Tikhonov Regularization Methods for Variational Inequality Problems

Motivated by the work of Facchinei and Kanzow (Ref. 1) on regularization methods for the nonlinear complementarity problem and the work of Ravindran and Gowda (Ref. 2) for the box variational inequality problem, we study regularization methods for the general variational inequality problem. A sufficient condition is given which guarantees that the union of the solution sets of the regularized problems is nonempty and bounded. It is shown that solutions of the regularized problems form a minimizing sequence of the D-gap function under a mild condition.     

Macro-Hybrid Variational Formulations of Constrained Boundary Value Problems

In the context of convex analysis, macro-hybrid variational formulations of constrained boundary value problems are presented. Monotone mixed variational inclusions are macro-hybridized on the basis of nonoverlapping domain decompositions, and corresponding three-field versions are derived. Then, for regularization purposes, augmented formulations are established via preconditioned exact penalizations and expressed in terms of proximation operators. Optimization interpretations are given for potential problems, recovering the classic two- and three-field augmented Lagrangian formulations. Furthermore, associated parallel two- and three-field proximal-point algorithms are discussed for numerical resolution of finite element discretizations. Applications to dual mixed variational formulations of problems from mechanics illustrate the theory.     

y-Regularization-Operator Splitting Method for the Numerical Solution of a Scalar Eikonal Equation

In this article, we discuss a numerical method for the computation of the minimal and maximal solutions of a steady scalar Eikonal equation. This method relies on a penalty treatment of the nonlinearity, a biharmonic regularization of the resulting variational problem, and the time discretization by operator-splitting of an initial value problem associated with the Euler-Lagrange equations of the regularized variational problem. A low-order finite element discretization is advocated since it is well-suited to the low regularity of the solutions. Numerical experiments show that the method sketched above can capture efficiently the extremal solutions of various two-dimensional test problems and that it has also the ability of handling easily domains with curved boundaries.     

Variational Bayesian Generative Topographic Mapping

General finite mixture models are powerful tools for the density-based grouping of multivariate i.i.d. data, but they lack data visualization capabilities, which reduces their practical applicability to real-world problems. Generative topographic mapping (GTM) was originally formulated as a constrained mixture of distributions in order to provide simultaneous visualization and clustering of multivariate data. In its inception, the adaptive parameters were determined by maximum likelihood (ML), using the expectation-maximization (EM) algorithm. The original GTM is, therefore, prone to data overfitting unless a regularization mechanism is included. In this paper, we define an alternative variational formulation of GTM that provides a full Bayesian treatment to a Gaussian process (GP)-based variation of the model. The generalization capabilities of the proposed Variational Bayesian GTM are assessed in some detail and compared with those of alternative GTM regularization approaches in terms of test log-likelihood, using several artificial and real datasets.     

Lagrange multipliers, (exact) regularization and error bounds for monotone variational inequalities

We examine two central regularization strategies for monotone variational inequalities, the first a direct regularization of the operative monotone mapping, and the second via regularization of the associated dual gap function. A key link in the relationship between the solution sets to these various regularized problems is the idea of exact regularization, which, in turn, is fundamentally associated with the existence of Lagrange multipliers for the regularized variational inequality. A regularization is said to be exact if a solution to the regularized problem is a solution to the unregularized problem for all parameters beyond a certain value. The Lagrange multipliers corresponding to a particular regularization of a variational inequality, on the other hand, are defined via the dual gap function. Our analysis suggests various conceptual, iteratively regularized numerical schemes, for which we provide error bounds, and hence stopping criteria, under the additional assumption that the solution set to the unregularized problem is what we call weakly sharp of order greater than one.     

Relaxation of Nonlocal Singular Integrals

In this paper we study well-posedness of a class of nonconvex variational principles arising in regularization theory for denoising of data with sampling errors and level set regularization methods for inverse problems. These models result in minimization of nonconvex, singular functionals involving (possibly) non-local operators.     

Augmented Lagrangian Homotopy Method for the Regularization of Total Variation Denoising Problems

This paper presents a homotopy procedure which improves the solvability of mathematical programming problems arising from total variational methods for image denoising. The homotopy on the regularization parameter involves solving a sequence of equality-constrained optimization problems where the positive regularization parameter in each optimization problem is initially large and is reduced to zero. Newton’s method is used to solve the optimization problems and numerical results are presented.     

