Combined Entropic Regularization and Path-Following Method for Solving Finite Convex Min-max Problems Subject to Infinitely Many Linear Constraints

In this paper, we study the minimization of the max function of q smooth convex functions on a domain specified by infinitely many linear constraints. The difficulty of such problems arises from the kinks of the max function and it is often suggested that, by imposing certain regularization functions, nondifferentiability will be overcome. We find that the entropic regularization introduced by Li and Fang is closely related to recently developed path-following interior-point methods. Based on their results, we create an interior trajectory in the feasible domain and propose a path-following algorithm with a convergence proof. Our intention here is to show a nice combination of minmax problems, semi-infinite programming, and interior-point methods. Hopefully, this will lead to new applications.     

一类线性与框式约束凸规划问题的原始-对偶内点算法

本文对一类具有线性和框式约束的凸规划问题给出了一个原始-对偶内点算法, 该算法可在任一原始-对偶可行内点启动, 并且全局收敛,当初始点靠近中心路径时, 算法成为中心路径跟踪算法。 数值实验表明, 算法对求解大型的这类问题是有效的。     

Value Estimation Approach to the Iri-Imai Method for Constrained Convex Optimization

In this paper we propose an extension of the so-called Iri-Imai method to solve constrained convex programming problems. The original Iri-Imai method is designed for linear programs and assumes that the optimal objective value of the optimization problem is known in advance. Zhang (Ref. 9) extends the method for constrained convex optimization but the optimum value is still assumed to be known in advance. In our new extension this last requirement on the optimal value is relaxed; instead only a lower bound of the optimal value is needed. Our approach uses a multiplicative barrier function for the problem with a univariate parameter that represents an estimated optimum value of the original optimization problem. An optimal solution to the original problem can be traced down by minimizing the multiplicative barrier function. Due to the convexity of this barrier function the optimal objective value as well as the optimal solution of the original problem are sought iteratively by applying Newtons method to the multiplicative barrier function. A new formulation of the multiplicative barrier function is further developed to acquire computational tractability and efficiency. Numerical results are presented to show the efficiency of the new method.His research supported by Hong Kong RGC Earmarked Grant CUHK4233/01E.Communicated by Z. Q. Luo     

Complexity analysis and numerical implementation of a short-step primal-dual algorithm for linear complementarity problems

In this paper we deal with the study of the polynomial complexity and numerical implementation for a short-step primal-dual interior point algorithm for monotone linear complementarity problems LCP. The analysis is based on a new class of search directions used by the author for convex quadratic programming (CQP) [M. Achache, A new primal-dual path-following method for convex quadratic programming, Computational and Applied Mathematics 25 (1) (2006) 97-110]. Here, we show that this algorithm enjoys the best theoretical polynomial complexity namely , iteration bound. For its numerical performances some strategies are used. Finally, we have tested this algorithm on some monotone linear complementarity problems.     

The extended supporting hyperplane algorithm for convex mixed-integer nonlinear programming

A new deterministic algorithm for solving convex mixed-integer nonlinear programming (MINLP) problems is presented in this paper: The extended supporting hyperplane (ESH) algorithm uses supporting hyperplanes to generate a tight overestimated polyhedral set of the feasible set defined by linear and nonlinear constraints. A sequence of linear or quadratic integer-relaxed subproblems are first solved to rapidly generate a tight linear relaxation of the original MINLP problem. After an initial overestimated set has been obtained the algorithm solves a sequence of mixed-integer linear programming or mixed-integer quadratic programming subproblems and refines the overestimated set by generating more supporting hyperplanes in each iteration. Compared to the extended cutting plane algorithm ESH generates a tighter overestimated set and unlike outer approximation the generation point for the supporting hyperplanes is found by a simple line search procedure. In this paper it is proven that the ESH algorithm converges to a global optimum for convex MINLP problems. The ESH algorithm is implemented as the supporting hyperplane optimization toolkit (SHOT) solver, and an extensive numerical comparison of its performance against other state-of-the-art MINLP solvers is presented.     

