A trust region affine scaling method for bound constrained optimization

We study a new trust region affine scaling method for general bound constrained optimization problems. At each iteration, we compute two trial steps. We compute one along some direction obtained by solving an appropriate quadratic model in an ellipsoidal region. This region is defined by an affine scaling technique. It depends on both the distances of current iterate to boundaries and the trust region radius. For convergence and avoiding iterations trapped around nonstationary points, an auxiliary step is defined along some newly defined approximate projected gradient. By choosing the one which achieves more reduction of the quadratic model from the two above steps as the trial step to generate next iterate, we prove that the iterates generated by the new algorithm are not bounded away from stationary points. And also assuming that the second-order sufficient condition holds at some nondegenerate stationary point, we prove the Q-linear convergence of the objective function values. Preliminary numerical experience for problems with bound constraints from the CUTEr collection is also reported.     

常微分方程近似解的LS-SVM改进求法

提出一种基于最小二乘支持向量机(LS-SVMs)的求解常微分方程近似解的改进方法.该方法首先通过离散计算域,将常微分方程转换为有约束条件的目标优化问题,然后利用径向基(RBF)核函数可导的性质,用带有导数形式的LS-SVM模型将此优化问题转化为LS-SVM回归问题,进而进行求解.最终得到的闭式近似解具有精度高、连续可微、结构简单且形式固定的特点.该方法适用于任意阶非刚性和奇异的线性常微分方程初值问题和边值问题,以及一阶非线性常微分方程问题.仿真结果验证了该方法具有良好的有效性.     

一类兰切斯特方程的广义变分迭代解

通过构造一组泛函,用变分迭代地方式研究了兰切斯特方程的各次近似解,这种近似解是解析解,还能继续进行各种解析运算.从而可进一步研究对兰切斯特方程进行定性及定量的研究.     

A study of the rate-controlled constrained-equilibrium dimension reduction method and its different implementations

The Rate-Controlled Constrained-Equilibrium (RCCE) method is a thermodynamic based dimension reduction method which enables representation of chemistry involving n _s species in terms of fewer n _r constraints. Here we focus on the application of the RCCE method to Lagrangian particle probability density function based computations. In these computations, at every reaction fractional step, given the initial particle composition (represented using RCCE), we need to compute the reaction mapping, i.e. the particle composition at the end of the time step. In this work we study three different implementations of RCCE for computing this reaction mapping, and compare their relative accuracy and efficiency. These implementations include: (1) RCCE/TIFS (Trajectory In Full Space): this involves solving a system of n _s rate-equations for all the species in the full composition space to obtain the reaction mapping. The other two implementations obtain the reaction mapping by solving a reduced system of n _r rate-equations obtained by projecting the n _s rate-equations for species evaluated in the full space onto the constrained subspace. These implementations include (2) RCCE: this is the classical implementation of RCCE which uses a direct projection of the rate-equations for species onto the constrained subspace; and (3) RCCE/RAMP (Reaction-mixing Attracting Manifold Projector): this is a new implementation introduced here which uses an alternative projector obtained using the RAMP approach. We test these three implementations of RCCE for methane/air premixed combustion in the partially-stirred reactor with chemistry represented using the n _s=31 species GRI-Mech 1.2 mechanism with n _r=13 to 19 constraints. We show that: (a) the classical RCCE implementation involves an inaccurate projector which yields large errors (over 50%) in the reaction mapping; (b) both RCCE/RAMP and RCCE/TIFS approaches yield significantly lower errors (less than 2%); and (c) overall the RCCE/TIFS approach is the most accurate, efficient (by orders of magnitude) and robust implementation.     

Homogenization of First Order Equations with (u/ε)-Periodic Hamiltonians Part II: Application to Dislocations Dynamics

This paper is concerned with a result of homogenization of a non-local first order Hamilton–Jacobi equation describing the dislocations dynamics. Our model for the interaction between dislocations involves both an integro-differential operator and a (local) Hamiltonian depending periodicly on u/ε. The first two authors studied in a previous work homogenization problems involving such local Hamiltonians. Two main ideas of this previous work are used: on the one hand, we prove an ergodicity property of this equation by constructing approximate correctors which are necessarily non periodic in space in general; on the other hand, the proof of the convergence of the solution uses here a twisted perturbed test function for a higher dimensional problem. The limit equation is a nonlinear diffusion equation involving a first order Lévy operator; the nonlinearity keeps memory of the short range interaction, while the Lévy operator keeps memory of long ones. The homogenized equation is a kind of effective plastic law for densities of dislocations moving in a single slip plane.     

An exponential matrix method for solving systems of linear differential equations

This paper presents an exponential matrix method for the solutions of systems of high‐order linear differential equations with variable coefficients. The problem is considered with the mixed conditions. On the basis of the method, the matrix forms of exponential functions and their derivatives are constructed, and then by substituting the collocation points into the matrix forms, the fundamental matrix equation is formed. This matrix equation corresponds to a system of linear algebraic equations. By solving this system, the unknown coefficients are determined and thus the approximate solutions are obtained. Also, an error estimation based on the residual functions is presented for the method. The approximate solutions are improved by using this error estimation. To demonstrate the efficiency of the method, some numerical examples are given and the comparisons are made with the results of other methods. Copyright © 2012 John Wiley & Sons, Ltd.     

Weighted Generalized Inverses,Oblique Projections,and Least-Squares Problems

A generalization with singular weights of Moore–Penrose generalized inverses of closed range operators in Hilbert spaces is studied using the notion of compatibility of subspaces and positive operators.     

A Depth Formula for Generic Singularities and Their Weak Normality

Let X be a smooth projective variety of dimension r and π:X → ?^m a generic projection with r + 1 ≤ m ≤ 2r. It is shown that, at any point on X′ = π(X) of multiplicity μ, off a closed subset of the triple locus of codimension four, the depth of the local ring is equal to r ? (μ ? 1)(m ? r ? 1). This leads to some improvements on the affirmation of a conjecture of Andreotti–Bombieri–Holm on the weak normality of X′ and a conjecture of Piene on the weak normality of Sing(X′).     

Development and testing of a CW-EPR apparatus for imaging of short-lifetime nitroxyl radicals in mouse head

This article describes a method for reducing the acquisition time in three-dimensional (3D) continuous-wave electron paramagnetic resonance (CW-EPR) imaging. To visualize nitroxyl spin probes, which have a short lifetime in living organisms, the acquisition time for a data set of spectral projections should be shorter than the lifetime of the spin probes. To decrease the total time required for data acquisition, the duration of magnetic field scanning was reduced to 0.5 s. Moreover, the number of projections was decreased by using the concept of a uniform distribution. To demonstrate this faster data acquisition, two kinds of nitroxyl radicals with different decay rates were measured in mice. 3D EPR imaging of 4-hydroxy-2,2,6,6-tetramethylpiperidine-d₁₇-1-¹⁵N-1-oxyl in mouse head was successfully carried out. 3D EPR imaging of nitroxyl spin probes with a half-life of a few minutes was achieved for the first time in live animals.