A globally and locally superlinearly convergent inexact Newton-GMRES method for large-scale variational inequality problem

In this paper, we propose an inexact Newton-generalized minimal residual method for solving the variational inequality problem. Based on a new smoothing function, the variational inequality problem is reformulated as a system of parameterized smooth equations. In each iteration, the corresponding linear system is solved only approximately. Under mild assumptions, it is proved that the proposed algorithm has global convergence and local superlinear convergence properties. Preliminary numerical results indicate that the method is effective for a large-scale variational inequality problem.     

On numerical approximation of a variational–hemivariational inequality modeling contact problems for locking materials

This paper is devoted to numerical analysis of a new class of elliptic variational–hemivariational inequalities in the study of a family of contact problems for elastic ideally locking materials. The contact is described by the Signorini unilateral contact condition and the friction is modeled by a nonmonotone multivalued subdifferential relation allowing slip dependence. The problem involves a nonlinear elasticity operator, the subdifferential of the indicator function of a convex set for the locking constraints and a nonconvex locally Lipschitz friction potential. Solution existence and uniqueness result on the inequality can be found in Migórski and Ogorzaly (2017) . In this paper, we introduce and analyze a finite element method to solve the variational–hemivariational inequality. We derive a Céa type inequality that serves as a starting point of error estimation. Numerical results are reported, showing the performance of the numerical method.     

A multigrid optimization algorithm for the numerical solution of quasilinear variational inequalities involving the p-Laplacian

In this paper we propose a multigrid optimization algorithm (MG/OPT) for the numerical solution of a class of quasilinear variational inequalities of the second kind. This approach is enabled by the fact that the solution of the variational inequality is given by the minimizer of a nonsmooth energy functional, involving the $p$-Laplace operator. We propose a Huber regularization of the functional and a finite element discretization for the problem. Further, we analyze the regularity of the discretized energy functional, and we are able to prove that its Jacobian is slantly differentiable. This regularity property is useful to analyze the convergence of the MG/OPT algorithm. In fact, we demonstrate that the algorithm is globally convergent by using a mean value theorem for semismooth functions. Finally, we apply the MG/OPT algorithm to the numerical simulation of the viscoplastic flow of Bingham, Casson and Herschel–Bulkley fluids in a pipe. Several experiments are carried out to show the efficiency of the proposed algorithm when solving this kind of fluid mechanics problems.     

Dynamic Network User Equilibrium with State-Dependent Time Lags

This paper recasts the Friesz et al. (1993) measure theoretic model of dynamic network user equibrium as a controlled variational inequality problem involving Riemann integrals. This restatement is done to make the model and its foundations accessible to a wider audience by removing the need to have a background in functional analysis. Our exposition is dependent on previously unavailable necessary conditions for optimal control problems with state-dependent time lags. These necessary conditions, derived in an Appendix, are employed to show that a particular variational inequality control problem has solutions that are dynamic network user equilibria. Our analysis also shows that use of proper flow propagation constraints obviates the need to explicitly employ the arc exit time functions that have complicated numerical implementations of the Friesz et al. (1993) model heretofore. We close by describing the computational implications of numerically determining dynamic user equilibria from formulations based on state-dependent time lags.     

Numerical analysis of a strain-adaptive bone remodelling problem

In this paper we study from the numerical point of view a strain-adaptive bone remodelling that couples the displacements and the apparent density (the porosity) of the bone. The rate of this density at a particular location is described as an objective function, which depends on a particular stimulus at that location. The variational problem is written as a coupled system of a nonlinear variational equation for the displacement field and a nonlinear parabolic variational inequality for the apparent density. Then, fully discrete approximations are provided by using the finite element method to approximate the spatial variable and an Euler scheme to discretize the time derivatives. Error estimates are proved, from which, under adequate regularity conditions, the linear convergence of the algorithm is deduced. Finally, some numerical simulations involving one- and two-dimensional test examples are presented to demonstrate the accuracy of the approximation and the behaviour of the solution.     

面向全变分图像复原的增广拉格朗日方法综述

图像复原旨在根据退化图像重建高品质原始图像,其复原的质量和速度问题一直都是图像处理领域研究的重要方向。由于其图像边缘保持特性,全变分（TV）最小化模型在图像复原领域取得了很大的成功。然而,全变分图像复原是一个典型的非光滑优化问题,需要发展相应的快速优化算法,而增广拉格朗日方法（ALM）则是近年来发展起来的一类代表性方法。结合相关进展,综述了全变分图像复原模型,变量分裂（VS）法和典型ALM算法,并通过实验从CPU运行时间、峰值信噪比（PSNR）和品质评价等方面分析了不同的变量分裂和ALM方法对图像复原性能的影响。     

A variational method for unconstrained optimization problems

We introduce a new method for solving an unconstrained optimization problem. The method is based on the solution of a variational inequality problem and may be considered a modified trust region algorithm. We prove the convergence of the method and present some numerical results.     

