遗传算法求解多模态优化问题的动力性

Dynamical Behavior of Genetic Algorithms on Multi-modal Optimization

李航[1,2] 李敏强[1] 寇纪淞[1]

多模态函数一股存在多个局部极值解，局部极值解处适应值的大小很大程度上影响了它们被遗传算法搜索到的概率．为了弄清楚这种影响机制，通过分析基因池遗传算法的无限种群动力系统，刻画了双峰函数局部极值解的适值差与系统不动点之间的解析关系，进一步分析推广了理论结果的适用范围．最后，提出针对多模态优化问题的两阶段遗传算法，给出了应用理论结果改善遗传搜索性能的范例，实验结果表明该算法对多模态函数的搜索性能有明显改善，从侧面证明了理论结果在实际应用中的正确性．[著者文摘]

There is more than one local optimum in multi-modal landscapes. The values of local optima can influence their opportunities to be exploited by GAs. By analyzing the infinite population dynamical system using the gene pool GA and the BINEEDLE fitness function, we characterize the analytic relation between the fall of local optima and the fixed points in the infinite population dynamical system of the gene pool GA. Further analysis shows that the relation is still held in more common cases. Enlightened by the result, we derive a method to improve GAs for the multi-modal landscapes and get satisfying effect. The experiment results also prove the correctness of the theory in this paper.[著者文摘]

社会符号学研究中的多模态化

Multimodalization in Social Semiotics

胡壮麟

索绪尔对语言符号学的建立和对普通符号学的展望有其不可磨灭的贡献，但他的有些观点在符号学研究者中有不同反响。代表英美传统的皮尔斯发展了符号范畴，但他忽视了作为社会成员的人，因此仍不能摆脱形式主义的研究方法。随后的社会符号学发端于韩礼德的“语言是社会符号”的理论，关注的是特定于某一文化某一社团的符号实践。这一理论推动了多模态表达的研究，从而形成了多模态符号学。本文首先讨论多模态符号学和多媒体符号学的区别；进一步介绍具有媒体和模态双重特性的计算机符号学；对一些具体问题分别作了论述，如作为符号系统的符号资源、以屏幕代替书本的多模态符号的设计以及由话语参与者实现的连贯。最后指出，在人类进入社会符号学的多模态化的新世纪，对多模态识读能力的培养应引起重视。[著者文摘]

Saussure＇ s linguistic semiotics has been an immense contribution to the development of general semiotics, but has met challenges from some researchers in this domain. Pierce, representing the British-American tradition has developed the notions of semiotic categories, but is tied to formal approach. Based on Halliday＇ s theory of language as social signs, social semiotics emphasizes the semiotic practices of a particular community of a particular culture, which pushes forward the study of muhimodal presentations, hence the muhimodal semiotics. This paper will first make a distinction between muhimodal semiotics and multimedia semiotics, the former based on the view that all discourses inherently possess the nature of muhimodality; then it goes on discussing the role of computer semotics, which possesses the characteristics of both media and mode. The paper also deals with some relevant problems, such as semiotic resources as semiotic system, the replacement of books with screens in the designing of muhimodal signs, and coherence realized by discourse par- ticipants. Finally, at the time when we are entering into a new era marked by muhimodal social semiotics, the nurturing of muhiliteracy should draw our attention.[著者文摘]

利用多端统一潮流控制器阻尼多模态振荡

Damping Power System Multi-mode Oscillations Using MUPFC

胡晓波 陈中 杜文娟 王海风 万秋兰 唐国庆

研究了多机系统中使用统一潮流控制器（UPFC）抑制电力系统多模态振荡的可行性。首先推导了作为UPFC概念延伸的多端统一潮流控制器（MUPFC）的非线性动态模型和线性化的Phillips—Heffron模型。利用UPFC和MUPFC具有多个控制回路的特点，提出在其多个控制回路上附加多个阻尼控制器来阻尼多模态振荡。通过模态分析，采用可控性指标对系统中的各个弱阻尼模态选取有效的控制回路，并设计了附加阻尼控制器。仿真结果表明，通过该方法确定的在一个UPFC或MUPFC上设置的多个附加阻尼控制器能够成功实现对电力系统多模态振荡的抑制。[著者文摘]