Extensions of Dinkelbach's algorithm for solving non-linear fractional programming problems

Dinkelbach's algorithm was developed to solve convex fractinal programming. This method achieves the optimal solution of the optimisation problem by means of solving a sequence of non-linear convex programming subproblems defined by a parameter. In this paper it is shown that Dinkelbach's algorithm can be used to solve general fractional programming. The applicability of the algorithm will depend on the possibility of solving the subproblems. Dinkelbach's extended algorithm is a framework to describe several algorithms which have been proposed to solve linear fractional programming, integer linear fractional programming, convex fractional programming and to generate new algorithms. The applicability of new cases as nondifferentiable fractional programming and quadratic fractional programming has been studied. We have proposed two modifications to improve the speed-up of Dinkelbachs algorithm. One is to use interpolation formulae to update the parameter which defined the subproblem and another truncates the solution of the suproblem. We give sufficient conditions for the convergence of these modifications. Computational experiments in linear fractional programming, integer linear fractional programming and non-linear fractional programming to evaluate the efficiency of these methods have been carried out.     

Dual Semidefinite Programs Without Duality Gaps for a Class of Convex Minimax Programs

In this paper, we introduce a new dual program, which is representable as a semidefinite linear programming problem, for a primal convex minimax programming problem, and we show that there is no duality gap between the primal and the dual whenever the functions involved are sum-of-squares convex polynomials. Under a suitable constraint qualification, we derive strong duality results for this class of minimax problems. Consequently, we present applications of our results to robust sum-of-squares convex programming problems under data uncertainty and to minimax fractional programming problems with sum-of-squares convex polynomials. We obtain these results by first establishing sum-of-squares polynomial representations of non-negativity of a convex max function over a system of sum-of-squares convex constraints. The new class of sum-of-squares convex polynomials is an important subclass of convex polynomials and it includes convex quadratic functions and separable convex polynomials. The sum-of-squares convexity of polynomials can numerically be checked by solving semidefinite programming problems whereas numerically verifying convexity of polynomials is generally very hard.     

Best Approximation on Convex Sets in Metric Linear Spaces

In this paper the concepts of strictly convex and uniformly convex normed linear spaces are extended to metric linear spaces. A relationship between strict convexity and uniform convexity is established. Some existence and uniqueness theorems on best approximation in metric linear spaces under different conditions are proved.     

On the linear convergence of the alternating direction method of multipliers

We analyze the convergence rate of the alternating direction method of multipliers (ADMM) for minimizing the sum of two or more nonsmooth convex separable functions subject to linear constraints. Previous analysis of the ADMM typically assumes that the objective function is the sum of only two convex functions defined on two separable blocks of variables even though the algorithm works well in numerical experiments for three or more blocks. Moreover, there has been no rate of convergence analysis for the ADMM without strong convexity in the objective function. In this paper we establish the global R-linear convergence of the ADMM for minimizing the sum of any number of convex separable functions, assuming that a certain error bound condition holds true and the dual stepsize is sufficiently small. Such an error bound condition is satisfied for example when the feasible set is a compact polyhedron and the objective function consists of a smooth strictly convex function composed with a linear mapping, and a nonsmooth \(\ell _1\) regularizer. This result implies the linear convergence of the ADMM for contemporary applications such as LASSO without assuming strong convexity of the objective function.     

A relaxed primal-dual path-following algorithm for linear programming

In this paper, we provide an easily satisfied relaxation condition for the primaldual interior path-following algorithm to solve linear programming problems. It is shown that the relaxed algorithm preserves the property of polynomial-time convergence. The computational results obtained by implementing two versions of the relaxed algorithm with slight modifications clearly demonstrate the potential in reducing computational efforts.Partially supported by the North Carolina Supercomputing Center, the 1993 Cray Research Award, and a National Science Council Research Grant of the Republic of China.     

A linear programming-based optimization algorithm for solving nonlinear programming problems

In this paper a linear programming-based optimization algorithm called the Sequential Cutting Plane algorithm is presented. The main features of the algorithm are described, convergence to a Karush–Kuhn–Tucker stationary point is proved and numerical experience on some well-known test sets is showed. The algorithm is based on an earlier version for convex inequality constrained problems, but here the algorithm is extended to general continuously differentiable nonlinear programming problems containing both nonlinear inequality and equality constraints. A comparison with some existing solvers shows that the algorithm is competitive with these solvers. Thus, this new method based on solving linear programming subproblems is a good alternative method for solving nonlinear programming problems efficiently. The algorithm has been used as a subsolver in a mixed integer nonlinear programming algorithm where the linear problems provide lower bounds on the optimal solutions of the nonlinear programming subproblems in the branch and bound tree for convex, inequality constrained problems.     

