Convergence conditions for Newton-type methods applied to complementarity systems with nonisolated solutions

We consider a class of Newton-type methods that are designed for the difficult case when solutions need not be isolated, and the equation mapping need not be differentiable at the solutions. We show that the only structural assumption needed for rapid local convergence of those algorithms applied to PC$$^1$$1-equations is the piecewise error bound, i.e., a local error bound holding for the branches of the solution set resulting from partitions of the bi-active complementarity indices. The latter error bound is implied by various piecewise constraint qualifications, including relatively weak ones. We apply our results to KKT systems arising from optimization or variational problems, and from generalized Nash equilibrium problems. In the first case, we show convergence if the dual part of the solution is a noncritical Lagrange multiplier, and in the second case convergence follows under a relaxed constant rank condition. In both cases, previously available results are improved.

[1]  Francisco Facchinei,et al.  A family of Newton methods for nonsmooth constrained systems with nonisolated solutions , 2013, Math. Methods Oper. Res..

[2]  O. Mangasarian,et al.  The Fritz John Necessary Optimality Conditions in the Presence of Equality and Inequality Constraints , 1967 .

[3]  Alexey F. Izmailov,et al.  Stabilized SQP revisited , 2012, Math. Program..

[4]  Francisco Facchinei,et al.  A new error bound result for Generalized Nash Equilibrium Problems and its algorithmic application , 2014, Comput. Optim. Appl..

[5]  Francisco Facchinei,et al.  Generalized Nash equilibrium problems and Newton methods , 2008, Math. Program..

[6]  Andreas Fischer,et al.  GENERALIZED NASH EQUILIBRIUM PROBLEMS - RECENT ADVANCES AND CHALLENGES , 2014 .

[7]  Jerzy Kyparisis,et al.  Sensitivity Analysis for Nonlinear Programs and Variational Inequalities with Nonunique Multipliers , 1990, Math. Oper. Res..

[8]  Mikhail V. Solodov,et al.  Stabilized sequential quadratic programming for optimization and a stabilized Newton-type method for variational problems , 2010, Math. Program..

[9]  C. Kanzow.  Levenberg-Marquardt methods for constrained nonlinear equations with strong local convergence properties , 2004 .

[10]  Alexey F. Izmailov,et al.  The Josephy–Newton Method for Semismooth Generalized Equations and Semismooth SQP for Optimization , 2013 .

[11]  C. Wojcik.  Springer international publishing switzerland , 2016 .

[12]  Alexey F. Izmailov,et al.  Newton-Type Methods for Optimization Problems without Constraint Qualifications , 2004, SIAM J. Optim..

[13]  Francisco Facchinei,et al.  An LP-Newton method: nonsmooth equations, KKT systems, and nonisolated solutions , 2013, Mathematical Programming.

[14]  F. Facchinei,et al.  Finite-Dimensional Variational Inequalities and Complementarity Problems , 2003 .

[15]  S. M. Robinson.  Stability Theory for Systems of Inequalities, Part II: Differentiable Nonlinear Systems , 1976 .

[16]  Francisco Facchinei,et al.  Generalized Nash Equilibrium Problems , 2010, Ann. Oper. Res..

[17]  J. Frédéric Bonnans,et al.  Perturbation Analysis of Optimization Problems , 2000, Springer Series in Operations Research.

[18]  M. Kojima,et al.  EXTENSION OF NEWTON AND QUASI-NEWTON METHODS TO SYSTEMS OF PC^1 EQUATIONS , 1986 .

[19]  Alexey F. Izmailov,et al.  A note on upper Lipschitz stability, error bounds, and critical multipliers for Lipschitz-continuous KKT systems , 2013, Math. Program..

[20]  Paulo J. S. Silva,et al.  A relaxed constant positive linear dependence constraint qualification and applications , 2011, Mathematical Programming.

[21]  Alexey F. Izmailov,et al.  On error bounds and Newton-type methods for generalized Nash equilibrium problems , 2014, Comput. Optim. Appl..

[22]  Andreas Fischer,et al.  A unified local convergence analysis of inexact constrained Levenberg–Marquardt methods , 2011, Optimization Letters.

[23]  Alexey F. Izmailov,et al.  Karush-Kuhn-Tucker systems: regularity conditions, error bounds and a class of Newton-type methods , 2003, Math. Program..

[24]  M. Fukushima,et al.  Erratum to Levenberg-Marquardt methods with strong local convergence properties for solving nonlinear equations with convex constraints , 2005 .

[25]  A. F. Izmailov,et al.  Newton-Type Methods for Optimization and Variational Problems , 2014 .

[26]  Andreas Fischer,et al.  Modified Wilson's Method for Nonlinear Programs with Nonunique Multipliers , 1999, Math. Oper. Res..

[27]  L. Minchenko,et al.  On relaxed constant rank regularity condition in mathematical programming , 2011 .

[28]  W. Hager.  Lipschitz Continuity for Constrained Processes , 1979 .

[29]  Stephen J. Wright.  An Algorithm for Degenerate Nonlinear Programming with Rapid Local Convergence , 2005, SIAM J. Optim..

[30]  Shu Lu,et al.  Relation between the constant rank and the relaxed constant rank constraint qualifications , 2012 .

[31]  Axel Dreves.  Improved error bound and a hybrid method for generalized Nash equilibrium problems , 2016, Comput. Optim. Appl..