Article ID: | iaor20113598 |
Volume: | 217 |
Issue: | 16 |
Start Page Number: | 6917 |
End Page Number: | 6925 |
Publication Date: | Apr 2011 |
Journal: | Applied Mathematics and Computation |
Authors: | Li Yuan Min, Wang Xing Tao, Wei De Yun |
Keywords: | Newton method, penalty functions, smoothing |
In this paper, we first investigate a two‐parametric class of smoothing functions which contains the penalized smoothing Fischer–Burmeister function and the penalized smoothing CHKS function as special cases. Then we present a smoothing Newton method for the nonlinear complementarity problem based on the class of smoothing functions. Issues such as line search rule, boundedness of the level set, global and quadratic convergence are studied. In particular, we give a line search rule containing the common used Armijo‐type line search rule as a special case. Also without requiring strict complementarity assumption at the P