求解一类特殊二次规划问题的分布式牛顿算法

针对二次规划问题,现有的基于对偶分解和梯度方法的分布式算法由于没有充分利用目标函数的二阶信息,算法并不高效.针对一类特殊二次规划问题提出分布式牛顿算法,算法在计算对偶向量时使用Jacobi迭代,使算法不仅能够分布式执行并且可以并行运算.通过证明Jacobi矩阵的谱半径小于1保证了迭代的收敛性.最后通过数值实验说明分布式牛顿算法在运行时间上的高效性.     

A Nonmonotone Smoothing Newton Algorithm for Weighted Complementarity Problem

Journal of Optimization Theory and Applications - The weighted complementarity problem (denoted by WCP) significantly extends the general complementarity problem and can be used for modeling a...     

A Global Newton Method for the Nonsmooth Vector Fields on Riemannian Manifolds

Journal of Optimization Theory and Applications - This paper proposes and analyzes a globalized version of the Newton method for finding a singularity of the nonsmooth vector fields. Basically, the...     

一种部分约束满足车辆路线问题及其求解算法

描述了一类过度约束车辆路线问题，其中可用车辆数较少而时间窗口等其它约束又不允许放松，因而导致不存在满足所有约束的可行解。此时问题求解可以转化为一类部分约束满足问题来处理，相应的优化目标是最小化未访问顾客的损失和。本给出了求解这类特殊问题的一种禁忌搜索算法设计，并通过规模不同的几个算例与其它常用方法进行了比较。最后分析了模型和算法的实用意义。     

Subgradient Algorithm on Riemannian Manifolds

The subgradient method is generalized to the context of Riemannian manifolds. The motivation can be seen in non-Euclidean metrics that occur in interior-point methods. In that frame, the natural curves for local steps are the geodesies relative to the specific Riemannian manifold. In this paper, the influence of the sectional curvature of the manifold on the convergence of the method is discussed, as well as the proof of convergence if the sectional curvature is nonnegative.     

一种求解约束优化问题的改进差分进化算法

针对在处理约束优化问题时约束条件难以处理的问题,提出了一种求解约束优化问题的改进差分进化算法.即在每代进化前将群体分为可行个体和不可行个体两类,对不可行个体,用差量法将其逐个转化为可行个体,并保持种群规模不变,经过一序列的进化后,计算所有可行个体的适应度并找到问题的最优解.对5个经典函数进行了优化测试,测试结果表明提出的算法对求解约束优化问题是有效的.     

An Eigenvalue Problem on Negatively Curved Manifolds

Consider the eigenvalue problem: Δgu - λKu = 0 \quad in D where D is the unit disc of the complex plane, g is a complete metric conformal to the Poincaré metric on D, and K is the Gaussian curvature. It is shown that if λ > \frac{1}{2} (λ > \frac{1}{4}in the case of K ≤ 0), then the above problem has no positive solutions.     

框式凸二次规划问题的非精确不可行内点算法

对框式凸二次规划问题提出了一种非精确不可行内点算法 ,该算法使用的迭代方向仅需要达到一个相对的精度 .在初始点位于中心线的某邻域内的假设下 ,证明了算法的全局收敛性     

非线性互补问题的一种全局收敛的显式光滑Newton方法

本针对Po函数非线性互补问题，给出了一种显式光滑Newton方法，该方法将光滑参数μ进行显式迭代而不依赖于Newton方向的搜索过程，并在适当的假设条件下，证明了算法的全局收敛性。     

一种双目标带约束工件成类的平行机器调度启发式规则

提出了一种快速而有效的启发式规则(fam ily slack,简称FSLACK),来求解极小化总延误时间和极小化最大完工时间两个目标,工件按产品类型成组,带模具数量约束的平行机器生产调度问题.本文提出的FSLACK与EDD、LPT及SLACK进行了比较.随机订单的测试结果表明,本文提出的启发式规则在求解双目标带约束工件成类的平行机器调度问题上是有效的.这表明该算法可以应用在成型加工业的现场作业调度.     

Brownian Motion and the Dirichlet Problem at Infinity on Two-dimensional Cartan-Hadamard Manifolds

After recalling the Dirichlet problem at infinity on a Cartan-Hadamard manifold, we describe what is known under various curvature assumptions and the difference between the two-dimensional and the higher-dimensional cases. We discuss the probabilistic formulation of the problem in terms of the asymptotic behavior of the angular component of Brownian motion. We then introduce a new (and appealing) probabilistic approach that allows us to prove that the Dirichlet problem at infinity on a two-dimensional Cartan-Hadamard manifold is solvable under the curvature condition K?≤?(1?+?ε)/(r ² logr) outside of a compact set, for some ε?>?0, in polar coordinates around some pole. This condition on the curvature is sharp, and improves upon the previously known case of quadratic curvature decay. Finally, we briefly discuss the issues which arise in trying to extend this method to higher dimensions.     

The Yamabe Problem and Applications on Noncompact Complete Riemannian Manifolds

We let (M,g) be a noncompact complete Riemannian manifold of dimension n 3 whose scalar curvature S(x) is positive for all x in M. With an assumption on the Ricci curvature and scalar curvature at infinity, we study the behavior of solutions of the Yamabe equation on –u+[(n–2)/(4(n–1))]Su=qu ^{(n+2)/(n–2)} on (M,g). This study finds restrictions on the existence of an injective conformal immersion of (M,g) into any compact Riemannian n -manifold. We also show the existence of a complete conformal metric with constant positive scalar curvature on (M,g) with some conditions at infinity.     

The Corona Problem in Carleman Algebras on Stein Manifolds

Hörmander L ² estimates for the $\bar\partial$ -operator on Stein Manifolds are applied to the solution of the Corona problem in Carleman Algebras on Stein Manifolds.     

调度问题中两类分离约束传播算法的比较及一种改进算法

约束传播算法是求解约束满足问题的一种重要方法。调度问题是一种特殊的约束满足问题。本介绍了调度问题中的Edge-Finding和Energy-Reasoning两种分离约束传播算法，并对它们进行了比较，中最后给出了一种结合Energy-Reasoning的Edge-Finding改进算法。     

黎曼流形上次梯度算法及其收敛性的研究

主要研究了黎曼流形M上一类最优化问题,并给出了解决该问题的一种ε次梯度算法.并在流形M是一个完备的且具有非负截面曲率的黎曼流形时,证明了算法得出的无限迭代点列的收敛性.     

基于牛顿插值算法的模糊控制器

提出一种基于牛顿在线插值算法的模糊控制器,介绍该控制器的设计方法。该方法既简化了合成推理运算,又能满足模糊控制规则的完整性要求,从本质上消除由于量化误差和调节死区给模糊控制系统带来的稳态误差与颤振现象。通过仿真证明系统的性能得到明显改善。     

Fenchel Duality Theory and a Primal-Dual Algorithm on Riemannian Manifolds

Foundations of Computational Mathematics - This paper introduces a new notion of a Fenchel conjugate, which generalizes the classical Fenchel conjugation to functions defined on Riemannian...     

Primal Newton Method for the Linear Cone Programming Problem

A linear cone programming problem containing among the constraints a second-order cone is considered. For solving this problem, a primal Newton method which is constructed with the help of the optimality conditions is proposed. Local convergence of this method is proven.