Newton's method for the nonlinear complementarity problem: a B- differentiable equation approach (Q2277159)

The nonlinear complementarity problem \(F^ T(x)x=0\), \(F(x)\in R^ n_+\) can be equivalently formulated as a system of nonlinear equations \(H(x)=\min (x,F(x))=0\), where the `min' operation is taken componentwise. The function H is not Fréchet differentiable in general, however, it is B-differentiable. S. Robinson was the first to study Newton's method for equations with such functions. The present paper develops a previous work by the first author and \textit{J.-S. Pang} [in: Computational solution of nonlinear systems of equations, Proc. SIAM-AMS Summer Semin., Ft. Collins/CO (USA) 1988, Lect. Appl. Math. 26, 265-284 (1990; Zbl 0699.65054)] converting the original problem into a system of equations through the use of a Minty map. At each step of the algorithm a linear system is solved and a line search for a given merit function is performed. Numerical results and a comparison with the traditional Josephy-Newton method are presented.