optimization that are known to converge only weakly. Exploiting an idea originally proposed
by Haugazeau, we present a simple modification of these methods that makes them strongly
convergent without additional assumptions. Several applications are discussed.