A nonsmooth version of Newton's method
Mathematical Programming: Series A and B
Convergence analysis of some algorithms for solving nonsmooth equations
Mathematics of Operations Research
Journal of Optimization Theory and Applications
Semismooth homeomorphisms and strong stability of semidefinite and Lorentz complementarity problems
Mathematics of Operations Research
SC1 optimization reformulations of the generalized Nash equilibrium problem
Optimization Methods & Software
Computational Optimization and Applications
Computational Optimization and Applications
Hi-index | 0.00 |
Let H:R^mxR^n-R^n be a locally Lipschitz function in a neighborhood of (y@?,x@?) and H(y@?,x@?)=0 for some y@?@?R^m and x@?@?R^n. The implicit function theorem in the sense of Clarke (Pacific J. Math. 64 (1976) 97; Optimization and Nonsmooth Analysis, Wiley, New York, 1983) says that if @p"x@?H(y@?,x@?) is of maximal rank, then there exist a neighborhood Y of y@? and a Lipschitz function G(.):Y-R^n such that G(y@?)=x@? and for every y in Y, H(y,G(y))=0. In this paper, we shall further show that if H has a superlinear (quadratic) approximate property at (y@?,x@?), then G has a superlinear (quadratic) approximate property at y@?. This result is useful in designing Newton's methods for nonsmooth equations.