A trust-region and affine scaling algorithm for linearly constrained optimization

A new trust-region and affine scaling algorithm for linearly constrained optimization is presented in this paper. Under no nondegenerate assumption, we prove that any limit point of the sequence generated by the new algorithm satisfies the first order necessary condition and there exists at least one limit point of the sequence which satisfies the second order necessary condition. Some preliminary numerical experiments are reported. The work was done while visiting Institute of Applied Mathematics, AMSS, CAS.     

Superlinear convergence of affine scaling interior point newton method for linear inequality constrained minimization without strict complementarity

In this paper we extend and improve the classical affine scaling interior-point Newton method for solving nonlinear optimization subject to linear inequality constraints in the absence of the strict complementarity assumption. Introducing a computationally efficient technique and employing an identification function for the definition of the new affine scaling matrix, we propose and analyze a new affine scaling interior-point Newton method which improves the Coleman and Li affine sealing matrix in [2] for solving the linear inequlity constrained optimization. Local superlinear and quadratical convergence of the proposed algorithm is established under the strong second order sufficiency condition without assuming strict complementarity of the solution.     

A trust region and affine scaling interior point method for nonconvex minimization with linear inequality constraints

A trust region and affine scaling interior point method (TRAM) is proposed for a general nonlinear minimization with linear inequality constraints [8]. In the proposed approach, a Newton step is derived from the complementarity conditions. Based on this Newton step, a trust region subproblem is formed, and the original objective function is monotonically decreased. Explicit sufficient decrease conditions are proposed for satisfying the first order and second order necessary conditions.?The objective of this paper is to establish global and local convergence properties of the proposed trust region and affine scaling interior point method. It is shown that the proposed explicit decrease conditions are sufficient for satisfy complementarity, dual feasibility and second order necessary conditions respectively. It is also established that a trust region solution is asymptotically in the interior of the proposed trust region subproblem and a properly damped trust region step can achieve quadratic convergence. Received: January 29, 1999 / Accepted: November 22, 1999?Published online February 23, 2000     

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.     

仿射变换内点Levenberg-Marquardt法解KKT系统

提供了一类新的结合非单调内点回代线搜索技术的仿射变换Levenberg-Marquardt法解Karush-Kuhn-Tucker（KKT）系统. 基于由KKT系统转化得到的等价的部分变量具有非负约束的最小化问题,建立了Levenberg-Marquardt方程. 证明了算法不仅具有整体收敛性,而且在合理的假设条件下,算法具有超线性收敛速率. 数值结果验证了算法的实际有效性.     

A filter-trust-region method for <Emphasis Type="Italic">LC</Emphasis><Superscript>1</Superscript> unconstrained optimization and its global convergence

In this paper we present a filter-trust-region algorithm for solving LC1 unconstrained optimization problems which uses the second Dini upper directional derivative.We establish the global convergence of the algorithm under reasonable assumptions.     

Global and local convergence of a new affine scaling trust region algorithm for linearly constrained optimization

Chen and Zhang [Sci.China,Ser.A,45,1390–1397(2002)] introduced an affine scaling trust region algorithm for linearly constrained optimization and analyzed its global convergence.In this paper,we derive a new affine scaling trust region algorithm with dwindling filter for linearly constrained optimization.Different from Chen and Zhang's work,the trial points generated by the new algorithm are accepted if they improve the objective function or improve the first order necessary optimality conditions.Under mild conditions,we discuss both the global and local convergence of the new algorithm.Preliminary numerical results are reported.     

INTERIOR POINT PROJECTED REDUCED HESSIAN METHOD WITH TRUST REGION STRATEGY FOR NONLINEAR CONSTRAINED OPTIMIZATION

§1　IntroductionIn this paper we analyze an interior point scaling projected reduced Hessian methodwith trust region strategy for solving the nonlinear equality constrained optimizationproblem with nonnegative constraints on variables:min　f(x)s.t.　c(x) =0 (1.1)x≥0where f∶Rn→R is the smooth nonlinear function,notnecessarily convex and c(x)∶Rn→Rm(m≤n) is the vector nonlinear function.There are quite a few articles proposing localsequential quadratic programming reduced Hessian methods…     

An Adaptive Mesh-Refining Algorithm Allowing for an <Emphasis Type="Italic">H</Emphasis><Superscript>1</Superscript> Stable <Emphasis Type="Italic">L</Emphasis><Superscript>2</Superscript> Projection onto Courant Finite Element Spaces

Suppose $\cal{S}^1({\cal T})\subset H^1(\Omega)$ is the $P_1$-finite element space of $\cal{T}$-piecewise affine functions based on a regular triangulation $\cal{T}$ of a two-dimensional surface $\Omega$ into triangles. The $L^2$ projection $\Pi$ onto $\cal{S}^1(\cal{T})$ is $H^1$ stable if $\norm{\Pi v}{H^1(\Omega)}\le C\norm{v}{H^1(\Omega)}$ for all $v$ in the Sobolev space $H^1(\Omega)$ and if the bound $C$ does not depend on the mesh-size in $\cal{T}$ or on the dimension of $\cal{S}^1(\cal{T})$. \hskip 1em A red–green–blue refining adaptive algorithm is designed which refines a coarse mesh $\cal{T}_0$ successively such that each triangle is divided into one, two, three, or four subtriangles. This is the newest vertex bisection supplemented with possible red refinements based on a careful initialization. The resulting finite element space allows for an $H^1$ stable $L^2$ projection. The stability bound $C$ depends only on the coarse mesh $\cal{T}_0$ through the number of unknowns, the shapes of the triangles in $\cal{T}_0$, and possible Dirichlet boundary conditions. Our arguments also provide a discrete version $\norm{h_\cal{T}^{-1}\,\Pi v}{L^2(\Omega)}\le C\norm{h_\cal{T}^{-1}\,v}{L^2(\Omega)}$ in $L^2$ norms weighted with the mesh-size $h_\T$.     

