约束优化无罚函数非单调SQP算法

本文研究求解非线性约束优化问题.利用非单调无罚函数方法,提出了一个新的序列二次规划算法.该算法在每次迭代过程中只需求解一个QP子问题和一个线性方程组.在一般条件下,算法具有全局收敛性,数值结果表明,计算量小于单调且含罚函数的传统算法.     

矩形网格上的二元切触有理插值

二元切触有理插值是有理插值的一个重要内容,而降低其函数的次数和解决其函数的存在性是有理插值的一个重要问题.二元切触有理插值算法的可行性大都是有条件的,且计算复杂度较大,有理函数的次数较高.利用二元Hermite（埃米特）插值基函数的方法和二元多项式插值误差性质,构造出了一种二元切触有理插值算法并将其推广到向量值情形.较之其它算法,有理插值函数的次数和计算量较低.最后通过数值实例说明该算法的可行性是无条件的,且计算量低.     

基本初等函数的高精度快速计算的加速算法

在现有的基本初等函数的高精度快速算法基础上,进一步研究基本初等函数的加速算法.现有的基本初等函数的高精度快速算法是通过对函数进行幂级数展开的方式来实现函数的任意精度快速计算.而其加速算法则是在幂级数展开之前,先利用函数的多种性质来缩减函数的参数,减少函数在进行幂级数展开时的计算难度,提高函数的计算速度.给出了加速算法,并从计算误差和算法复杂性两方面对该算法进行了分析,给出了误差最小,算法复杂性最低的最优加速算法.然后,对于三角函数、双曲函数、指数函数以及它们的反函数,在实数域上给出了的具体的加速过程和计算结果.     

基于增广Lagrange函数的约束优化问题的一个信赖域方法

本文提出一个求解不等式约束优化问题的基于指数型增广Lagrange函数的信赖域方法.基于指数型增广Lagrange函数,将传统的增广Lagrange方法的精确求解子问题转化为一个信赖域子问题,从而减少了计算量,并建立相应的信赖域算法.在一定的假设条件下,证明了算法的全局收敛性,并给出相应经典算例的数值实验结果.     

任意初始点下的广义梯度投影滤子算法

提出了一个任意初始点的广义梯度滤子方法. 该方法不使用罚函数以避免由此带来的缺陷并可以减少计算量. 方法的另一个特点是不因使用了滤子技术而使算法早熟或陷入循环. 算法对初始点没有要求并在比较合理的条件下具有全局收敛性.     

基于Taylor级数的矩阵双曲余弦函数的数值算法

提出了一种基于Taylor级数的矩阵双曲余弦函数的数值逼近算法,为减少计算量使用了Paterson-Stockmeyer方法来计算矩阵Taylor多项式,对逼近误差进行了绝对后向误差分析以减少误差,并设计了算法可以较为快速且准确地求解矩阵双曲余弦函数,最后进行了数值实验,验证了算法的有效性.     

具有承袭性的高阶导数有理插值算法

切触有理插值是函数逼近的一个重要内容,而降低切触有理插值的次数和解决切触有理插值函数的存在性是有理插值的一个重要问题.切触有理插值函数的算法大都是基于连分式进行的,其算法可行性是有条件的,且计算量较大.利用Newton(牛顿)多项式插值的承袭性和分段组合的方法,构造出了一种无极点且满足高阶导数插值条件的切触有理插值函数,并推广到向量值切触有理插值情形;既解决了切触有理插值函数存在性问题,又降低了切触有理插值函数的次数.最后给出误差估计,并通过数值实例说明该算法具有承袭性、计算量低、便于编程等特点.     

一类无需线性搜索的记忆梯度法

提出一类新的求解无约束优化问题的记忆梯度法,证明了算法的全局收敛性.当目标函数为一致凸函数时,对其线性收敛速率进行了分析.新算法在迭代过程中无需对步长进行线性搜索,仅需对算法中的一些参数进行预测估计,从而减少了目标函数及梯度的迭代次数,降低了算法的计算量和存储量.数值试验表明算法是有效的.     

