Logarithmic-quadratic Proximal Research Articles

Iteration complexity analysis of a partial LQP-based alternating direction method of multipliers

In this paper, we consider a prototypical convex optimization problem with multi-block variables and separable structures. By adding the Logarithmic Quadratic Proximal (LQP) regularizer with suitable proximal parameter to each of the first grouped subproblems, we develop a partial LQP-based Alternating Direction Method of Multipliers (ADMM-LQP). The dual variable is updated twice with relatively larger stepsizes than the classical region (0,1+52). Using a prediction-correction approach to analyze properties of the iterates generated by ADMM-LQP, we establish its global convergence and sublinear convergence rate of O(1/T) in the new ergodic and nonergodic senses, where T denotes the iteration index. We also extend the algorithm to a nonsmooth composite convex optimization and establish similar convergence results as our ADMM-LQP.

A self-adaptive descent LQP alternating direction method for the structured variational inequalities

In this paper, by combining the logarithmic-quadratic proximal (LQP) method and alternating direction method, we proposed an LQP alternating direction method for solving structured variational inequalities. The new iterate is generated by searching the optimal step size along a descent direction with a new step size αk. The choice of the descent direction and the step size selection strategies are important for the algorithm’s efficiency. The O(1/t) convergence rate of the proposed method is studied, and its efficiency is also verified by some numerical experiments.

An LQP-Based Symmetric Alternating Direction Method of Multipliers with Larger Step Sizes

Symmetric alternating direction method of multipliers (ADMM) is an efficient method for solving a class of separable convex optimization problems. This method updates the Lagrange multiplier twice with appropriate step sizes at each iteration. However, such step sizes were conservatively shrunk to guarantee the convergence in recent studies. In this paper, we are devoted to seeking larger step sizes whenever possible. The logarithmic-quadratic proximal (LQP) terms are applied to regularize the symmetric ADMM subproblems, allowing the constrained subproblems to then be converted to easier unconstrained ones. Theoretically, we prove the global convergence of such LQP-based symmetric ADMM by specifying a larger step size domain. Moreover, the numerical results on a traffic equilibrium problem are reported to demonstrate the advantage of the method with larger step sizes.

An LQP-Based Two-Step Method for Structured Variational Inequalities

The logarithmic quadratic proximal (LQP) regularization is a popular and powerful proximal regularization technique for solving monotone variational inequalities with nonnegative constraints. In this paper, we propose an implementable two-step method for solving structured variational inequality problems by combining LQP regularization and projection method. The proposed algorithm consists of two parts. The first step generates a pair of predictors via inexactly solving a system of nonlinear equations. Then, the second step updates the iterate via a simple correction step. We establish the global convergence of the new method under mild assumptions. To improve the numerical performance of our new method, we further present a self-adaptive version and implement it to solve a traffic equilibrium problem. The numerical results further demonstrate the efficiency of the proposed method.

A Strictly Contractive Peaceman-Rachford Splitting Method with Logarithmic-Quadratic Proximal Regularization for Convex Programming

Recently, a strictly contractive Peaceman-Rachford splitting method (PRSM) was proposed for a separable convex minimization model whose variables are subject to some linear constraints and two additional generic constraints. In general, the strictly contractive PRSM requires to solve two constrained minimization subproblems at each iteration. In this paper, we consider the case where the additional constraints on variables are positive orthants and apply the well-developed logarithmic-quadratic proximal (LQP) regularization to regularize the subproblems of the strictly contractive PRSM. A new algorithm in combination with the strictly contractive PRSM and the LQP regularization is thus proposed. The new algorithm only needs to solve two unconstrained subproblems at each iteration. An inexact version allowing the unconstrained subproblems to be solved approximately subject to certain inexactness criterion is also studied. We prove the global convergence and establish a worst-case convergence rate measured by the iteration complexity for both the exact and inexact versions of the new algorithm.

An inexact LQP alternating direction method for solving a class of structured variational inequalities

In this paper, we present an inexact alternating direction method. Each iteration of the proposed method contains a prediction and a correction, the predictor is obtained via solving the logarithmic-quadratic proximal (LQP) system approximately under significantly relaxed accuracy criterion and the new iterate is computed directly by an explicit formula derived from the original LQP method. Under certain conditions, the global convergence of the proposed method is proved.

An LQP-Based Decomposition Method for Solving a Class of Variational Inequalities

The alternating direction method (ADM) is an influential decomposition method for solving a class of variational inequalities with block-separable structures. In the literature, the subproblems of the ADM are usually regularized by quadratic proximal terms to ensure a more stable and attractive numerical performance. In this paper, we propose to apply the logarithmic-quadratic proximal (LQP) terms to regularize the ADM subproblems, and thus develop an LQP-based decomposition method for solving a class of variational inequalities. Global convergence of the new method is proved under standard assumptions.

An improved LQP-based method for solving nonlinear complementarity problems

The well-known logarithmic-quadratic proximal (LQP)method has motivated a number of efficient numerical algorithms for solving nonlinear complementarity problems (NCPs). In this paper,we aim at improving one of them, i.e., the LQP-based interior prediction-correction method proposed in [He, Liao and Yuan, J. Comp. Math., 2006, 24(1): 33–44], via identifying more appropriate step-sizes in the correction steps. Preliminary numerical results for solving some NCPs arising in traffic equilibrium problems are reported to verify the theoretical assertions.

