避免引入人工变量求线性规划可行基的一个新方法

讨论了线性规划的单纯形解法,给出了不须加人工变量就可得到一个可行基的算法.通过大量的算例表明此法比传统的单纯形方法具有算法结构简单,计算量小的优点.     

线性规划全方位变化的灵敏度分析

本文对变量目标函数系数、变量约束系数向量以及约束右端项向量同时变化进行灵敏度分析。不仅对变化后可能出现的各种情况进行分析处理,尤其在对偶可行性和可行性都不满足时,利用联合算法进行处理,并通过算例加以说明。     

n维欧氏空间En中Child不等式的推广

利用几何不等式理论与解析方法,研究了n维欧氏空间En中n维单形的内点到各顶点的距离与到各侧面距离之间的关系,获得相关的几个几何不等式,推广了Child不等式．     

线性规划的最钝角松弛算法

本文提出一个基于最钝角原理的松弛算法求解线性规划问题。该算法依据最钝角原理略去部分约束得到一个规模较小的子问题,用原始单纯形算法解之;再添加所略去的约束恢复原问题,若此时全部约束条件均满足则已获得一个基本最优解,否则用对偶单纯形算法继续求解。初步的数值试验表明,新算法比传统两阶段单纯形算法快得多。     

An algorithm for three-dimensional Voronoi S-network

The paper presents an algorithm for calculating the three-dimensional Voronoi-Delaunay tessellation for an ensemble of spheres of different radii (additively-weighted Voronoi diagram). Data structure and output of the algorithm is oriented toward the exploration of the voids between the spheres. The main geometric construct that we develop is the Voronoi S-network (the network of vertices and edges of the Voronoi regions determined in relation to the surfaces of the spheres). General scheme of the algorithm and the key points of its realization are discussed. The principle of the algorithm is that for each determined site of the network we find its neighbor sites. Thus, starting from a known site of the network, we sequentially find the whole network. The starting site of the network is easily determined based on certain considerations. Geometric properties of ensembles of spheres of different radii are discussed, the conditions of applicability and limitations of the algorithm are indicated. The algorithm is capable of working with a wide variety of physical models, which may be represented as sets of spheres, including computer models of complex molecular systems. Emphasis was placed on the issue of increasing the efficiency of algorithm to work with large models (tens of thousands of atoms). It was demonstrated that the experimental CPU time increases linearly with the number of atoms in the system, O(n).     

En 中n维单形二面角的角平分面的性质

利用距离几何的理论与方法研究了En中n维单形内二面角与外二面角的角平分面的性质,获得了n维单形内二面角与外二面角平分面的2个定理.     

基于5阶球面最简相径的改进型容积卡尔曼滤波在SINS/DVL组合导航中的应用

为提高水下SINS/DVL组合导航系统的精度,建立了捷联惯性导航系统(SINS)的非线性误差模型,并建立多普勒测速仪的误差方程,以SINS为主导航设备建立SINS/DVL组合导航系统模型。设计了5阶球面最简相径容积卡尔曼滤波器,采用了球面最简相径采样规则改进容积卡尔曼滤波,并应用于SINS/DVL组合导航系统中。通过数学平台仿真验证了5阶球面最简相径容积卡尔曼滤波方法有效性,仿真结果表明:该方法能够有效提高SINS/DVL组合导航系统的精度,且稳定性好。     

简化超球体SRUKF的GPS/MIMU组合姿态估计

针对加速度计、磁强计易受机动加速度和周围环境影响的问题,提出了一种单基线GPS/MIMU组合姿态估计方法。该方法利用单基线GPS天线代替磁强计提供航向信息,并利用GPS的速度信息对加速度计输出的机动加速度分量进行补偿,构成单基线GPS/MIMU组合姿态确定单元。同时考虑实时性要求,在SRUKF的基础上,引入加性非扩展形式和超球体单形采样,提出了简化超球体平方根UKF算法,通过减少状态维数和采样点的数量,降低算法计算量。建立加性噪声下的单基线GPS/MIMU姿态模型,并采用简化超球体平方根UKF算法进行姿态估计。实验结果表明,单基线GPS的速度信息可以有效提高加速度计对水平倾角的确定精度,姿态估计算法收敛后的最大误差小于0.8°,估计精度与UKF相当,且执行时间仅是UKF的42%。     

Czerny-Turner正交型光栅单色仪的波长校准

研究了Czerny-Turner正交型光栅单色仪的波长校准技术。结合单色仪的结构参数和特点,提出以符合光栅方程的正弦曲线作为仪器出射波长的校准方程,基于最小二乘法原理给出校准方程的拟合残差表达式,由于校准方程的非线性,应用二维Nelder-Mead单纯形法求解拟合残差的待定系数,建立了波长与光栅转角的精确表达式,并通过实验验证了该算法的准确性。结果表明,经过校准的单色仪波长定位精度小于0.1nm,比设计要求提高一个数量级。该方法在Czerny-Turner正交型光栅单色仪的波长校准过程中,应用简单,容易实现,只需稍加修改步进电机的控制程序,即可完成对仪器出射波长的快速实时校准,实用性强。     

Solving a multiobjective traveling salesman problem by dynamic programming

In this paper Klötzler's method of multiobjective dynamic programming is applied to the solution of a two-dimensional traveling salesman problem. In this way Bellman's and Held/Karp's dynamic programming approach to one-dimensional traveling salesman problems is extended to the multiobjective case.