学分制下排课的数据组织与算法

高校实施学分制过程中,学生先选课再排课,排课方案搜索变得相当复杂.传统的排课算法常采用字符串比较的方式,因此排课效率较低.以二进制位为单元的算法和数据结构,用位的表示方式构建出数据表,既缩短了数据量又提高了运算速度.通过在MS-SQL上建立存储过程并编写用于实现学校排课的应用程序,能大大提高排课效率,同时也使数据的交换量仅为原来的1/8.通过在拥有两万多学生的高校进行排课实测,结果显示该算法简练、逻辑清晰、快速高效.     

竞争PSO在多序列联配中的应用

本文提出了一种求解多序列联配的竞争粒子群优化算法,算法根据适应值分类设计了粒子群的惯性权重以及其飞行速度范围,并进行了动态调整,提高了算法的收敛速度和精度;引入了重新初始化机制,有效的避免粒子群优化算法可能出现的早熟现象;提出了一种全新的速度更新模式和竞争策略,增强了算法的能力。实验表明该算法是有效的。     

基于混合遗传算法的网络拓扑设计

提出了一种由启发式算法和遗传算法混合使用的混合遗传算法用于通信网络中的骨干网拓扑设计。文中骨干网拓扑设计问题是在满足R边连通和跳数约束的情况下使得网络费用最小。在遗传算法中,交叉和变异操作会产生不可行解,可通过增加链路来使不可行解变为可行解。增加链路后,其费用一般要比父代个体大,并且有多余的链路。该文的混合遗传算法是在遗传算法中加入启发式策略,来消除多余的链路,降低子代的费用,加快算法的收敛速度。仿真结果验证了算法的有效性。     

Hidden Information Revealed Using the Orthogonal System of Nucleic Acids

In this study, the organization of genetic information in nucleic acids is defined using a novel orthogonal representation. Clearly defined base pairing in DNA allows the linear base chain and sequence to be mathematically transformed into an orthogonal representation where the G–C and A–T pairs are displayed in different planes that are perpendicular to each other. This form of base allocation enables the evaluation of any nucleic acid and predicts the likelihood of a particular region to form non-canonical motifs. The G4Hunter algorithm is currently a popular method of identifying G-quadruplex forming sequences in nucleic acids, and offers promising scores despite its lack of a substantial rational basis. The orthogonal representation described here is an effort to address this incongruity. In addition, the orthogonal display facilitates the search for other sequences that are capable of adopting non-canonical motifs, such as direct and palindromic repeats. The technique can also be used for various RNAs, including any aptamers. This powerful tool based on an orthogonal system offers considerable potential for a wide range of applications.     

一种新的全姿态指示画面填充算法

现代飞机座舱主飞行仪显示系统中,全姿态指示天地区域的填充是件非常费时的工作。提出了一种新的区域填充算法。该算法第一步引入分块查找的思想以及基于填充区域的特点确定采用水平线或者竖直线扫描方式,并且对区域轮廓中满足关键码字段的像素点建立坐标数组;第二步根据已确定的扫描方式进行区域填充。此外,求取新旧地平线交点并没有采用通过大量数值计算直接求取的方法,而是采用了比较法。实验结果表明该算法显著提高了填充速度和显示效果,降低了数据存储空间量。     

传感器网络 BRA 路由算法的设计与实现

为了提高数据传输的可靠性以适合某些特殊应用,设计并实现了传感器网络备用路径可靠路由算法(BRA)。该路由算法通过建立备用路径,并在主路径失效的情况下采用备用路径传输数据的方法来提高数据传输的可靠性。仿真测试结果表明,文中所提出的备用路径可靠算法在主干路径不可用的情况下,可以有效减小路径断裂引起的丢包现象,进而提高网络整体数据传输的可靠性。     

基于模糊算法的移动机器人行为融合理论研究

机器人体系结构是机器人的基础。本文采用预先设计好的四种机器人基本行为设计了基于功能分解的机器人体系结构,在此基础上指出了此结构的缺点,提出了新的基于模糊算法的行为融合理论。最后通过机器人专用仿真软件进行仿真,结果表明模糊算法用于机器人的体系结构中能有效改善机器人的性能。     

A novel Volt-VAR Optimization engine for smart distribution networks utilizing Vehicle to Grid dispatch

In recent years, Smart Grid technologies such as Advanced Metering, Pervasive Control, Automation and Distribution Management have created numerous control and optimization opportunities and challenges for smart distribution networks. Availability of Co-Gen loads and/or Electric Vehicles (EVs) enable these technologies to inject reactive power into the grid by changing their inverter’s operating mode without considerable impact on their active power operation. This feature has created considerable opportunity for distribution network planners to explore if EVs could be used in the distribution network as reliable VAR suppliers. It may be possible for network operators to employ some EVs as VAR suppliers for future distribution grids. This paper proposes an innovative Smart Grid-based Volt-VAR Optimization (VVO) engine, capable of minimizing system power loss cost as well as the operating cost of switched Capacitor Banks, while optimizing the system voltage using an improved Genetic Algorithm (GA) with two levels of mutation and two levels of crossover. The paper studies the impact of EVs with different charging and penetration levels on VVO in different operating scenarios. Furthermore, the paper demonstrates how a typical VVO engine could benefit from V2G’s reactive power support. In order to assess V2G impacts on VVO and test the applicability of the proposed VVO, revised IEEE-123 Node Test Feeder in presence of various load types is used as case study.     

基于神经网络的机械结构系统优化问题的分解算法

基于前馈神经网络的权重分析，针对优化问题提出了一种分解算法，探讨了神经网络在桁架结构系统优化中的应用。实例分析表明此方法是可行和有效的。     

神经网络技术在摩擦中的应用

摩擦系数是一个很复杂的参数，它受到多种因素的影响和制约。因此，对摩擦系数进行神经网络分析，采用神经网络优化算法，建立以轧件温度、粘度、轧制速度为输入，以摩擦系数为输出的神经网络结构并对其进行神经网络预报。