PSO-SVM模型的构建与应用

为了使支持向量机(SVM)获得更好的分类效果,针对人为选择参数的随机性,提出了利用粒子群算法(PSO)进行参数自动选取的优化方法,构建了PSO-SVM模型.在个人信用评估中,通过对粒子适应度函数的设置来控制造成较大损失的第二类误判,应用结果表明:模型在训练和测试样本中的分类精度可以达到95%,第二类误判率分别仅为0.78%和2.02%.利用PSO对SVM中的参数进行优化,可以避免人为选择的随机性,并且在解决分类问题中表现出较好的稳健性.     

K8型单层球面网壳支座减震

为了对跨度为60 m的K8型单层球面网壳进行了减震分析,分别将结构的支座设置成弹性支座和弹性阻尼支座,对结构进行减震控制.结合对结构在有控和无控状态下自振特性的分析,阐述了单层球面网壳结构的减震控制原理.在常遇地震作用下,得到网壳结构的最大水平位移、减震支座的最大水平位移、最大支座反力和最大杆件内力.通过改变支座的刚度和阻尼系数,对两种减震方案的减震效果进行比较.研究结果表明:两种减震装置均有效地控制了K8型单层球面网壳结构的支座反力,弹性阻尼支座还可以明显地降低减震支座的水平位移.     

PCNN模型在彩色图像分割中的应用

针对脉冲耦合神经网络(PCNN)模型主要应用于灰度图像处理的局限性,利用脉冲发生器将颜色信息引入模型作为输入,与灰度信息共同控制神经元的内部行为,控制等灰度值的不同颜色区域分期点火,实现彩色图像的精确分割.双输入PCNN模型实现了彩色图像的分割,同时保持了PCNN模型对噪声的鲁棒性,从简单的仿真图像和实际图像两方面验证了此分割方法的有效性.     

桁架结构形状优化的粒子群优化算法

为解决有应力约束、几何约束以及局部稳定性约束的桁架结构的形状优化设计,将粒子群优化(PSO)算法应用于桁架结构的形状优化设计.首先详细介绍了原始PSO算法的基本原理,然后引入压缩因子改进了原始的PSO算法,并提出了合理的参数设置值.优化计算过程中,综合考虑了节点坐标和截面面积等两类不同性质的设计变量.最后对几个经典问题进行了求解,并与传统的优化算法进行了比较.数值结果表明,改进的PSO算法具有良好的收敛性和稳定性,可以有效地进行桁架结构的形状优化设计.     

TH-PPM UWB通信系统多址性能研究

介绍了TH-PPM UWB信号,在AWGN信道和多用户干扰条件下对系统的性能进行了分析,包括误码率分析和系统所能接入的最大用户数,对其误码率进行了仿真.结果表明,随着接入用户数目的增加,系统误码率也增加,而增大重复码长度和脉冲重复时间则可降低系统误码率,但增加脉冲重复时间会降低信息传输速率.     

大口径光学透镜瞬态温度场模拟

为考察热辐射效应对光学元件稳、瞬态温度场的影响,给出了考虑辐射和导热耦合换热情况下复杂几何结构光学透镜稳、瞬态热特性的适体坐标下离散坐标法.对喷管形和圆柱形介质进行计算并与其他方法比较表明,建立的适体坐标系下辐射与导热耦合换热计算方法精度较好.对两端为球缺中间为圆柱的大口径透镜内的辐射与导热耦合换热过程进行了数值模拟.计算结果表明,为了精确地预测半透明介质如透镜内的温度分布,应该考虑辐射与导热耦合作用,且应当考虑介质的光谱特性.     