On the Numerical Solution of Some Eikonal Equations: An Elliptic Solver Approach

The steady Eikonal equation is a prototypical first-order fully nonlinear equation. A numerical method based on elliptic solvers is presented here to solve two different kinds of steady Eikonal equations and compute solutions, which are maximal and minimal in the variational sense. The approach in this paper relies on a variational argument involving penalty, a biharmonic regularization, and an operator-splitting-based time-discretization scheme for the solution of an associated initial-value problem. This approach allows the decoupling of the nonlinearities and differential operators. Numerical experiments are performed to validate this approach and investigate its convergence properties from a numerical viewpoint.     

Progressive Regularization of Variational Inequalities and Decomposition Algorithms

For nonsymmetric operators involved in variational inequalities, the strong monotonicity of their possibly multivalued inverse operators (referred to as the Dunn property) appears to be the weakest requirement to ensure convergence of most iterative algorithms of resolution proposed in the literature. This implies the Lipschitz property, and both properties are equivalent for symmetric operators. For Lipschitz operators, the Dunn property is weaker than strong monotonicity, but is stronger than simple monotonicity. Moreover, it is always enforced by the Moreau–Yosida regularization and it is satisfied by the resolvents of monotone operators. Therefore, algorithms should always be applied to this regularized version or they should use resolvents: in a sense, this is what is achieved in proximal and splitting methods among others. However, the operation of regularization itself or the computation of resolvents may be as complex as solving the original variational inequality. In this paper, the concept of progressive regularization is introduced and a convergent algorithm is proposed for solving variational inequalities involving nonsymmetric monotone operators. Essentially, the idea is to use the auxiliary problem principle to perform the regularization operation and, at the same time, to solve the variational inequality in its approximately regularized version; thus, two iteration processes are performed simultaneously, instead of being nested in each other, yielding a global explicit iterative scheme. Parallel and sequential versions of the algorithm are presented. A simple numerical example demonstrates the behavior of these two versions for the case where previously proposed algorithms fail to converge unless regularization or computation of a resolvent is performed at each iteration. Since the auxiliary problem principle is a general framework to obtain decomposition methods, the results presented here extend the class of problems for which decomposition methods can be used.     

The Tikhonov regularization extended to equilibrium problems involving pseudomonotone bifunctions

We extend the Tikhonov regularization method widely used in optimization and monotone variational inequality studies to equilibrium problems. It is shown that the convergence results obtained from the monotone variational inequality remain valid for the monotone equilibrium problem. For pseudomonotone equilibrium problems, the Tikhonov regularized subproblems have a unique solution only in the limit, but any Tikhonov trajectory tends to the solution of the original problem, which is the unique solution of the strongly monotone equilibrium problem defined on the basis of the regularization bifunction.     

Composition Duality Methods for Mixed Variational Inclusions

Composition duality methods for mixed variational inclusions are studied in a functional framework of reflexive Banach spaces. On the basis of duality principles, the solvability of maximal monotone and subdifferential mixed variational inclusions is established. For computational purposes, mass-preconditioned augmented formulations are introduced for regularization, as well as three-field and macro-hybrid variational versions. At a finite-dimensional level, corresponding discrete mixed and macro-hybrid internal approximations are discussed, as well as proximal-point iterative algorithms. Primal and dual mixed variational inclusions from contact mechanics illustrate the theory.     

解一类结构变分不等式问题的非精确并行交替方向法

带线性约束的具有两分块结构的单调变分不等式问题, 出现在许多现代应用中, 如交通和经济问题等. 基于该问题良好的可分结构, 分裂型算法被广泛研究用于其求解. 提出新的带回代的非精确并行交替方向法解该类问题, 在每一步迭代中,首先以并行模式通过投影得到预测点, 然后对其校正得到下一步的迭代点. 在压缩型算法的理论框架下, 在适当条件下证明了所提算法的全局收敛性. 数值结果表明了算法的有效性. 此外, 该算法可推广到求解具有多分块结构的问题.     

