Abstract
In this paper, we introduce a family of new NCP functions which is smooth, coercive and strongly semismooth. Based on new NCP functions, we propose an inexact Levenberg-Marquardt method for solving Nonlinear Complementarity Problem (NCP). Different from existing exact/inexact Levenberg-Marquardt methods, the proposed method adopts a derivative-free line search to ensure its globalization. Moreover, by using the strong semismoothness of new NCP functions, wc prove that the proposed method is locally superlinearly/quadratically convergent under a local error bound condition. Some numerical results are reported.
Original language | English |
---|---|
Pages (from-to) | 2361-2385 |
Number of pages | 25 |
Journal | Journal of Nonlinear and Convex Analysis |
Volume | 24 |
Issue number | 11 |
Publication status | Published - 2023 |
Keywords
- Levenberg-Marquardt method
- NCP function
- Nonlinear complementarity problem
- superlinear/quadratic convergence
ASJC Scopus subject areas
- Analysis
- Geometry and Topology
- Control and Optimization
- Applied Mathematics