Page 97 - Complementarity and Variational Inequalities in Electronics
P. 97

88  Complementarity and Variational Inequalities in Electronics


                           Noting that   ∞ (e) < +∞ since e ∈ D(  ∞ ), we may divide this last relation
                           by ||u i || to get
                                        1               q        1
                                         z i ,e + Mz i +   ,e +       ∞ (e) ≥ 0.
                                        i             ||u i ||  ||u i ||
                           Taking the limit as i →+∞, we get  Mz,e ≥ 0. This holds for any e ∈
                           D(  ∞ ), and thus

                                                     z ∈ K(M, ).                     (4.44)
                           Setting now v = x 0 in (4.42), we obtain
                             1    2              1
                              ||u i || + Mu i ,u i  ≤ ( I + M)u i ,x 0  + q,x 0 − u i  +  (x 0 ) −  (u i ).
                             i                   i
                                                                                     (4.45)

                           The function   is proper, convex, and lower semicontinuous, and thus (see e.g.
                           Theorem 1.1.11 in [46]) there exist a ≥ 0 and b ∈ R such that
                                                    n
                                             (∀x ∈ R ) :  (x) ≥−a||x|| + b.
                           Thus

                                                      1
                                Mu i ,u i  ≤ a||u i || − b + ( I + M)u i ,x 0  + q,x 0 − u i  +  (x 0 ).
                                                      i
                                                      2
                           Dividing this last relation by ||u i || , we get
                                                 a       b      1          x 0
                                       Mz i ,z i  ≤  −      + ( I + M)z i ,
                                                ||u i ||  ||u i || 2  i   ||u i ||
                                                    q    x 0        (x 0 )
                                                +      ,     − z i  +    .
                                                   ||u i || ||u i ||  ||u i || 2
                           Taking the limit as i →+∞, we get  Mz,z ≤ 0. Thus

                                                     z ∈ N − (M).                    (4.46)
                           Using (4.43), (4.44), and (4.46), we obtain that

                                                     z ∈ B(M, ).

                           We get a contradiction since we have proved that z 
= 0 and z ∈ B(M, ).
                              The sequence {u i } is thus bounded. There exists a converging subsequence,
                           again denoted {u i }.Let

                                                     u = lim u i .
                                                        i→+∞
   92   93   94   95   96   97   98   99   100   101   102