A SPLITTING METHOD FOR COMPOSITE MAPPINGS
Teemu Pennanen · Numerical Functional Analysis and Optimization · 2002
This paper studies the numerical solution of variational inclusions in the composite form , where is linear, is monotone, and is the adjoint of A. Such problems arise frequently in practice, but it seems that their numerical solution has not been considered (explicitly) before. This is in sharp contrast with the special case of sums for which several splitting methods are available. It turns out that the method of partial inverses due to Spingarn yields a simple splitting method for composite mappings, that achieves a decomposition between A and T. Several applications of this method are given.