Multibody Dynamic Response Optimization with ALM and Approximate Line Search

This paper presents an efficient approach for dynamic responseoptimization based on the ALM method. In this approach, an approximateaugmented Lagrangian is employed for line searches while an exactaugmented Lagrangian is used for finding search directions. An importantfeature of this study is that the approximate augmented Lagrangian forline search is composed of the linearized cost and constraint functionsprojected on the search direction. The quality of this approximationshould be good since an approximate penalty term is found to have almostsecond-order accuracy near the optimum. Quasi-Newton and conjugategradient algorithms are used to find exact search directions and a goldensection method followed by a cubic polynomial approximation is employedfor line search. The numerical performance of the proposed approach isinvestigated by solving eight typical dynamic response optimizationproblems and comparing the results with those in the literature. Thiscomparison shows that the suggested approach is robust and efficient.     

An augmented Lagrangian optimization method for inflatable structures analysis problems

This paper describes the development of an augmented Lagrangian optimization method for the numerical simulation of the inflation process in the design of inflatable space structures. Although the Newton–Raphson scheme was proven to be efficient for solving many nonlinear problems, it can lead to lack of convergence when it is applied to the simulation of the inflation process. As a result, it is recommended to use an optimization algorithm to find the minimum energy configuration that satisfies the equilibrium equations characterizing the final shape of the inflated structure subject to an internal pressure. On top of that, given that some degrees of freedom may be linked, the optimum may be constrained, and specific optimization methods for constrained problems must be considered. The paper presents the formulation and the augmented Lagrangian method (ALM) developed in SAMCEF Mecano for inflatable structures analysis problems. The related quasi-unconstrained optimization problem is solved with a nonlinear conjugate gradient method. The Wolfe conditions are used in conjunction with a cubic interpolation for the line search. Equality constraints are considered and can be easily treated by the ALM formulation. Numerical applications present simulations of unconstrained and constrained inflation processes (i.e., where the motion of some nodes is ruled by a rigid body element restriction and/or problems including contact conditions).Part of this paper was presented at the sixth world congress of Structural and Multidisciplinary Optimization held in Rio de Janeiro, June 2005.     

Structured eigenvalue methods for the computation of corner singularities in 3D anisotropic elastic structures

This paper is concerned with the computation of three-dimensional vertex singularities of anisotropic elastic fields. The singularities are described by eigenpairs of a corresponding operator pencil on a subdomain of the sphere. The solution approach is to introduce a modified quadratic variational boundary eigenvalue problem which consists of two self-adjoint, positive definite sesquilinear forms and a skew-Hermitian form. This eigenvalue problem is discretized by the finite element method. The resulting quadratic matrix eigenvalue problem is then solved with the skew Hamiltonian implicitly restarted Arnoldi method which is specifically adapted to the structure of this problem. Some numerical examples are given that show the performance of this approach.     

A symmetric variational finite element-boundary integral equation coupling method

This paper is concerned with the development of a mixed variational principle for coupling finite element and boundary integral methods in interface problems, using the generalized Poisson's equation as a prototype situation. One of its primary objectives is to compare the performance of fully variational procedures with methods that use collocation for the treatment of boundary integral equations. A distinctive feature of the new variational principle is that the discretized algebraic equations for the coupled problem are automatically symmetric since they are all derived from a single functional. In addition, the condition that the flux remain continuous across interfaces is satisfied naturally. In discretizing the problem within inhomogeneous or loaded regions, domain finite elements are used to approximate the field variable. On the other hand, only boundary elements are used for regions where the medium is homogeneous and free of external agents. The corresponding integral equations are discretized both by fully variational and by collocation techniques. Results of numerical experiments indicate that the accuracy of the fully variational procedure is significantly greater than that of collocation for the complete interface problem, especially for complex disturbances, at little additional computational cost. This suggests that fully variational procedures may be preferable to collocation, not only in dealing with interface problems, but even for solving integral equations by themselves.     

Rigid body attitude estimation based on the Lagrange–d’Alembert principle

Estimation of rigid body attitude and angular velocity without any knowledge of the attitude dynamics model is treated using the Lagrange–d’Alembert principle from variational mechanics. It is shown that Wahba’s cost function for attitude determination from two or more non-collinear vector measurements can be generalized and represented as a Morse function of the attitude estimation error on the Lie group of rigid body rotations. With body-fixed sensor measurements of direction vectors and angular velocity, a Lagrangian is obtained as the difference between a kinetic energy-like term that is quadratic in the angular velocity estimation error and an artificial potential obtained from Wahba’s function. An additional dissipation term that depends on the angular velocity estimation error is introduced, and the Lagrange–d’Alembert principle is applied to the Lagrangian with this dissipation. A Lyapunov analysis shows that the state estimation scheme so obtained provides stable asymptotic convergence of state estimates to actual states in the absence of measurement noise, with an almost global domain of attraction. These estimation schemes are discretized for computer implementation using discrete variational mechanics. A first order Lie group variational integrator is obtained as a discrete-time implementation. In the presence of bounded measurement noise, numerical simulations show that the estimated states converge to a bounded neighborhood of the actual states.     

