Accelerating the convergence in the single-source and multi-source Weber problems

The modified Weiszfeld method [Y. Vardi, C.H. Zhang, A modified Weiszfeld algorithm for the Fermat-Weber location problem, Mathematical Programming 90 (2001) 559-566] is perhaps the most widely-used algorithm for the single-source Weber problem (SWP). In this paper, in order to accelerate the efficiency for solving SWP, a new numerical method, called Weiszfeld-Newton method, is developed by combining the modified Weiszfeld method with the well-known Newton method. Global convergence of the new Weiszfeld-Newton method is proved under mild assumptions. For the multi-source Weber problem (MWP), a new location-allocation heuristic, Cooper-Weiszfeld-Newton method, is presented in the spirit of Cooper algorithm [L. Cooper, Heuristic methods for location-allocation problems, SIAM Review 6 (1964) 37-53], using the new Weiszfeld-Newton method in the location phase to locate facilities and adopting the nearest center reclassification algorithm (NCRA) in the allocation phase to allocate the customers. Preliminary numerical results are reported to verify the evident effectiveness of Weiszfeld-Newton method for SWP and Cooper-Weiszfeld-Newton method for MWP.     

A Penalty Approach for Generalized Nash Equilibrium Problem

The generalized Nash equilibrium problem (GNEP) is a generalization of the standard Nash equilibrium problem (NEP),in which both the utility function and the strategy space of each player depend on the strategies chosen by all other players.This problem has been used to model various problems in applications.However,the convergent solution algorithms are extremely scare in the literature.In this paper,we present an incremental penalty method for the GNEP,and show that a solution of the GNEP can be found by solving a sequence of smooth NEPs.We then apply the semismooth Newton method with Armijo line search to solve latter problems and provide some results of numerical experiments to illustrate the proposed approach.     

A smoothing-type algorithm for the second-order cone complementarity problem with a new nonmonotone line search

The second-order cone complementarity problem (denoted by SOCCP) can be effectively solved by smoothing-type algorithms, which in general are designed based on some monotone line search. In this paper, based on a new smoothing function of the Fischer–Burmeister function, we propose a smoothing-type algorithm for solving the SOCCP. The proposed algorithm uses a new nonmonotone line search scheme, which contains the usual monotone line search as a special case. Under suitable assumptions, we show that the proposed algorithm is globally and locally quadratically convergent. Some numerical results are reported which indicate the effectiveness of the proposed algorithm.     

非线性方程组的Newton流线法

为求解非线性方程组F(x)=0, 研究了Newton流方程x_t=V(x)=-(DF(x))^-1F(x),x(0)=x⁰,及数值Newton流x^j+1=x^j+hV(x^j),h∈(0,1].导出了减幅指标g_j(h)=||F(x^j+1)||/||F(x^j)||=1-h+h²d_jh<1和m重根x^*附近的表示g_j(h)=(1-h/m)^m+h²O(||x^j-x^*||).最后基于4个可计算量g_j,d_j,K_j,q_j,提出了新的Newton流线法,如果投入大量的随机初始点, 能找到所有实根、重根和复根.     

用矩阵符号函数解(广义)周期Sylvester方程

(广义)周期Sylvester方程来源于周期离散线性系统. 本文主要研究这类方程满足特征值分别位于开左半复平面和开右半复平面或位于单位圆周内和单位圆周外条件时用矩阵符号函数求解的数值方法.并通过数值例子说明我们的结论.     

The newton method for C1αmaps In anach Spaceis$fr1;∗$f:∗

On the basis of a work by B.Döring for ?¹→ ?¹maps we give a thourough proof of the convergence and on the convergence speed of the Newton Method for C ¹αmaps in Banach spaces. We describe the application of this classical method for the existence (and local uniqueness and approximation) of nonlinear eigenvalue problems - abstractly - and for the concrete case of nonlinear Dirichlet problem which was for- merly attacked by variational methods     

复合凸优化的快速邻近点算法

在大数据时代,随着数据采集手段的不断提升,大规模复合凸优化问题大量的出现在包括统计数据分析,机器与统计学习以及信号与图像处理等应用中.本文针对大规模复合凸优化问题介绍了一类快速邻近点算法.在易计算的近似准则和较弱的平稳性条件下,本文给出了该算法的全局收敛与局部渐近超线性收敛结果.同时,我们设计了基于对偶原理的半光滑牛顿法来高效稳定求解邻近点算法所涉及的重要子问题.最后,本文还讨论了如何通过深入挖掘并利用复合凸优化问题中由非光滑正则函数所诱导的非光滑二阶信息来极大减少半光滑牛顿算法中求解牛顿线性系统所需的工作量,从而进一步加速邻近点算法.     

Characterization of Levi-Civita and Newton–Cartan connections in dimension 2

In dimension 2, we give a local characterization of Levi-Civita connections, improving a result by G. Thompson (1991), and of Newton–Cartan connections (connections which are limit of Levi-Civita connections).     

二阶锥约束随机变分不等式问题的数值方法研究

本文研究二阶锥约束随机变分不等式(SOCCSVI)问题,运用样本均值近似(SAA)方法结合光滑Fischer-Burmeister互补函数来求解该问题.首先,将SOCCSVI问题的Karush-Kuhn-Tucker系统转化为与之等价的方程组,并证明了该方程组的雅可比矩阵的非奇异性.其次,构造了光滑牛顿算法求解该方程组.最后,文章给出了两个数值实验证明了算法的有效性.