A new logarithmic-quadratic proximal method for nonlinear complementarity problems

In this paper, we propose a new modified logarithmic-quadratic proximal (LQP) method for solving nonlinear complementarity problems (NCP). We suggest using a prediction-correction method to solve NCP. The predictor is obtained via solving the LQP system approximately under significantly relaxed accuracy criterion and the new iterate is computed by using a new step size α k . Under suitable conditions, we prove that the new method is globally convergent. We report preliminary computational results to illustrate the efficiency of the proposed method. This new method can be considered as a significant refinement of the previously known methods for solving nonlinear complementarity problems.

A new predicto-corrector method for pseudomonotone nonlinear complementarity problems

Based on the Logarithmic-Quadratic Proximal (LQP) method 2, we propose a new prediction-correction method for nonlinear complementarity problem (NCP). We obtain the predictor through a simplified inexact logarithmic-quadratic proximal method under a relaxed inexact criterion. The corrector is obtained by the improved extragradient method. Under certain conditions, the global convergence of the proposed method is proved. Preliminary numerical results indicate the efficiency of the proposed method.

Extended LQP Method for Monotone Nonlinear Complementarity Problems

To solve nonlinear complementarity problems (NCP), the logarithmic-quadratic proximal (LQP) method solves a system of nonlinear equations at each iteration. In this paper, the iterates generated by the original LQP method are extended by explicit formulas and thus an extended LQP method is presented. It is proved theoretically that the lower bound of the progress obtained by the extended LQP method is greater than that by the original LQP method. Preliminary numerical results are provided to verify the theoretical assertions and the effectiveness of both the original and the extended LQP method.

A new criterion for the inexact logarithmic-quadratic proximal method and its derived hybrid methods

To solve nonlinear complementarity problems, the inexact logarithmic-quadratic proximal (LQP) method solves a system of nonlinear equations (LQP system) approximately at each iteration. Therefore, the efficiencies of inexact-type LQP methods depend greatly on the involved inexact criteria used to solve the LQP systems. This paper relaxes inexact criteria of existing inexact-type LQP methods and thus makes it easier to solve the LQP system approximately. Based on the approximate solutions of the LQP systems, a descent method, and a prediction---correction method are presented. Convergence of the new methods are proved under mild assumptions. Numerical experiments for solving traffic equilibrium problems demonstrate that the new methods are more efficient than some existing methods and thus verify that the new inexact criterion is attractive in practice.

An self-adaptive LQP method for constrained variational inequalities

In this paper, we present a logarithmic-quadratic proximal (LQP) type prediction–correction methods for solving constrained variational inequalities VI ( S , f ) , where S is a convex set with linear constraints. The computational load in each iteration is quite tiny. However, the number of iterations is significantly dependent on a parameter which balances the primal and dual variables. We then propose a self-adaptive prediction–correction method that adjusts the scalar parameter automatically. Under certain conditions, the global convergence of the proposed method is established. In order to demonstrate the efficiency of the proposed method, we provide numerical results for a convex nonlinear programming and traffic equilibrium problems.

An LQP Method for Pseudomonotone Variational Inequalities

In this paper, we proposed a modified Logarithmic-Quadratic Proximal (LQP) method [Auslender et al.: Comput. Optim. Appl. 12, 31---40 (1999)] for solving variational inequalities problems. We solved the problem approximately, with constructive accuracy criterion. We show that the method is globally convergence under that the operator is pseudomonotone which is weaker than the monotonicity and the solution set is nonempty. Some preliminary computational results are given.

A Logarithmic-Quadratic Proximal Prediction-Correction Method for Structured Monotone Variational Inequalities

Inspired by the Logarithmic-Quadratic Proximal (LQP) method for variational inequalities, we present a prediction-correction method for structured monotone variational inequalities. Each iteration of the new method consists of a prediction and a correction. Both the predictor and the corrector are obtained easily with tiny computational load. In particular, the LQP system that appears in the prediction is approximately solved under significantly relaxed inexactness restriction. Global convergence of the new method is proved under mild assumptions. In addition, we present a self-adaptive version of the new method that leads to easier implementations. Preliminary numerical experiments for traffic equilibrium problems indicate that the new method is effectively applicable in practice.

Lagrangian Duality and Related Multiplier Methods for Variational Inequality Problems

We consider a new class of multiplier interior point methods for solving variational inequality problems with maximal monotone operators and explicit convex constraint inequalities. Developing a simple Lagrangian duality scheme which is combined with the recent logarithmic-quadratic proximal (LQP) theory introduced by the authors, we derive three algorithms for solving the variational inequality (VI) problem. This provides a natural extension of the methods of multipliers used in convex optimization and leads to smooth interior point multiplier algorithms. We prove primal, dual, and primal-dual convergence under very mild assumptions, eliminating all the usual assumptions used until now in the literature for related algorithms. Applications to complementarity problems are also discussed.