Nonmonotone Line Search for Minimax Problems

Abstract

It was recently shown that, in the solution of smooth constrained optimization problems by sequential programming (SQP), the Maratos effect can be prevented by means of a certain nonmonotone (more precisely, four-step monotone) line search. Using a well known transformation, this scheme can be readily extended to the case of minimax problems. It turns out however that, due to the structure of these problems, one can use a simpler scheme. Such a scheme is proposed and analyzed in this paper. It is also shown that a three-step monotone (rather than four-step monotone) line search, with a relaxed decrease requirement, can be used without losing the theoretical convergence properties. Numerical experiments indicate a significant advantage of the proposed line search over the (monotone) Armijo search.