基于改进粒子群算法的匹配追踪分解优化研究

文章针对使用牛顿法进行匹配追踪分解信号的速度慢、精度低等问题,在具有全局优化能力的粒子群算法基础上,提出了一种结合局部单纯形搜索并引入变异操作的改进粒子群算法实现信号匹配追踪分解.利用单纯形搜索增加了算法的局部开发能力,通过变异操作控制种群多样性以避免早熟收敛,增强了算法全局探测能力;并以描述机械系统的振动冲击响应作为基原子与单一粒子群算法实现匹配追踪分解信号的结果进行对比,证明了使用改进粒子群算法的匹配追踪分解能够快速准确提取信号特征参数,同时成功识别出某轴承发生外圈损伤时隐含在振动信号中的周期性冲击脉冲故障特征.结果表明,加入单纯形和变异的改进粒子群算法有效降低了匹配追踪计算复杂度,提高了信号特征提取准确度.     

Support vector machine forecasting method improved by chaotic particle swarm optimization and its application

By adopting the chaotic searching to improve the global searching performance of the particle swarm optimization (PSO), and using the improved PSO to optimize the key parameters of the support vector machine (SVM) forecasting model, an improved SVM model named CPSO-SVM model was proposed. The new model was applied to predicting the short term load, and the improved effect of the new model was proved. The simulation results of the South China Power Market’s actual data show that the new method can effectively improve the forecast accuracy by 2.23% and 3.87%, respectively, compared with the PSO-SVM and SVM methods. Compared with that of the PSO-SVM and SVM methods, the time cost of the new model is only increased by 3.15 and 4.61 s, respectively, which indicates that the CPSO-SVM model gains significant improved effects. Foundation item: Project(70572090) supported by the National Natural Science Foundation of China     

Average cluster sizes and cluster size distributions of superfine nickel particles in light media

Effects of shear rates on average cluster sizes (ACSs) and cluster size distributions (CSDs) in uni- and bi-systems of partly charged superfine nickel particles were investigated by Brownian dynamics, and clustering properties in these systems were compared with those in non-polar systems. The results show that the ACSs in bi-polar systems are larger than those in the non-polar systems. In uni-polar systems the behavior of clustering property differs: at the lower ionic concentration (10%), repulsive force is not strong enough to break clusters, but may greatly weaken them. The clusters are eventually cracked into smaller ones only when concentration of uni-polar charged particles is large enough. In this work, the ionic concentration is 20%. The relationship between ACS and shear rates follows power law in a exponent range of 0.176–0.276. This range is in a good agreement with the range of experimental data, but it is biased towards the lower limit slightly. Foundation item: Projects(50474037, 50874087) supported by the National Natural Science Foundation of China; Project (BK2006078) supported by the Natural Scientific Funds of Jiangsu Province, China     

Novel fluorescence sensor based on covalent immobilization of 3-amino-9-ethylcarbazole via electrostatically assembled gold nanoparticle layer

A novel technique of immobilizing indicator dyes by electrostatic adsorption and covalent bonding to fabricate optical sensors was developed. 3-Amino-9-ethylcarbazole (AEC) was attached to the outmost surface of quartz glass slide via aminosilanizing the slide, crosslinking chitosan, adsorbing Au nanoparticle, self-assembling HS(CH₂)₁₁OH, and coupling AEC. Thus, an AEC-immobilized optical sensor was obtained. The sensor exhibits a wide linear response range from 7.0×10⁻⁷ to 1.0×10⁻⁴ mol/L and a correlation coefficient of 0.995 9 for the detection of 2-nitrophenol. The detection limit and response time of the sensor are 1.0×10⁻⁷ mol/L and less than 10 s, respectively. The fluorescence intensity of the used sensor can be restored to the blank value by simply rinsing with blank buffer. A very effective matrix for immobilizing indicator dye is provided by the proposed technique, which is adaptable to other indicator dyes with amino groups besides AEC. Foundation item: Project(20775010) supported by the National Natural Science Foundation of China; Project(208095) supported by the Key Project of Ministry of Education, China; Project(07A006) supported by the Scientific Research Fund of Hunan Provincial Education Department, China; Project(07JJ3020) supported by Hunan Provincial Natural Science Foundation of China