算法

According to disadvantage of AS that the speed low of convergentiss and easily entering the stagnation,improving the AS in some aspect by the way that importing pheromone windows that the value of pheromone cannot exceed the window,only deposit the pheromone on the edge that belongs to the iteration shortest solution path,releasing additional pheromone to the edge belongs to the so-far shortest solution path,judging the situation of convergent and to reinitialize pheromone,optimizing the solutions by local search procedure at the end of every iteration,modifying the transition probability by add some problem specific parameters and so on.

本文主要运用蚁群优化算法(Ant Colony Optimization，ACO)来求解VRP。针对蚁群算法收敛性比较差，易于停滞的缺陷，通过使用信息素窗口限制、信息素的最大最小值，只更新信息素迭代中最好解，增加了精英的蚂蚁，信息素重新初始化之前先判断出汇聚情况，在每次迭代过程中加入局部的搜索进行优化，在对蚁群算法进行了优化时考虑选择概率当中加入与问题的相关参数等措施，从而使蚁群算法的收敛性得到大大提高，避免了算法的停滞现象。

On global optimal algorithms, a heuristic 0-1 Integer Liner Programming algorithm is achieved, which combines heuristic enumeration method and revised simplex method. Simulation experiments show that the degree and optimization efficiency in our algorithm is obviously improved compared with the MATLAB 7 bintprog function that is a 0-1 integer programming algorithm. Moreover our algorithm isn\'t relevant to the scale changes of Web services composition and candidate service sets, thus it is applicable for the scenes that the scales of Web services composition changed dynamically. But for Web services composition with service alliance, the algorithm is unable because the non-linear global optimal model of the fitness function and QoS constraints can\'t be expressed in linear mode.

在全局优化算法研究上，基于启发式0-1整线性规划算法结合了修正单纯形法和启发式枚举法，仿真实验表明该算法与MATLAB 7的0-1整数规划方法比较，在优化程度及优化效率上得到了显著改进，并且对服务组合和候选服务集合的规模变化没有明显的相关性，因而该算法可用于服务组合规模动态变化的场景，但该算法不能应用于存在服务联盟的非线性全局优化模型。

Riemannian Space. The parameter space in blind source separation is Riemannian Space. Natural gradient has many better characters compared with normal gradient.

自适应盲分离算法中的系数空间是黎曼空间，本文将自然梯度下降法应用于Infomax算法可以改进算法的性能，有利于算法的数值稳定性和提高算法的收敛速度。

An improved scheduling method based on exponential division of deadline is presented after study on the problem of prevenient average division of deadline method in this paper.The paper also proves when the relative deadline of each node is too large,the schedular performance of messages transmission task using average division of deadline will be decreased by introducing notion of quantization cost.The new method solves the problem well,and digital simulation validate the new method.

摘 要：针对平均分区的EDF算法在CAN总线中的应用出现的问题，本文提出了一种改进的基于指数分区的EDF算法，并通过引入量化误差的概念，推导证明当CAN网络中各节点相对截至期分布时间过大时，平均分区的EDF算法会导致CAN总线中信息传输任务可调度性的下降，而基于指数分区的EDF算法保证了信息传输任务的实时性，仿真试验验证了算法的有效性。

Firstly, the paper expatiates the essence principle of the motion estimation based block matching, introduces Full Search method and several typical motion estimation fast algorithms, and analyzes each characteristic, then to sum up the key technology for increasing the motion estimation search efficiency; Secondly, it stress introduces three motion estimation fast improved algorithms: the improved algorithm based Full Search method, it imports successive elimination algorithm、subsample and termination condition to increase search speed for enhancing the real time of Full Search.