Identifying the coefficient of first-order in parabolic equation from final measurement data

This paper studies an inverse problem of recovering the first-order coefficient in parabolic equation when the final observation is given. Such problem has important application in a large field of applied science. The original problem is transformed into an optimal control problem by the optimization theory. The existence, uniqueness and necessary condition of the minimum for the control functional are established. By an elliptic bilateral variational inequality which is deduced from the necessary condition, an algorithm and some numerical experiments are proposed in the paper. The numerical results show that the proposed method is an accurate and stable method to determine the coefficient of first-order in the inverse parabolic problems.     

An immersed boundary technique for simulating complex flows with rigid boundary

A new immersed boundary (IB) technique for the simulation of flow interacting with solid boundary is presented. The present formulation employs a mixture of Eulerian and Lagrangian variables, where the solid boundary is represented by discrete Lagrangian markers embedding in and exerting forces to the Eulerian fluid domain. The interactions between the Lagrangian markers and the fluid variables are linked by a simple discretized delta function. The numerical integration is based on a second-order fractional step method under the staggered grid spatial framework. Based on the direct momentum forcing on the Eulerian grids, a new force formulation on the Lagrangian marker is proposed, which ensures the satisfaction of the no-slip boundary condition on the immersed boundary in the intermediate time step. This forcing procedure involves solving a banded linear system of equations whose unknowns consist of the boundary forces on the Lagrangian markers; thus, the order of the unknowns is one-dimensional lower than the fluid variables. Numerical experiments show that the stability limit is not altered by the proposed force formulation, though the second-order accuracy of the adopted numerical scheme is degraded to 1.5 order. Four different test problems are simulated using the present technique (rotating ring flow, lid-driven cavity and flows over a stationary cylinder and an in-line oscillating cylinder), and the results are compared with previous experimental and numerical results. The numerical evidences show the accuracy and the capability of the proposed method for solving complex geometry flow problems both with stationary and moving boundaries.     

Hybrid Discretization Methods with Adaptive Yield Surface Detection for Bingham Pipe Flows

We devise a hybrid low-order method for Bingham pipe flows, where the velocity is discretized by means of one unknown per mesh face and one unknown per mesh cell which can be eliminated locally by static condensation. The main advantages are local conservativity and the possibility to use polygonal/polyhedral meshes. We exploit this feature in the context of adaptive mesh refinement to capture the yield surface by means of local mesh refinement and possible coarsening. We consider the augmented Lagrangian method to solve iteratively the variational inequalities resulting from the discrete Bingham problem, using piecewise constant fields for the auxiliary variable and the associated Lagrange multiplier. Numerical results are presented in pipes with circular and eccentric annulus cross-section for different Bingham numbers.     

A novel neural network for nonlinear convex programming

In this paper, we present a neural network for solving the nonlinear convex programming problem in real time by means of the projection method. The main idea is to convert the convex programming problem into a variational inequality problem. Then a dynamical system and a convex energy function are constructed for resulting variational inequality problem. It is shown that the proposed neural network is stable in the sense of Lyapunov and can converge to an exact optimal solution of the original problem. Compared with the existing neural networks for solving the nonlinear convex programming problem, the proposed neural network has no Lipschitz condition, no adjustable parameter, and its structure is simple. The validity and transient behavior of the proposed neural network are demonstrated by some simulation results.     

基于路段元胞传输模型的动态用户最优配流问题

利用基于路段的元胞传输模型进行模拟, 给出了一种计算实际路段出行阻抗的方法, 并在此基础上构造了基于路段变量的动态用户最优变分不等式模型. 模型采用针对迄节点的路段变量, 在每一个小时段都能给出路段流入率、流出率、路段流量和实际路段阻抗, 为用户提供较为全面的诱导信息打下了较好的理论基础. 采用了修正投影算法来进行求解. 数值算例表明模型具有的实用性和优越性, 使道路交通流宏观模型与动态网络交通配流问题得到较好的结合.     

Numerical solutions for system of third-order boundary value problems

We develop a numerical method for computing approximations for the solutions of a system of third order boundary value problems associated with odd order obstacle problems. Such a problem arise in physical oceanography and can be studied in the framework of variational inequality theory. We study the convergence analysis of the present method and we show that it gives numerical results which are better than the other available results. Numerical example is presented to illustrate the applicability of the new method.     

FETI based algorithms for contact problems: scalability,large displacements and 3D Coulomb friction

Theoretical and experimental results concerning FETI based algorithms for contact problems of elasticity are reviewed. A discretized model problem is first reduced by the duality theory of convex optimization to the quadratic programming problem with bound and equality constraints. The latter is then optionally modified by means of orthogonal projectors to the natural coarse space introduced by Farhat and Roux in the framework of their FETI method. The resulting problem is then solved either by special algorithms for bound constrained quadratic programming problems combined with penalty that imposes the equality constraints, or by an augmented Lagrangian type algorithm with the inner loop for the solution of bound constrained quadratic programming problems. Recent theoretical results are reported that guarantee certain optimality and scalability of both algorithms. The results are confirmed by numerical experiments. The performance of the algorithm in solution of more realistic engineering problems by basic algorithm is demonstrated on the solution of 3D problems with large displacements or Coulomb friction.