This paper investigates the feasibility of damping multi-mode oscillations using unified power flow controller （UPFC） in a multi-machine power system. Firstly, the nonlinear dynamic model and the linearized Phillips-Heffron model of multiterminal unified power flow controller （MUPFC） are derived, which is the extension of UPFC. Based on the characteristic of having more than one control channels, the means of superposing multi-auxiliary damping controllers on the multi-control loops of UPFC or MUPFC is proposed. By modal analysis, controllability indices are used to choose efficient control loops for weak modes respectively and auxiliary damping controllers are designed. Simulation results indicate that the multi-auxiliary damping controllers of one UPFC or MUPFC designed by the proposed method can achieve the suppression of multi-mode oscillations in power systems successfully.[著者文摘]

弹射座椅多参数多模态控制仿真分析

Multi-parameter and Multi-mode Control Simulation Analyses of Ejection Seat

封文春 林贵平

提出了安全高度影响因子法,比较分析了弹射启动时飞机的几种飞行参数对救生高度的影响。建立了多模态控制仿真模型,确定了多模态控制输入参数,划分了30种弹射工作模式。通过仿真,比较分析了HTY-8多模态控制模型与K36D-3.5A和ACESⅡ座椅在最低安全救生高度上的差异。并比较了国军标（GJB1800-93）所规定的68种弹射条件所需最低安全救生高度。能够实现0高度0速度横倾90°条件下的弹射救生;在倒飞条件下,最低安全救生高度平均降低约40m左右。分析表明,多模态控制能够达到改善中低速低空不利姿态条件下救生性能的目的,而对于速度大于1000km/h的高速弹射条件,则起不到改善的目的。[著者文摘]

The method of safe altitude impact factor was proposed and the infection of the plane parameters at ejection start to safe altitude was put forward. The simulation model of multi-mode control was built and determined the input parameters of multi-mode control and thirty ejection modes were constituted. Through the simulation, the minimum safe altitude of HTY-8 multi-mode was compared with K36D-3.5A, ACES II and sixty-eight ejection conditions in GJB 1800-93. The ejection at the condition of zero altitude, zero velocity and roll 90 degree was successful. At the condition of roll 180 degree, the minimum safe altitude reduced about 40 meter in average. Analyses result shows that the goal of improved escape performance at lower and medium speed and lower altitude by multi-mode control can be obtained. About the condition of speed over 1000 km/h, otherwise, multi-mode control can not attained the goal and need to find other methods.[著者文摘]

多模态广告语篇的互动意义的构建

The Construal of Interactive Meaning in Multimodal Advertising Discourse

王红阳[1,2]

数字化信息时代的到来使语篇世界日趋呈多模态化,应运而生的多模态语篇分析方法为由多种符号组成的语篇的分析提供了新的视角和有效途径。用韩礼德的系统功能语言学基础上发展起来的视觉图像分析语法框架分析一则宣传广告画中的互动意义的构建和体现,验证了系统功能语言学理论对分析带有图像的多模态广告语篇的可行性和可操作性。[著者文摘]

Based on the theoretical framework of the visual communication grammar, which is developed by Kress ＆ van Leeuwen from Halliday＇ s systemic functional linguistics, this paper studies how the interactive meaning is construed in a poster. The aim of this study is to explore the possibilities of analyzing multimodal advertising discourse from a different perspective. The analysis in this paper shows that important interactive and interpersonal meanings can be identified from a multimodal perspective of discourse analysis.[著者文摘]

桥梁多模态耦合抖振控制方法研究

Research on passive control theory for multi-mode coupled buffeting control of bridges

曾宪武 韩大建

目前已证实调谐质量阻尼器（TMD）可以有效控制桥梁抖振响应，并已在工程中得到应用。然而，传统桥梁抖振被动控制理论是基于单模态叠加SRSS法，无法考虑多模态参与作用和模态间气动耦合效应，本文基于Scanlan多模态耦合抖振理论和多重调谐质量阻尼器（MTMD）被动控制理论，提出一种桥梁多模态耦合抖振MTMD控制方法，该方法可以考虑多模态参与作用、模态间气动耦合效应和单模态中各模态位移分量的气动耦合，且对各TMD在主梁上的安装位置没有任何限制。本文最后采用时域仿真方法对该方法进行了验证，两者计算结果吻合良好，表明本文所提出的方法的正确性。[著者文摘]

Tuned mass damper（TMD） is proven to be effective against wind-induced buffeting response and has been installed on some existed bridges. However, the traditional buffeting control method using TMD or MTMD has predominantly been accomplished by single-mode-based method or SRSS method in which the aerodynamic coupling effects among the modes are neglected. Based on multi-mode coupled buffeting theory proposed by Scanlan and passive control theory of multiple tuned mass dampers （MTMD）, a method for muhiTmode coupled buffeting control using multiple tuned mass dampers for bridges is established in this paper. Multi-mode participation and aerodynamic coupling between modes or modal displacement components in a mode are all taken into account in the formulation and each TMD can be installed anywhere on the bridge deck. Finally, time-domain approach is used to verify the proposed method and the results agree well with those of time-domain approach, which shows that the method is correct.[著者文摘]

