The exact l1 penalty function method for constrained nonsmooth invex optimization problems

Abstract

The exactness of the penalization for the exact l 1 penalty function method used for solving nonsmooth constrained optimization problems with both inequality and equality constraints is considered. Thus, the equivalence between the sets of optimal solutions in the nonsmooth constrained optimization problem and its associated penalized optimization problem with the exact l 1 penalty function is established under locally Lipschitz invexity assumptions imposed on the involved functions. © 2013 IFIP International Federation for Information Processing.