求解旅行商问题的一种改进粒子群算法

本文研究了求解旅行商问题的粒子群算法。针对标准粒子群算法在求解旅行商问题过程中容易出现早熟和停滞现象的缺点,提出了一种改进的粒子群算法。首先,在初始种群的选取过程中,利用改进的贪婪策略直接获得具有较高性能的初始种群以提高算法的搜索效率。其次,通过引入次优吸引子,使粒子在搜索过程中可以更加充分地利用群体的信息来提高自身的性能,有效抑制收敛过程中的停滞现象,提高算法的搜索能力。最后为了验证所提出的方法的有效性和可行性,对TSPLIB标准库中的多个实例进行了测试,并给出了数值结果。     

A comparative study of recursive methods

Summary Recursive methods represent a well-established and wide-spread technique for generation of dynamical equations of mechanical systems. Their main objective is to obtain computational schemes involving as few operations as possible. In this paper, we show that, besides the consideration of simplification resulting from a careful representation of geometric relationships between contiguous links, only two factors determine the overall efficiency of these methods. These factors are: the choice of an appropriate reference point for the kinematic and kinetic equations, and the frame of decomposition used to represent the involved vectors and tensors. Further, we derive a computational scheme which compares advantageously with existing methods. As only elementary laws of mechanics are applied, the exposition is also suitable for practising engineers seeking to understand the main differences between existing methods. A comparison with one of the reportedly most effective non-recursive method elucidates the advantages and bounds of the present approach.     

关于体上矩阵方程A_（mn）X_（ns）＝B_（ms）之解的注记

本文改进了文［１］的结果，利用体上矩阵的初等行和列变换，给出了任意体上的矩阵方程ＡｍｓＸａｓ＝Ｂｍｓ的一种更为实用的简便解法。     

Optical Module for Modified Signed-Digit Computing Based on Bit Plane Encoding and Pattern Operations

We propose here a new optical modified signed-digit (MSD) adder module based on bit plane pattern encoding of MSD digits and pattern operations. The pattern operations required in the algorithm are duplication, combination and shifting. They are simply performed by using optical components such as beam splitters, mirrors and parallel plates, instead of optical logic arrays. An optical MSD adder module comprised of six properly interacting blocks with all optical components packaged on a common substrate is presented in detail. We analyze the system errors caused by manufacture and alignment of optical components, the information throughput, the intensity nonuniformity and the system volume of the adder module.     

打点计时器电源的改进

根据振荡电路产生不同周期的计时信号，对打点计时器的供电电源进行了改进。改进后，实验者可根据不同的运动速度来选择适宜的打点计时单位，使纸带上的点疏密程度较为合理，从而减少测量误差。     

单机排序问题|r_j,prmp| ∑w_j(1-e~(-ac_j)) 的动态在线调度

本文首先一般化了可中断的概念,并建立了相应的中断—安装重复模型,然后研究了单机排序问题1|rj,prmp| wj(1-e-acj)在中断—重复和中断—安装重复模型下的动态在线排序问题,给出了只考虑当前可用信息而不是考虑全部任务信息的在线调度规则。     

像素1:1匹配的晶体全息存储系统中像素位置偏移的补偿算法

在体全息存储系统的数据读出过程中,为了提高读出速度,需要做到SLM与CCD像素的1∶1匹配。但在实现1∶1匹配的过程中,如何降低误码率是一个难点。利用傅里叶光学成像理论,推导出了在4f系统中单个SLM像素成像的光强分布,并由此定量地分析了像素位置偏移对读出误码率的影响。提出了针对像素偏移的数值补偿算法,并获得了较理想的实验结果。     

Gauss periods: orders and cryptographical applications

Experimental results on the multiplicative orders of Gauss periods in finite fields are presented. These results indicate that Gauss periods have high order and are often primitive (self-dual) normal elements in finite fields. It is shown that Gauss periods can be exponentiated in quadratic time. An application is an efficient pseudorandom bit generator.

     

单源模糊数及其运算

本文首先指出模糊数的运算存在模糊源问题，然后定义了单源模糊数的一些基本概念，并建立了单源模糊数的运算方法，最后给出了单源模糊数线性方程组的求解方法。     

On tertiary exotic characteristic classes

We prove the existence and nontriviality of tertiary exotic characteristic classes extending the results of Peterson and Ravenel for secondary exotic classes.