Abstract
The plain Newton-min algorithm for solving the linear complementarity problem (LCP) “\(0\leqslant x\perp (Mx+q)\geqslant 0\)” can be viewed as an instance of the plain semismooth Newton method on the equational version “\(\min (x,Mx+q)=0\)” of the problem. This algorithm converges for any q when M is an \(\mathbf{M }\)-matrix, but not when it is a \(\mathbf{P }\)-matrix. When convergence occurs, it is often very fast (in at most n iterations for an \(\mathbf{M }\)-matrix, where n is the number of variables, but often much faster in practice). In 1990, Harker and Pang proposed to improve the convergence ability of this algorithm by introducing a stepsize along the Newton-min direction that results in a jump over at least one of the encountered kinks of the min-function, in order to avoid its points of nondifferentiability. This paper shows that, for the Fathi problem (an LCP with a positive definite symmetric matrix M, hence a \(\mathbf{P }\)-matrix), an algorithmic scheme, including the algorithm of Harker and Pang, may require n iterations to converge, depending on the starting point.
Similar content being viewed by others
References
Aganagić M (1984) Newton’s method for linear complementarity problems. Math Program 28:349–362. https://doi.org/10.1007/BF02612339
Armijo L (1966) Minimization of functions having Lipschitz continuous first partial derivatives. Pac J Math 16:1–3
Ben Gharbia I (2012) Résolution de Problèmes de Complémentarité: application à un Écoulement Diphasique Dans un Milieu Poreux. Ph.D. Thesis, Université Paris-Dauphine
Ben Gharbia I, Gilbert JC (2012) Nonconvergence of the plain Newton-min algorithm for linear complementarity problems with a P-matrix. Math Program 134:349–364. https://doi.org/10.1007/s10107-010-0439-6
Ben Gharbia I, Gilbert JC (2013) An algorithmic characterization of P-matricity. SIAM J Matrix Anal Appl 34(3):904–916. https://doi.org/10.1137/120883025
Ben Gharbia I, Gilbert JC (2019) An algorithmic characterization of P-matricity II: adjustments, refinements, and validation. SIAM J Matrix Anal Appl 40:800–813. https://doi.org/10.1137/18M1168522
Bergounioux M, Ito K, Kunisch K (1999) Primal–dual strategy for constrained optimal control problems. SIAM J Control Optim 37:1176–1194. https://doi.org/10.1137/S0363012997328609
Bergounioux M, Haddou M, Hintermüller M, Kunisch K (2000) A comparison of a Moreau–Yosida-based active set strategy and interior point methods for constrained optimal control problems. SIAM J Optim 11:495–521. https://doi.org/10.1137/S1052623498343131
Bonnans JF, Gilbert JC, Lemaréchal C, Sagastizábal C (2006) Numerical optimization: theoretical and practical aspects, 2nd edn. Springer, Berlin. https://doi.org/10.1007/978-3-540-35447-5
Burke JV, Xu S (1999) A polynomial time interior-point path-following algorithm for LCP based on Chen–Harker–Kanzow smoothing techniques. Math Program 86(1):91–103. https://doi.org/10.1007/s101070050081
Burke JV, Xu S (2002) Complexity of a noninterior path-following method for the linear complementarity problem. J Optim Theory Appl 112(1):53–76. https://doi.org/10.1023/A:1013040428127
Chandrasekaran R (1970) A special case of the complementary pivot problem. Opsearch 7:263–268
Chen B, Harker PT (1993) A non-interior continuation method for linear complementarity problems. SIAM J Matrix Anal Appl 14:1168–1190. https://doi.org/10.1137/0614081
Chung SJ (1989) NP-completeness of the linear complementarity problem. J Optim Theor Appl 60:393–399. https://doi.org/10.1007/BF00940344
Conn AR, Gould NIM, Toint PhL (2000) Trust-region methods. MPS-SIAM series on optimization 1. SIAM and MPS, Philadelphia. https://doi.org/10.1137/1.9780898719857
Cottle RW, Pang J-S, Stone RE (2009) The linear complementarity problem. Classics in applied mathematics, vol 60. SIAM, Philadelphia
Curtis FE, Han Z, Robinson DP (2015) A globally convergent primal–dual active-set framework for large-scale convex quadratic optimization. Comput Optim Appl 60(2):311–341. https://doi.org/10.1007/s10589-014-9681-9
Dennis JE, Schnabel RB (1983) Numerical methods for unconstrained optimization and nonlinear equations. Prentice-Hall, Englewood Cliffs
Dussault J-P, Frappier M, Gilbert JC (2018) Nmhp, version 1.0. https://who.rocq.inria.fr/Jean-Charles.Gilbert/codes/nmhp/nmhp.zip
