SELF-DUAL CLASSES OF NONEXPANSIVE OPERATORS IN NONLINEAR
                      ANALYSIS AND OPTIMIZATION 

                       Patrick L. Combettes
                      Department of Mathematics
                   North Carolina State University
                           Raleigh, NC 27695

ABSTRACT: Duality plays a fundamental role in standard 
optimization theory and in its extensions to more general
nonlinear analysis problems such as variational inequalities and
monotone inclusions. The algorithmic structures underlying the
numerical solution of such problems rely implicitly on classes of
operators that possess certain self-duality features. Examples of
self-dual classes of nonexpansive operators will be presented and
their properties will be discussed. Various applications will be covered.