大跨度桥梁多模态耦合颤振二分法全自动搜索

曾宪武 韩大建

在分析和评价国内外对大跨度桥梁风致颤振研究的成果和目前各种常用颤振分析方法优缺点的基础上，基于结构的固有模态坐标，进一步完善了大跨度桥梁多模态耦合颤振的双参数搜索状态空间法。通过将结构和气流作为一个动态系统，最后将颤振问题归结为数学上一个非对称实矩阵的广义特性值问题。在求解中，风速搜索采用了二分法加速收敛技术，频率采用搜索迭代法，该方法可以按指定的搜索精度，迅速搜索到颤振临界风速和颤振频率，无须设定步长，且能真实地给出各阶模态频率和阻尼比随风速的全过程。最后对具有理想平板截面的悬臂梁进行颤振分析验证了本文方法的正确性，同时对总跨度1177m的香港汀九大桥进行颤振分析，计算得到的颤振临界风速基本与风洞试验结果基本一致，进一步表明本文方法的可靠性和高效性。

高层建筑模态抗震设计中的多模态静力推覆分析

Application of Performance-Based Seismic Design Method for the Seismic Evaluation of Hybrid Structures

夏心红 沈蒲生 龚胡广 方辉

将模态推覆分析方法与用自适应的加栽方式相结合，提出了一种改进的多模态静力推覆分析方法．运用模态推覆分析将多自由度体系解耦成每阶模态对应的单自由度体系，对于每阶模态下的推覆分析，当结构进入塑性阶段后，采用自适应的加载方式，得到每阶模态下结构的反应．将结构推至每阶振型下的目标位移，采用SRSS组合的方式得到结构的响应．对应每阶振型下推覆分析的目标位移采用非弹性能力谱方法来计算．对不同高度的多高层钢框架-混凝土核心筒混合结构在8度罕遇地震下抗震性能进行了评估．结果显示，本文的静力推覆分析所得的结构响应同非线性动力时程分析所得的结果吻合较好．[著者文摘]

An improved multi-mode pushover analysis, which combines the advantage of the modal pushover analysis （MPA） and the adaptive pushover analysis （APA）, has been proposed. In the multi-mode pushover analysis procedure, the multiple degree-of-freedom （MDOF） system can approximately be decomposed into equivalent single degree-of-freedom （ESDOF） system of multi-mode. The response of each modal pushover analysis can be attained by using adaptive load pattern when the structures enter into plasticity. The target displacement of each mode pushover analysis can be obtained by inelastic spectrum. Numerical examples of different tall hybrid structures are calculated to evaluate the accuracy by comparing the structural responses of the proposed analysis procedure with the results of the nonlinear history analysis （NHA）. It shows that the structural responses of the proposed pushover analysis are generally similar to the results from the NHA, especially in the story drift.[著者文摘]

ISO联手IEC发布信息技术多模态生物特征识别技术方案

日前，由ISO和IEC联合公布的《信息技术——生物特征——多模态和其他多生物特征融合》（ISO／IECTR24722：2007）方案能融合多种生物指令，将相同身份的多种证据输入传感器。[第一段]

基于分箱核密度估计的非参数多模态背景模型

A multimodal background model based on binning kernel density estimation

王国良 梁德群 王新年 王彦春

提出了一种用于检测运动目标的非参数多模态背景模型。该模型采用分箱核密度估计算法从训练图像序列中得到背景的密度函数。分箱核密度估计算法利用基于网格数据重心的分箱规则，很好地提取了训练图像序列的关键信息，避免了采用全样本数据点的重复计算，大大提高了运动目标检测算法的实时性。通过与全样本算法进行对比，发现该背景模型在运动目标检测中的有效性，可用于户外的实时交通监控系统。[著者文摘]

A novel nonparametric multimodal background model was proposed to detect moving objects. The binned kernel density estimators were exploited to estimate the probability density function of background intensity in training sequence. Based on the gravity center of the data points, the binned kernel density estimators described the key information of the original whole sample set and avoided the repetition computation in the evaluation phase. Compared with algorithm based on the whole samples, the proposed approach is proved to be efficient in tragic surveillance systems, and it can be used in outdoor environment surveillance systems.[著者文摘]