Dussault J-P, Frappier M, Gilbert JC (2019) A globally convergent polyhedral Newton-min algorithm for solving a nonlinear complementarity problem. Technical report (to appear)
Fathi Y (1979) Computational complexity of LCPs associated with positive definite symmetric matrices. Math Program 17:335–344
Fischer A, Kanzow C (1996) On finite termination of an iterative method for linear complementarity problems. Math Program 74:279–292. https://doi.org/10.1007/BF02592200
Harker PT, Pang J-S (1990) A damped-Newton method for the linear complementarity problem. In: Allgower EL, Georg K (eds) Computational solution of nonlinear systems of equations, vol 26. Lecture in Applied Mathematics. AMS, Providence
Hintermüller M, Ito K, Kunisch K (2003) The primal–dual active set strategy as a semismooth Newton method. SIAM J Optim 13:865–888. https://doi.org/10.1137/S1052623401383558
Hotta K, Inaba M, Yoshise A (2000) A complexity analysis of a smoothing method using CHKS-functions for monotone linear complementarity problems. Comput Optim Appl 17(2–3):183–201. https://doi.org/10.1023/A:1026550331760
Izmailov AF, Solodov MV (2014) Newton-type methods for optimization and variational problems. Springer series in operations research and financial engineering. Springer, Berlin
Kanzow C (1996) Some noninterior methods for linear complementarity problems. SIAM J Matrix Anal Appl 17(4):851–868. https://doi.org/10.1137/S0895479894273134
Kanzow C (2004) Inexact semismooth Newton methods for large-scale complementarity problems. Optim Methods Softw 19:309–325. https://doi.org/10.1080/10556780310001636369
Kojima M, Shindo S (1986) Extension of Newton and quasi-Newton methods to systems of PC\(^1\) equations. J Oper Res Soc Jpn 29:352–375
Kojima M, Megiddo N, Noma T, Yoshise A (1991) A unified approach to interior point algorithms for linear complementarity problems, vol 538. Lecture Notes in Computer Science. Springer, Berlin. https://doi.org/10.1007/3-540-54509-3
Kojima M, Megiddo N, Noma T, Yoshise A (1991) A unified approach to interior point algorithms for linear complementarity problems: a summary. Oper Res Lett 10(5):247–254. https://doi.org/10.1016/0167-6377(91)90010-M
Kostreva MM (1976) Direct algorithms for complementarity problems. Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy, New York
Mangasarian OL (1977) Solution of symmetric linear complementarity problems by iterative methods. J Optim Theor Appl 22:465–485. https://doi.org/10.1007/BF01268170
Mizuno S, Yoshise A, Kikuchi T (1989) Practical polynomial time algorithms for linear complmentarity problems. J Oper Res Soc Jpn 32(1):75–92. https://doi.org/10.15807/jorsj.32.75
Murty KG (1974) Note on a Bard-type scheme for solving the complementarity problem. Opsearch 11:123–130
Murty KG (1978) Computational complexity of complementarity pivot methods. Math Program Stud 7:61–73
Murty KG (1988) Linear complementarity, linear and nonlinear programming (Internet edition, prepared by Feng-Tien Yu, 1997). Heldermann Verlag, Berlin
Pang J-S (1990) Newton’s method for B-differentiable equations. Math Oper Res 15:311–341. https://doi.org/10.1287/moor.15.2.311
Pang J-S (1991) A B-differentiable equation-based, globally and locally quadratically convergent algorithm for nonlinear programs, complementarity and variational inequality problems. Math Program 51:101–131. https://doi.org/10.1007/BF01586928
Qi L, Sun J (1993) A nonsmooth version of Newton’s method. Math Program 58:353–367. https://doi.org/10.1007/BF01581275
Samelson H, Thrall RM, Wesler O (1958) A partition theorem for the Euclidean \(n\)-space. Proc Am Math Soc 9:805–807
Acknowledgements
We have benefited from a careful reading of the manuscript by the referee, who is warmly thanked.
Author information
Authors and Affiliations
Corresponding authors
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Dussault, JP., Frappier, M. & Gilbert, J.C. A lower bound on the iterative complexity of the Harker and Pang globalization technique of the Newton-min algorithm for solving the linear complementarity problem. EURO J Comput Optim 7, 359–380 (2019). https://doi.org/10.1007/s13675-019-00116-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13675-019-00116-6
Keywords
- Iterative complexity
- Linear complementarity problem
- Fathi and Murty problems
- Globalization
- Harker and Pang algorithm
- Line search
- Newton-min algorithm
- Nondegenerate matrix
- P-matrix
- Semismooth Newton method