New examples of Cantor sets in <Emphasis Type="Italic">S</Emphasis><Superscript>1</Superscript> that are not <Emphasis Type="Italic">C</Emphasis><Superscript>1</Superscript>-minimal

Although every Cantor subset of the circle (S¹) is the minimal set of some homeomorphism of S¹, not every such set is minimal for a C¹ diffeomorphism of S¹. In this work, we construct new examples of Cantor sets in S¹ that are not minimal for any C¹-diffeomorphim of S¹.     

Regular interval Cantor sets of <Emphasis Type="Italic">S</Emphasis><Superscript>1</Superscript> and minimality

It is known that not every Cantor set of S ¹ is C ¹-minimal. In this work we prove that every member of a subfamily of what we here call regular interval Cantor set is not C ¹-minimal. We also prove that no member of a class of Cantor sets that includes this subfamily is C ^1+∈-minimal, for any ∈ > 0. Partially supported by CNPq-Brasil and PEDECIBA-Uruguay.     

Nonconforming <Emphasis Type="Italic">H</Emphasis><Superscript>1</Superscript>-Galerkin mixed FEM for Sobolev equations on anisotropic meshes

A nonconforming H^1-Calerkin mixed finite element method is analyzed for Sobolev equations on anisotropic meshes. The error estimates are obtained without using Ritz-Volterra projection.     

Geometric interpolation by planar cubic <Emphasis Type="Italic">G</Emphasis><Superscript>1</Superscript> splines

In this paper, geometric interpolation by G ¹ cubic spline is studied. A wide class of sufficient conditions that admit a G ¹ cubic spline interpolant is determined. In particular, convex data as well as data with inflection points are included. The existence requirements are based upon geometric properties of data entirely, and can be easily verified in advance. The algorithm that carries out the verification is added. AMS subject classification (2000)  65D05, 65D07, 65D17     

<Emphasis Type="Italic">W</Emphasis><Superscript>1,<Emphasis Type="Italic">p</Emphasis></Superscript>(R<Superscript><Emphasis Type="Italic">n</Emphasis></Superscript>)-boundedness for maximal functions

In this paper, we get W ^1,p (R ⁿ )-boundedness for tangential maximal function and nontangential maximal function, which improves J.Kinnunen, P.Lindqvist and Tananka’s results. Supported by the key Academic Discipline of Zhejiang Province of China under Grant No.2005 and the Zhejiang Provincial Natural Science Foundation of China.     

Properties of the <Emphasis Type="Italic">C</Emphasis><Superscript>1</Superscript>-smooth functions with nowhere dense gradient range

One of the main results of the present article is as follows Theorem. Let v: Ω → ? be a C¹-smooth function on a domain Ω ? ?². Suppose that Int?v(Ω) = ?. Then, for every point z ∈ Ω, there is a straight line L ? z such that ?v ≡ const on the connected component of the set L ? Ω containing z.Also, we prove that, under the conditions of the theorem, the range of the gradient ?v(Ω) is locally a curve and this curve has tangents in the weak sense and the direction of these tangents is a function of bounded variation.     

Fundamental groups of manifolds with <Emphasis Type="Italic">S</Emphasis><Superscript>1</Superscript>-category 2

A closed topological n-manifold M ⁿ is of S ¹-category 2 if it can be covered by two open subsets W ₁,W ₂ such that the inclusions W _i → M ⁿ factor homotopically through maps W _i → S ¹ → M ⁿ . We show that the fundamental group of such an n-manifold is a cyclic group or a free product of two cyclic groups with nontrivial amalgamation. In particular, if n = 3, the fundamental group is cyclic.      

Necessary and sufficient conditions for a curve to be the gradient range of a <Emphasis Type="Italic">C</Emphasis><Superscript>1</Superscript>-smooth function

We find some necessary and sufficient conditions for a plane curve to be the gradient range of a C ¹-smooth function of two variables. As one of the consequences we give the necessary and sufficient conditions on a continuous function ? under which the differential equation \(\frac{{\partial v}}{{\partial t}} = \varphi \left( {\frac{{\partial v}}{{\partial x}}} \right)\) has nontrivial C ¹-smooth solutions.     

Large time behavior of solutions for critical and subcritical complex Ginzburg-Landau equations in H<Superscript>1</Superscript>

Considering the Cauchy problem for the critical complex Ginzburg-Landau equation in H¹(Rⁿ), we shall show the asymptotic behavior for its solutions in C(0, ∖;H¹(Rⁿ)) ∩ L²(0, ∖;H^1,2n/(n-2)(R²)), n≥3. Analogous results also hold in the case that the nonlinearity has the subcritical power in H¹(Rⁿ), n≥1. Dedicated to Professor Zhou Yulin for his 80th birthday.     

Non-linear elliptic problems having natural growth and <Emphasis Type="Italic">L</Emphasis><Superscript>1</Superscript> data in Orlicz spaces

An existence result for a strongly non-linear elliptic equation with natural growth condition on the non-linearity and L¹-data is proved. Mathematics Subject Classification (2000) 35J25, 35J60     

An example of a <Emphasis Type="Italic">C</Emphasis><Superscript>1</Superscript>-smooth function whose gradient range is an arc with no tangent at any point

We construct an example of a C ¹-smooth real function of two variables whose gradient range is an arc with no tangent at any point.