AN EXTENSION OF PREDICTOR-CORRECTOR ALGORITHM TO A CLASS OF CONVEX SEPARABLE PROGRAM

1.IntroductionIn[1]Mizuno,ToddandYepresentedapredictor-correctoralgorithmforlinearpramgrammingwhichpossessesaquadraticconvergencerateofthedualgaptozero.GuoandWul6]gaveamodificationofthisalgorithmforsolvingconvexquadraticprogramwithupperbounds.Itisshownthatthemodifiedmethodnotonlypreservesalltheoriginalmerits,butalsoreducesthedualgapbyaconstantfactorineachcorrectorstep,incontrasttotheMizuno,TOddandYe'soriginalpredictor--correctormethodwherethedualgapremainsunchanged.Thealgorithmdiscussedint…     

Applications of symmetric derivatives in mathematical programming

In recent times the Kuhn—Tucker optimality conditions and the duality theorems for convex programming have been extended by generalizations of the convexity concept. In this paper the notion of a symmetric derivative for a function of several variables is introduced and used to provide extensions of some fundamental optimality and duality theorems of convex programming. Symmetric derivatives are also used to extend some optimality and duality theorems involving pseudoconvexity and differentiable quasiconvexity.     

Weakly Convex Sets and Their Properties

In this paper, the notion of a weakly convex set is introduced. Sharp estimates for the weak convexity constants of the sum and difference of such sets are given. It is proved that, in Hilbert space, the smoothness of a set is equivalent to the weak convexity of the set and its complement. Here, by definition, the smoothness of a set means that the field of unit outward normal vectors is defined on the boundary of the set; this vector field satisfies the Lipschitz condition. We obtain the minimax theorem for a class of problems with smooth Lebesgue sets of the goal function and strongly convex constraints. As an application of the results obtained, we prove the alternative theorem for program strategies in a linear differential quality game.     

Parallel Interior Point Schemes for Solving Multistage Convex Programming

The predictor–corrector interior-point path-following algorithm is promising in solving multistage convex programming problems. Among many other general good features of this algorithm, especially attractive is that the algorithm allows possibility to parallelise the major computations. The dynamic structure of the multistage problems specifies a block-tridiagonal system at each Newton step of the algorithm. A wrap-around permutation is then used to implement the parallel computation for this step.     

Shadow prices in nonconvex mathematical programming

In this paper a definition is proposed for the concept of shadow prices in nonconvex programming. For a nonlinear program with equality and inequality constraints, existence of these prices and bounds for their possible values are obtained under the Mangasarian—Fromowitz regularity condition. Their exact values and some continuity properties are obtained under the more restrictive linear independence regularity condition. A definition of equilibrium prices is also proposed. Under convexity assumptions, all definitions and results coincide with those already known on this subject in convex programming.This research was supported by the Natural Sciences and Engineering Research Council of Canada under Grant A-9273.     

A Primal-Dual Decomposition Algorithm for Multistage Stochastic Convex Programming

This paper presents a new and high performance solution method for multistage stochastic convex programming. Stochastic programming is a quantitative tool developed in the field of optimization to cope with the problem of decision-making under uncertainty. Among others, stochastic programming has found many applications in finance, such as asset-liability and bond-portfolio management. However, many stochastic programming applications still remain computationally intractable because of their overwhelming dimensionality. In this paper we propose a new decomposition algorithm for multistage stochastic programming with a convex objective and stochastic recourse matrices, based on the path-following interior point method combined with the homogeneous self-dual embedding technique. Our preliminary numerical experiments show that this approach is very promising in many ways for solving generic multistage stochastic programming, including its superiority in terms of numerical efficiency, as well as the flexibility in testing and analyzing the model.Research supported by Hong Kong RGC Earmarked Grant CUHK4233/01E.     

凸二次半定规划一个长步原始对偶路径跟踪算法

本文基于Nesterov-Todd方向,并引进中心路径测量函数以及原始对偶对数障碍函数,建立了一个求解凸二次半定规划的长步路径跟踪法.算法保证当迭代点落在中心路径附近时步长1被接受.算法至多迭代O(n|lnε|)次可得到一个ε最优解.论文最后报告了初步的数值试验结果.     

凸规划的新算法

对一般凸目标函数和一般凸集约束的凸规划问题新解法进行探讨,它是线性规划一种新算法的扩展和改进,此算法的基本思想是在规划问题的可行域中由所建－的一个切割面到另一个切割面的不断推进来求取最优的。文章对目标函数是二次的且约束是一般凸集和二次目标函数且约束是线性的情形,给出了更简单的算法。