基于指数型核函数的线性规划原始对偶内点算法

给出线性规划原始对偶内点算法的一个单变量指数型核函数.首先研究了这个指数型核函数的性质以及其对应的障碍函数.其次,基于这个指数型核函数,设计了求解线性规划问题的原始对偶内点算法,得到了目前小步算法最好的理论迭代界.最后,通过数值算例比较了基于指数型核函数的原始对偶内点算法和基于对数型核函数的原始对偶内点算法的计算效果.     

不等式约束极大极小问题的可行下降束方法

本文提出一个求解不等式约束极大极小问题的可行下降束方法.该方法的主要特点有(1)借助于函数的次梯度及束方法思想,不需要假设原问题的分量函数具备光滑性;(2)利用部分割平面模型技术,每次无效步迭代仅需利用一个分量函数的函数值和次梯度产生新的割平面,从而有效减少了计算量;(3)能够保证有效迭代点的可行性及目标函数的下降性;(4)引入次梯度聚集技术,对束集中的次梯度进行聚集,克服了数值计算和存储的困难;(5)算法具备全局收敛性,且初步的数值试验表明算法是有效的.     

Factors affecting the process of proficient mental addition and subtraction: case studies of flexible and inflexible computers

The relationship between mental computation and number sense is complex: mental computation can facilitate number sense when students are encouraged to be flexible, but flexibility and number sense is neither sufficient nor necessary for accuracy in mental computation. It is possible for familiarity with a strategy to compensate for a lack of number sense and inefficient processes. This study reports on six case studies exploring Year 3 students’ procedures for and understanding of mental addition and subtraction, and understanding of number sense and other cognitive, metacognitive, and affective factors associated with mental computation. The case studies indicate that the mental computation process is composed of four stages in which cognitive, metacognitive and affective factors operate differently for flexible and inflexible computers. The authors propose a model in which the differences between computer types are seen in terms of the application of different knowledges in number facts, numeration, effect of operation on number, and beliefs and metacognition on strategy choice and strategy implementation.     

非奇异阵特征极值和条件数的近似计算

本文从共轭梯度法的公式推导出对称正定阵A与三对角阵B的相似关系,B的元素由共轭梯度法的迭代参数确定.因此,对称正定阵的条件数计算可以化成三对角阵条件数的计算,并且可以在共轭梯度法的计算中顺带完成.它只需增加O(s)次的计算量,s为迭代次数.这与共轭梯度法的计算量相比是可以忽略的.当A为非对称正定阵时,只要A非奇异,即可用共轭梯度法计算ATA的特征极值和条件数,从而得出A的条件数.对不同算例的计算表明,这是一种快速有效的简便方法.     

Complexity of linear programming

The complexity of linear programming is discussed in the “integer” and “real number” models of computation. Even though the integer model is widely used in theoretical computer science, the real number model is more useful for estimating an algorithm's running time in actual computation.Although the ellipsoid algorithm is a polynomial-time algorithm in the integer model, we prove that it has unbounded complexity in the real number model. We conjecture that there exists no polynomial-time algorithm for the linear inequalities problem in the real number model. We also conjecture that linear inequalities are strictly harder than linear equalities in all “reasonable” models of computation.     

An efficient algorithm for the computation of the metric average of two intersecting convex polygons, with application to morphing

Motivated by the method for the reconstruction of 3D objects from a set of parallel cross sections, based on the binary operation between 2D sets termed “metric average”, we developed an algorithm for the computation of the metric average between two intersecting convex polygons in 2D. For two 1D sets there is an algorithm for the computation of the metric average, with linear time in the number of intervals in the two 1D sets. The proposed algorithm has linear computation time in the number of vertices of the two polygons. As an application of this algorithm, a new technique for morphing between two convex polygons is developed. The new algorithm performs morphing in a non-intuitive way.     