本文首先阐述了基于块匹配法的运动估计基本原理，介绍了全搜索法和几种典型的运动估计快速算法，分析了各算法的特点，在此基础上总结了提高运动估计搜索效率的几个关键技术。接着重点介绍了本文提出的三种运动估计快速改进算法。基于全搜索的改进算法，是在全搜索的基础上尝试引入了连续排除算法、子抽样和终止条件来提高搜索速度，增强全搜索的实时性。

The constructing process of the algorithms itself indicate that they are brevity and have definite geometric meaning, scientific iterative terminal criterions and no strict requirement to function. Finally, some digital experimentation is presented to indicate the algorithms application in solving the problem of maximum of function, transcendental equation,integer programming and so on. The results indicate that the algorithms have more global and more adaptability compared to traditional algorithms and have more calculating efficiency and more scientific iterative terminal criterions compared to some present popular artificial intelligence algorithms.

算法的构造过程本身表明了该算法简单明了、几何意义明确、迭代终止准则科学、对函数的要求较低；最后举了几个数值试验的例子以说明该类算法在求函数的最值，解超越方程以及整数规划等问题中的应用；数值试验的结果标明该算法比单纯由"微分"思想演化而来的传统算法有更好的全局性和对问题的更广泛的适应性，比现在流行的一些人工智能方法有更高的运算效率和更适当的迭代中止准则。

In order to make the new recursive method convergence fast, the RLS-like forgetting factor is introduced to construct the new data Hankel matrices. In the framework of subspace methods, the gradient type subspace tracking method is used to implement the update of the state subspace, which results the unbiased estimate. Different from traditional offline method, the system matrices are recursively computed using RLS method. The convergence analysis of the proposed method is given and finally, the efficiency of this method is illustrated with a simulation example.A new closed-loop identification method is proposed to implement consistency estimation of closed-loop subspace identification.

引入RLS-like遗忘因子机制构造新的数据矩阵，更好地跟踪系统时变信息，提高递推算法的收敛速度；在子空间辨识框架内，采用梯度型子空间跟踪算法实现对状态子空间的更新，避免了子空间近似带来的估计有偏性；与传统的离线算法不同，在得到系统状态子空间后，基于递推最小二乘方法提出新的算法，实现对系统矩阵的递推估计；并给出递推子空间辨识算法的收敛性分析；最后通过仿真实例验证算法的有效性。

Efficient detector generation algorithm is the kernel of anomaly detection. Aiming at low true positive value, unhandy matching threshold value and large detector set size of existent algorithms, a novel detector generation algorithm based on multiple populations genetic algorithm is put forward in this paper. According to morphologic analysis of intrusion detection system and covering problem principle, self set is divided into several partitions on the basis of their characters.

摘要有效的检测器生成算法是异常检测的核心问题，针对现有算法存在检测率低、匹配阈值固定、检测器集合庞大等问题，本文提出了基于多种群遗传算法的检测器生成算法，根据形态学空间的分析和覆盖问题原理，自体集根据特征进行划分，各个种群根据划分独立按遗传算法进化，最后求得所有检测器种群的并集得到成熟的检测器。

Most of community detection algorithms are used in dealing with unipartite networks, and there are few algorithms designed for detecting communities in bipartite networks, and these algorithms have some weaknesses respectively. For example, the community detection algorithm proposed by Guimera can not utilize weight information of a bipartite network; the algorithm proposed by Barber can not deal with the complex mapping situation of communities in bipartite network.

目前社团检测算法大多用于处理单元网络，而专门用于处理二分网络的社团检测算法不多，而且这些算法都存在着各自的不足之处，如 Guimera 所提的二分图社团检测算法无法处理边含权重信息的二分网络； Barber 所提的社团检测算法无法处理二分网络中具有复杂社团对应关系的情况。

Polygon Filling Algorithm typical scan line mainly include filling algorithm and contour filling algorithm domain signs for the document vector polygon filling algorithm for the scan line fill algorithm.

详细说明：典型的多边形填充算法主要包括扫描线填充算法和轮廓标志域填充算法，适用于矢量多边形文件的填充算法为扫描线填充算法。

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。