A Mehrotra-type predictor-corrector algorithm with polynomiality andQ-subquadratic convergence

Mehrotra's predictor-corrector algorithm [3] is currently considered to be one of the most practically efficient interior-point methods for linear programming. Recently, Zhang and Zhang [18] studied the global convergence properties of the Mehrotra-type predictor-corrector approach and established polynomial complexity bounds for two interior-point algorithms that use the Mehrotra predictor-corrector approach. In this paper, we study the asymptotic convergence rate for the Mehrotra-type predictor-corrector interior-point algorithms. In particular, we construct an infeasible-interior-point algorithm based on the second algorithm proposed in [18] and show that while retaining a complexity bound ofO(n ^1.5 t)-iterations, under certain conditions the algorithm also possesses aQ-subquadratic convergence, i.e., a convergence ofQ-order 2 with an unboundedQ-factor.Research supported in part by NSF DMS-9102761 and DOE DE-FG02-93ER25171.     

P*(k)线性互补问题的Mehrotra型预估-校正算法复杂性分析

本文提出一种求解单调非线性互补问题的Mehrotra型预估-校正算法.新算法采用不同的自适应更新策略.在尺度化的Lipschitz条件下,证明了新算法的迭代复杂性为O(n2 log (x0)T s0/ε)),其中(x0,s0)为初始点,ε为精度.     

A finite termination Mehrotra-type predictor–corrector algorithm

Mehrotra-type predictor–corrector algorithms are the backbone of most Interior Point Methods based software packages. Salahi et al. [M. Salahi, J. Peng, T. Terlaky, On Mehrotra-type predictor–corrector algorithms, Technical Report 2005/4, Advanced Optimization Lab., McMaster University, Hamilton, ON, Canada. http://www.cas.mcmaster.ca/~oplab/publication] in their recent works have shown some ill behaviors of Mehrotra’s original algorithm which motivated them to modify it in order to achieve the polynomial iteration complexity while preserving its practical efficiency. In this paper we analyze the same algorithm from a different perspective and give a condition on the maximum feasible step size in the predictor step, violation of which might lead to a very small or even zero step size in the corrector step. If the maximum step size in the predictor step is above a certain threshold, then we cut it to satisfy the derived condition. This enables us to prove that the algorithm terminates finitely.     

Postponing the choice of the barrier parameter in Mehrotra-type predictor–corrector algorithms

Recently Salahi et al. have considered a variant of Mehrotra’s celebrated predictor–corrector algorithm. By a numerical example they showed that this variant might make very small steps in order to keep the iterate in a certain neighborhood of the central path, that itself implies the inefficiency of the algorithm. This observation motivated them to incorporate a safeguard in their algorithmic scheme that gives a lower bound for the step size at each iteration and thus imply polynomial iteration complexity. In this paper we propose a different approach that enables us to have control on the iterates.     

$P_*(\kappa)$线性互补问题的Mehrotra型预估-校正算法

Mehrotra-type predictor-corrector algorithm,as one of most efficient interior point methods,has become the backbones of most optimization packages.Salahi et al.proposed a cut strategy based algorithm for linear optimization that enjoyed polynomial complexity and maintained its efficiency in practice.We extend their algorithm to P_*（κ）linear complementarity problems.The way of choosing corrector direction for our algorithm is different from theirs. The new algorithm has been proved to have an ο(（1+4κ）（17+19κ） √（1+2κn）^3/2log[（x⁰）^Ts⁰/ε] worst case iteration complexity bound.An numerical experiment verifies the feasibility of the new algorithm.     

A Mehrotra-Type Predictor-Corrector Algorithm for P*(κ) Linear Complementarity Problems

Mehrotra-type predictor-corrector algorithm,as one of most efficient interior point methods,has become the backbones of most optimization packages.Salahi et al.proposed a cut strategy based algorithm for linear optimization that enjoyed polynomial complexity and maintained its efficiency in practice.We extend their algorithm to P_*（κ）linear complementarity problems.The way of choosing corrector direction for our algorithm is different from theirs. The new algorithm has been proved to have an ο(（1+4κ）（17+19κ） √（1+2κn）^3/2log[（x⁰）^Ts⁰/ε] worst case iteration complexity bound.An numerical experiment verifies the feasibility of the new algorithm.     

基于自适应参数校正策略求解SDP的Mehrotra型内点算法

最近,Salahi对线性规划提出了一个基于新的自适应参数校正策略的Mehrotra型预估-校正算法,该策略使其在不使用安全策略的情况下,证明了算法的多项式迭代复杂界.本文将这一算法推广到半定规划的情形.通过利用Zhang的对称化技术,得到了算法的多项式迭代复杂界,这与求解线性规划的相应算法有相同的迭代复杂性阶.     

Infeasible Mehrotra-Type Predictor-Corrector Interior-Point Algorithm for the Cartesian P *(κ)-LCP Over Symmetric Cones

We propose an infeasible Mehrotra-type predictor-corrector algorithm with a new center parameter updating scheme for Cartesian P _*(κ)-linear complementarity problem over symmetric cones. Based on the Nesterov-Todd direction, we show that the iteration-complexity bound of the proposed algorithm is 𝒪((1 + κ)³ r ²log ε^?1), where r is the rank of the associated Euclidean Jordan algebras and κ is the handicap of the problem and ε > 0 is the required precision. Some numerical results are reported as well.     

单调线性互补问题的Mehrotra型预估-校正算法的迭代复杂性

Mehrotra型预估-校正算法是很多内点算法软件包的算法基础,但它的多项式迭代复杂性直到2007年才被Salahi等人证明.通过选择一个固定的预估步长及与Salahi文中不同的校正方向,本文把Salahi等人的算法拓展到单调线性互补问题,使得新算法的迭代复杂性为O(n log((x0)T s0/ε)),同时,初步的数值实验证明了新算法是有效的.