具不等式约束变分不等式的信赖域算法

1　引　　言令Ｘ是Ｒｎ 中的非空闭凸集 ,Ｆ :Ｘ→Ｒｎ 是连续映射 ,〈· ,·〉表示Ｒｎ 中的内积 有限维变分不等式问题 (以下简称变分不等式问题 ,记为ＶＩＰ或ＶＩ(Ｘ ,Ｆ) ) :就是求ｘ ∈Ｒｎ,使ｘ ∈Ｘ且 ｘ ∈Ｘ ,〈Ｆ(ｘ ) ,ｘ -ｘ 〉≥ 0 . ( 1 )在Ｘ =Ｒｎ+ 的特殊情形下 ,( 1 )变为非线性互补问题 (记为ＮＣＰ或ＮＣＰ(Ｆ) ) :就是求ｘ ∈Ｒｎ,使ｘ ≥ 0 ,Ｆ(ｘ ) ≥ 0 ,且〈ｘ ,Ｆ(ｘ )〉 =0 . ( 2 )　　变分不等式长期以来一直用于阐述和研究经济学、控制论、交通运输等领域中出现的各种平衡模型 近二十年来 ,变分不等式及其…     

A General Iterative Method for Solving Constrained Convex Minimization Problems

It is well known that the gradient-projection algorithm plays an important role in solving minimization problems. In this paper, we will use the idea of regularization to establish a general method so that the sequence generated by the general method can be strongly convergent to a minimizer of constrained convex minimization problems, which solves a variational inequality under suitable conditions.     

A regularized point cloud registration approach for orthogonal transformations

An important part of the well-known iterative closest point algorithm (ICP) is the variational problem. Several variants of the variational problem are known, such as point-to-point, point-to-plane, generalized ICP, and normal ICP (NICP). This paper proposes a closed-form exact solution for orthogonal registration of point clouds based on the generalized point-to-point ICP algorithm. We use points and normal vectors to align 3D point clouds, while the common point-to-point approach uses only the coordinates of points. The paper also presents a closed-form approximate solution to the variational problem of the NICP. In addition, the paper introduces a regularization approach and proposes reliable algorithms for solving variational problems using closed-form solutions. The performance of the algorithms is compared with that of common algorithms for solving variational problems of the ICP algorithm. The proposed paper is significantly extended version of Makovetskii et al. (CCIS 1090, 217–231, 2019).

     

Regularization of Nonmonotone Variational Inequalities

In this paper we extend the Tikhonov-Browder regularization scheme from monotone to rather a general class of nonmonotone multivalued variational inequalities. We show that their convergence conditions hold for some classes of perfectly and nonperfectly competitive economic equilibrium problems.     

Regularization of quasi-variational inequalities

An ill-posed quasi-variational inequality with contaminated data can be stabilized by employing the elliptic regularization. Under suitable conditions, a sequence of bounded regularized solutions converges strongly to a solution of the original quasi-variational inequality. Moreover, the conditions that ensure the boundedness of regularized solutions, become sufficient solvability conditions. It turns out that the regularization theory is quite strong for quasi-variational inequalities with set-valued monotone maps but restrictive for generalized pseudo-monotone maps. The results are quite general and are applicable to ill-posed variational inequalities, hemi-variational inequalities, inverse problems, and split feasibility problem, among others.     

Regularization in the Mosolov and Myasnikov problem with boundary friction

We propose an iterative algorithm for solving a semicoercive nonsmooth variational inequality. The algorithm is based on the stepwise partial smoothing of the minimized functional and an iterative proximal regularization method.We obtain a solution to the variational Mosolov and Myasnikov problem with boundary friction as a limit point of a sequence of solutions to stable auxiliary problems.     

THE KACANOV METHOD FOR A NONLINEAR VARIATIONAL INEQUALITY OF THE SECOND KIND ARISING IN ELASTOPLASTICITY

The authors first prove a convergence result on the Ka(?)anov method for solving generalnonlinear variational inequalities of the second kind and then apply the Kacanov method tosolve a nonlinear variational inequality of the second kind arising in elastoplasticity. In additionto the convergence result, an a posteriori error estimate is shown for the Kacanov iterates. Ineach step of the Ka(?)anov iteration, one has a (linear) variational inequality of the secondkind, which can be solved by using a regularization technique. The Ka(?)anov iteration andthe regularization technique together provide approximations which can be readily computednumerically. An a posteriori error estimate is derived for the combined effect of the Ka(?)anoviteration and the regularization.     

Iterative solution of linear equations with unbounded operators

A convergent iterative process is constructed for solving any solvable linear equation in a Hilbert space, including equations with unbounded, closed, densely defined linear operators. The method is proved to be stable towards small perturbation of the data. Some abstract results are established and used in an analysis of variational regularization method for equations with unbounded linear operators. The dynamical systems method (DSM) is justified for unbounded, closed, densely defined linear operators. The stopping time is chosen by a discrepancy principle. Equations with selfadjoint operators are considered separately. Numerical examples, illustrating the efficiency of the proposed method, are given.