Deren Han

A new inexact alternating directions method for monotone variational inequalities

Abstract.The alternating directions method (ADM) is an effective method for solving a class of variational inequalities (VI) when the proximal and penalty parameters in sub-VI problems are properly selected. In this paper, we propose a new ADM method which needs to solve two strongly monotone sub-VI problems in each iteration approximately and allows the parameters to vary from iteration to iteration. The convergence of the proposed ADM method is proved under quite mild assumptions and flexible parameter conditions.

A Note on the Alternating Direction Method of Multipliers

We consider the linearly constrained separable convex programming, whose objective function is separable into m individual convex functions without coupled variables. The alternating direction method of multipliers has been well studied in the literature for the special case m=2, while it remains open whether its convergence can be extended to the general case m≥3. This note shows the global convergence of this extension when the involved functions are further assumed to be strongly convex.

Solving non-additive traffic assignment problems: A descent method for co-coercive variational inequalities

Abstract This paper developed a descent direction of the merit function for co-coercive variational inequality (VI) problems. The descent approach is closely related to Fukushimas method for strongly monotone VI problems and Hes method for linear VI problems, and can be viewed as an extension for the more general case of co-coercive VI problems. This extension is important for route-based traffic assignment problems as the associated VI is often neither strongly monotone nor linear. This study then implemented the solution method for traffic assignment problems with non-additive route costs. Similar to projection-based methods, the computational effort required per iteration of this solution approach is modest. This is especially so for traffic equilibrium problems with elastic demand, where the solution method consists of a function evaluation and a simple projection onto the non-negative orthant.

Modified Goldstein-Levitin-Polyak projection method for asymmetric strongly monotone variational inequalities

In this paper, we present a modified Goldstein–Levitin–Polyak projection method for asymmetric strongly monotone variational inequality problems. A practical and robust stepsize choice strategy, termed self-adaptive procedure, is developed. The global convergence of the resulting algorithm is established under the same conditions used in the original projection method. Numerical results and comparison with some existing projection-type methods are given to illustrate the efficiency of the proposed method.

New alternating direction method for a class of nonlinear variational inequality problems

The alternating direction method is an attractive method for a class of variational inequality problems if the subproblems can be solved efficiently. However, solving the subproblems exactly is expensive even when the subproblem is strongly monotone or linear. To overcome this disadvantage, this paper develops a new alternating direction method for cocoercive nonlinear variational inequality problems. To illustrate the performance of this approach, we implement it for traffic assignment problems with fixed demand and for large-scale spatial price equilibrium problems.

Two new self-adaptive projection methods for variational inequality problems

Abstract In this paper, we propose two new projection methods for solving variational inequality problems (VI). The method is simple; it uses only function evaluation and projection onto the feasible set. Under the conditions that the underlying function is continuous and satisfies some generalized monotonicity assumption, the methods are proven to converge to a solution of variational inequality globally. Some preliminary computational results are reported to illustrate the efficiency of the methods.

A New Hybrid Generalized Proximal Point Algorithm for Variational Inequality Problems

In this paper, we propose a modified Bregman-function-based proximal point algorithm for solving variational inequality problems. The algorithm adopts a similar constructive approximate criterion as the one developed by Solodov and Svaiter (Set Valued Analysis 7 (1999) 323) for solving the classical proximal subproblems. Under some suitable conditions, we can get an approximate solution satisfying the accuracy criterion via a single Newton-type step. We obtain the Fejér monotonicity to solutions of VIP for paramonotone operators. Some preliminary computational results are also reported to illustrate the method.

A new stepsize rule in He and Zhou's alternating direction method

In this paper, we propose a new stepsize rule in He and Zhous alternating direction method. Under this new stepsize strategy, we extend their method for solving convex quadratic minimization problems to also monotone linear variational inequality problems.

A simple self-adaptive alternating direction method for linear variational inequality problems

In this study, we propose a new alternating direction method for solving linear variational variational inequality problems (LVIP). It is simple in the sense that, at each iteration, it needs only to perform a projection onto a simple set and some matrix-vector multiplications. The simplicity of the solution method makes it attractive for solving large-scale problems. To further improve its efficiency, we devise a self-adaptive strategy for choosing the necessary parameters of the solution procedure. We prove the global convergence of this new method under some mild conditions. Finally, some computational results are reported to demonstrate the properties and efficiency of the method.

A Parallel Splitting Method for Separable Convex Programs

In this paper, we propose a new parallel splitting augmented Lagrangian method for solving the nonlinear programs where the objective function is separable with three operators and the constraint is linear. The method is an improvement of the method of He (Comput. Optim. Appl., 2(42):195–212, 2009), where we generate a predictor using the same parallel splitting augmented Lagrangian scheme as that in He (Comput. Optim. Appl., 2(42):195–212, 2009), while adopting a new strategy to get the next iterate. Under the mild assumptions of convexity of the underlying mappings and the non-emptiness of the solution set, we prove that the proposed algorithm is globally convergent. We apply the new method in the area of image processing and to solve some quadratic programming problems. The preliminary numerical results indicate that the new method is efficient.