An Alternative Regularization Method for Nonexpansive Mappings with Applications

Hong-Kun Xu
DOI: https://doi.org/10.1090/conm/513/10087
2010-01-01
Abstract:We introduce a new regularization method for nonexpansive mappings in a Banach space and prove that if the underlying space is either uniformly smooth or has a weakly continuous duality map, then the regularized solutions converge in norm if and only if the relevant fixed point equation is consistent. We then discretize the regularization method to define iterative algorithms which are proved to be strongly convergent. Moreover, we apply these algorithms to find zeros of maximal monotone and m-accretive operators and solve a quadratic minimization problem.
What problem does this paper attempt to address?