New complexity analysis of a full-Newton step feasible interior-point algorithm for $$P_*(\kappa )$$P∗(κ)-LCP

In this paper, we consider a full-Newton step feasible interior-point algorithm for $$P_*(\kappa )$$P∗(κ)-linear complementarity problem. The perturbed complementarity equation $$xs=\mu e$$xs=μe is transformed by using a strictly increasing function, i.e., replacing $$xs=\mu e$$xs=μe by $$\psi (xs)=\psi (\mu e)$$ψ(xs)=ψ(μe) with $$\psi (t)=\sqrt{t}$$ψ(t)=t, and the proposed interior-point algorithm is based on that algebraic equivalent transformation. Furthermore, we establish the currently best known iteration bound for $$P_*(\kappa )$$P∗(κ)-linear complementarity problem, namely, $$O((1+4\kappa )\sqrt{n}\log \frac{n}{\varepsilon })$$O((1+4κ)nlognε), which almost coincides with the bound derived for linear optimization, except that the iteration bound in the $$P_{*}(\kappa )$$P∗(κ)-linear complementarity problem case is multiplied with the factor $$(1+4\kappa )$$(1+4κ).

[1]  G. Lesaja,et al.  A Class of Large-Update and Small-Update Primal-Dual Interior-Point Algorithms for Linear Optimization , 2008 .

[2]  Zsolt Darvay New Interior Point Algorithms in Linear Programming , 2003 .

[3]  Shaohua Pan,et al.  An infeasible primal–dual interior point algorithm for linear programs based on logarithmic equivalent transformation , 2006 .

[4]  Michael C. Ferris,et al.  Engineering and Economic Applications of Complementarity Problems , 1997, SIAM Rev..

[5]  Hossein Mansouri,et al.  Polynomial interior-point algorithm for P*${(\kappa)}$ horizontal linear complementarity problems , 2012, Numerical Algorithms.

[6]  Shuzhong Zhang,et al.  An O(\sqrtn L) Iteration Primal-dual Path-following Method, Based on Wide Neighborhoods and Large Updates, for Monotone LCP , 2005, SIAM J. Optim..

[7]  Hossein Mansouri,et al.  A Polynomial Interior-Point Algorithm for Monotone Linear Complementarity Problems , 2013, J. Optim. Theory Appl..

[8]  Behrouz Kheirfam,et al.  A New Complexity Analysis for Full-Newton Step Infeasible Interior-Point Algorithm for Horizontal Linear Complementarity Problems , 2014, J. Optim. Theory Appl..

[9]  Goran Lesaja,et al.  Unified Analysis of Kernel-Based Interior-Point Methods for P*(?)-Linear Complementarity Problems , 2010 .

[10]  Kees Roos,et al.  A Comparative Study of Kernel Functions for Primal-Dual Interior-Point Algorithms in Linear Optimization , 2004, SIAM J. Optim..

[11]  Kees Roos,et al.  Unified Analysis of Kernel-Based Interior-Point Methods for P*(Kappa)-Linear Complementarity Problems , 2010, SIAM J. Optim..

[12]  Jianming Miao A quadratically convergent $$O((\kappa + 1)\sqrt n L)$$ -iteration algorithm for theP*(κ)-matrix linear complementarity problem , 1995, Math. Program..

[13]  Nimrod Megiddo,et al.  A Unified Approach to Interior Point Algorithms for Linear Complementarity Problems , 1991, Lecture Notes in Computer Science.

[14]  Florian A. Potra,et al.  Interior Point Methods for Sufficient Horizontal LCP in a Wide Neighborhood of the Central Path with Best Known Iteration Complexity , 2014, SIAM J. Optim..

[15]  Mohamed Achache,et al.  Complexity analysis and numerical implementation of a short-step primal-dual algorithm for linear complementarity problems , 2010, Appl. Math. Comput..

[16]  Lipu Zhang,et al.  A full-Newton step interior-point algorithm based on modified-Newton direction , 2011 .

[17]  Shuzhong Zhang,et al.  iteration primal-dual path-following method , based on wide neighborhoods and large updates , for monotone linear complementarity problems , 2004 .

[18]  Florian A. Potra,et al.  Predictor-corrector algorithm for solvingP*(κ)-matrix LCP from arbitrary positive starting points , 1996, Math. Program..

[19]  Kok Lay Teo,et al.  A full-Newton step feasible interior-point algorithm for P∗(κ)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$P_*(\kap , 2013, Journal of Global Optimization.