This paper is concerned with the theory of the convergence of discretizations f o r set-valued operator equations. It is a continuation of investigations in 121, 171, [5], some results for one-valued operators have been generalrzed herein. As an application, two examples for set-valued differential equations are discussed.Define set convergence by S, + S, if any &-neighborhood of S contains S, for all n sufficiently large. Theorem 2.1: In equations (l), (2) assume that y , -+ y, [{A,}, A] is d-closed, and {S,} is d-compact. Then {S,}* c S and S' ,z + S. If S, # 0 for n E N', then S # 0. P r o o f : Let z E {SrL}*, then there are N' and x, E S,, n E N', such that x, ---f 2 . Since [{A,,}, A] is d-closed and y, + y, it follows that x E D(A) and y E Ax, that is z E S and {Sn}* c S. The rest follows from 111.Now we consider the weak formulation of problem (1). Let Z be a Banach space, J be an index set, {y(@), @ E J } c Z be a given set, and { A ( @ ) , @ E J } be a set of operators that map D = n D(A(@)) c X into 22, here D(A(@)) is the domain of A ( @ ) .
@ E J