非线性极大极小系统全局优化算法的分析

非线性极大极小系统的全局优化可用于柔性制造和智能交通的决策与控制.实现了非线性极大极小系统的全局优化算法的仿真,并进行了计算时间分析.数值实验表明了全局优化算法的可行性.算法的计算时间主要由系统的优化极大射影矩阵数目决定,而优化极大射影矩阵数目与系统解析式中单极大式的系数紧密相关,系数取值越分散,简约极大射影矩阵的效果越好,计算效率越高.     

On the Distributions of the State Sizes of Closed Continuous Time Homogeneous Markov Systems

The evolution of the state sizes of a closed continuous-time homogeneous Markov system is determined by the convolution of multinomial distributions expressing the number of transitions between the states of the system. In order to investigate the distributions of the state sizes, we provide the computation of their moments, at any time point, via a recursive formula concerning the derivative of the moments. The basic result is given by means of a new vector product which is similar to the Kronecker product. Finally, a formula for the computation of the state sizes distributions is given.     

An efficient boundary element method for two-dimensional transient wave propagation problems

The boundary element method (BEM) has been recognized by its unique feature of requiring neither internal cells nor their associated domain integrals in the computation. The method preserves its elegance for transient problems by means of a certain time-stepping scheme that initiates the time integration always from the initial time. Unfortunately, this time-marching scheme becomes rather difficult to apply, because the computation time and storage requirement grow dramatically with the increasing number of time steps. This paper shows that a reduction of one half of the computation time as well as the storage requirement can be achieved by an efficient truncation scheme for two-dimensional transient wave propagation problems. In particular, a guiding parameter for the determination of the truncation limit is proposed, and the overall measure of the error with respect to the truncation guide parameter is established.     

Numerical aspects of the XFEM — The XFEM p–Version

The XFEM is known to approximate the displacements and stresses around a crack tip in a very efficient way. But as we will present in this paper we have to deal with a phenomenon coming along with this method that compels us to use higher order shape functions for those elements that are enriched by the crack tip functions. For the computation of the stress–intensity–factors we are using a J–integral over a circular domain Ω. The accuracy of the results depend on • the radius of Ω • the number of elements used in the XFEM computation • the number of nodes which were enriched by the crack tip functions (number of layers) and • the shape functions which were used for the standard FE term For more information about the XFEM we refer to [1]. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Multi-Path Approach for Reliability-Redundancy Allocation Using a Scaling Method

The reliability-redundancy allocation problem is an optimization problem that achieves better system reliability by determining levels of component redundancies and reliabilities simultaneously. The problem is classified with the hardest problems in the reliability optimization field because the decision variables are mixed-integer and the system reliability function is nonlinear, non-separable, and non-convex. Thus, iterative heuristics are highly recommended for solving the problem due to their reasonable solution quality and relatively short computation time. At present, most iterative heuristics use sensitivity factors to select an appropriate variable which significantly improves the system reliability. The sensitivity factor represents the impact amount of each variable to the system reliability at a designated iteration. However, these heuristics are inefficient in terms of solution quality and computation time because the sensitivity factor calculations are performed only at integer variables. It results in degradation of the exploration and growth in the number of subsequent continuous nonlinear programming (NLP) subproblems. To overcome the drawbacks of existing iterative heuristics, we propose a new scaling method based on the multi-path iterative heuristics introduced by Ha (2004). The scaling method is able to compute sensitivity factors for all decision variables and results in a decreased number of NLP subproblems. In addition, the approximation heuristic for NLP subproblems helps to avoid redundant computation of NLP subproblems caused by outlined solution candidates. Numerical experimental results show that the proposed heuristic is superior to the best existing heuristic in terms of solution quality and computation time.     

The Bezout Number for Piecewise Algebraic Curves

The computation of the Bezout number, the maximum number of intersection points between two piecewise algebraic curves whose common points are finite, is considered. A piecewise algebraic curve is a curve determined by a bivariate spline function. It is found that the maximum number of intersections depends not only on the degrees and the differentiability of the spline functions, but also on the structure of the partition on which the spline functions are defined.