A Potential Reduction Newton Method for Constrained Equations

From MaRDI portal





The goal of this useful paper is to present a potential reduction Newton method for solving a constrained equations (a constrained system of nonlinear equations), without assuming the existence of the partitioning property. The constrained equation (CE) defined by the pair \((\Omega, H)\) (\(H:\mathbb{R}^n\to \mathbb{R}^n\) be a given mapping, \(\Omega\) be a given closed subset of \(\mathbb{R}^n\)) is to find a vector \(x\in\mathbb{R}^n\) such that: \(H(x)= 0\), \(x\in\Omega\). The method for solving the \(\text{CE}(\Omega, H)\) combines ideas from the classical damped Newton method for solving the unconstrained system of equations \(H(x)= 0\), \(x\in\mathbb{R}^n\), and from the family of interior point methods for solving constrained optimization and complementarity problems. The work is based on a set of broad hypotheses on the pair \((\Omega, H)\). Main result: A major convergence result for the method is established. Specializations of the method to a convex semidefinite program and a monotone complementarity problem in symmetric matrices are discussed. Strengthened convergence results are established in the context of these specializations.




Cited in
(20)








This page was built for publication: A Potential Reduction Newton Method for Constrained Equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4702307)