进展，给出机器噪声声场与盲分离的数学描述，讨论了声振辐射、声全息、盲分离模型、盲分离可解性、盲分离结果不确定性、分离算法等基本问题，在此基础上，论文的研究工作分为以下四个部分：第一部分研究适用于声学特征分离的盲分离算法，提出基于特征提取的联合近似对角化盲分离算法，该算法以频谱特征或时频特征作为分离目标，从混合信号中分离源信号频谱特征或时频特征，最大限度地保留了与声学特征提取有关的频谱特征或时频特征，采用模型变换把卷积混合模型变换为一个高维瞬时混合模型，通过联合近似分块对角化算法实现源信号频谱特征与时频特征的盲反卷积。

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。