Subwavelength-resolution direct writing using submicron-diameter fibers

<正>A novel direct writing technique using submicron-diameter fibers is presented.This technique adopts contact mode in the process of writing,and submicron lines with different widths have been obtained. Experimental results demonstrate that the resolution of this technique can be smaller than the exposure wavelength of 442 nm,and 380-nm-wide line is achieved.In addition,the distribution of light fields in the photoresist layer is analyzed by finite-difference time-domain method.     

Simultaneous multichannel NRZ-to-RZ format conversion of 4-ASK signal based on phase-intensity hybrid modulation and dispersion compensation fiber

We propose and investigate a novel technique to realize non-return-to-zero （NRZ） to return-to-zero （RZ） format conversion of a multichannel 4-ary amplitude shift keying （4-ASK） signal. The proposed format converter composed of two modulators and a dispersion compensating fiber （DCF） is theoretically inves-tigated using numerical simulation as basis. Research shows that a 4-channel 20-Gb/s NRZ-4-ASK signal can be converted to a RZ-4-ASK signal simultaneously without wavelength shifting and signal quality degradation, with the converted signal multiplexed to 40 Gb/s.     

MS-ADoG域结合ReNLU与VGG-16的矿井双波段图像融合算法

为满足矿井视频监控需求,针对现有图像融合算法获取的融合图像存在伪目标、模糊目标、晕光遮挡目标等问题,提出了一种基于多尺度和自适应高斯差分变换,结合修正非线性单元和VGG-16的矿井双波段图像融合算法.设计了基于多尺度和自适应高斯差分变换的源图像分解模型,将红外和可见光图像分解为基础图像和细节图像;构造了一种修正非线性单...     

基于Goose LDW-PSO的管壳式换热器优化设计

管壳式换热器在工业中的应用越来越广泛,对于设计者和使用者来说使其费用最小化是一个很重要的目标。本文以最小化管壳式换热器的总费用为优化目标,提出了运用基于雁群启示的惯性权重线性下降粒子群优化算法(Goose LDW-PSO)对其进行优化设计。为了验证这种方法的有效性,用Goose LDW-PSO对管壳式换热器进行了优化设计,并与原始设计结果进行了比较。比较结果表明,虽然管子数目增加,但是换热器的总费用比传统方法少了很多。     

Preserving Coherence of Coherent State Studied via Weyl-Wigner Correspondence

By means of the Weyl correspondence and the explicit normally ordered expression of the Wigner operator we convert the time evolution equation of coherent states, governed by some Hamiltonian operators, into seeking for consistent solution of a set of evolution equtions of classical variables which can meet the requirment that an initial coherent state remains coherent all the time.     

An Optical Labeling Scheme with Novel DPSK/PPM Orthogonal Modulation

A novel differential-phase-shift-keying （DPSK）/pulse-position-modulation （PPM） orthogonal modulation is pro- posed for optical labeling applications, with PPM data as the high-speed payload and DPSK signal as the optical label. The systematic setup for the proposed scheme and research on bit-error rate is presented. The results show that at the bit-error rate of 10-~, the power penalties for 70kin fiber transmission are in the range of 1-3 dB to DPSK label and 5-6 dB to PPM payload in our simulation system. This is under the condition that the large extinction ratio of 18-19 dB is used for the intensity modulators of the PPM payload, with a well received DPSK label. Thus the feasibility of the proposed scheme for all optical transmission networks is clearly verified.     

演示实验制作促进动手动脑的大学物理探索式学习的改革尝试

介绍了我们通过制作演示实验装置促进动手动脑的大学物理探索式学习的教学改革实践.详细说明了"动手动脑"学习的试验过程、试验结果.自制物理演示实验装置给学生提供了一个可以自主选题的探索式学习平台.     

不可压磁流体力学方程组的高精度数值解法

为数值求解低雷诺数下不可压流体在电磁场作用下的流动,提出一种四阶紧致差分方法.由二维原始变量的MHD方程组出发,推导出具有较少未知量的电流密度-涡量-流函数形式MHD方程组.建立了求解二维非定常不可压MHD方程组的电流密度-涡量-流函数形式的四阶精度紧致差分格式.为验证本文提出的高精度紧致差分方法的精确性和可靠性,对有...     

线性比式和规划问题的输出空间分支定界算法

本文旨在针对线性比式和规划这一NP-Hard非线性规划问题提出新的全局优化算法.首先,通过引入p个辅助变量把原问题等价的转化为一个非线性规划问题,这个非线性规划问题的目标函数是乘积和的形式并给原问题增加了p个新的非线性约束,再通过构造凸凹包络的技巧对等价问题的目标函数和约束条件进行相应的线性放缩,构成等价问题的一个下界线性松弛规划问题,从而提出了一个求解原问题的分支定界算法,并证明了算法的收敛性.最后,通过数值结果比较表明所提出的算法是可行有效的.     

基于辛格式的深度哈密尔顿神经网络

HNN是一类基于物理先验学习哈密尔顿系统的神经网络.本文通过误差分析解释使用不同积分器作为超参数对HNN的影响.如果我们把网络目标定义为在任意训练集上损失为零的映射,那么传统的积分器无法保证HNN存在网络目标.我们引进反修正方程,并严格证明基于辛格式的HNN具有网络目标,且它与原哈密尔顿量之差依赖于数值格式的精度.数值实验表明,由辛HNN得到的哈密尔顿系统的相流不能精确保持原哈密尔顿量,但保持网络目标;网络目标在训练集、测试集上的损失远小于原哈密尔顿量的损失;在预测问题上辛HNN较非辛HNN具备更强大的泛化能力和更高的精度.因此,辛格式对于HNN是至关重要的.