Page 120 - Complementarity and Variational Inequalities in Electronics
P. 120

A Variational Inequality Theory Chapter | 4 111


                           It is clear that u i ∈ D( ) (i ∈ N). Moreover, for i large enough, ||u i || 
= 0, and
                           we may set
                                                            u i
                                                      z i =    .
                                                           ||u i ||
                           There exist subsequences, again denoted by {t i } and {z i }, such that lim t i =
                                                                                  i→+∞
                           t ∈[0,1] and lim z i = z with ||z|| = 1.
                                      i→+∞
                              Setting now v = 0in (4.82), we obtain
                                              2
                                    (1 − t i )||u i || + t i  F(u i ),u i  + q,u i  ≤  (0) −  (u i ).
                           For i great enough, we have  u i  ≥ σ, and thus

                                               2
                                                                               2
                                     (1 − t i )||u i || + t i  F(u i ),u i  ≥ ((1 − t i ) + αt i )||u i || .
                           If α ≥ 1, then ((1−t i )+αt i ) ≥ ((1−t i )+t i ) = 1, and if α ≤ 1, then ((1−t i )+
                           αt i ) ≥ (α(1 − t i ) + αt i ) ≥ α. Therefore

                                                    2
                                        min{1,α}||u i || + q,u i  −  (0) +  (u i ) ≤ 0.
                           The function   is proper, convex, and lower semicontinuous, and thus there
                           exist a ≥ 0 and b ∈ R such that
                                                    n
                                              (∀x ∈ R ) :  (x) ≥−a||x|| + b.
                           We thus have

                                                  2
                                      min{1,α}||u i || + q,u i  −  (0) − a||u i || + b ≤ 0.
                                                       2
                           Dividing this last relation by ||u i || , we get
                                                   q         (0)     a      b
                                              2
                                  min{1,α}||z i || +   ,z i  −   −      +       ≤ 0.
                                                  ||z i ||  ||u i || 2  ||u i ||  ||u i || 2
                           Taking the limit as i →+∞, we get

                                                              2
                                                   min{1,α}||z|| ≤ 0,
                           a contradiction. Thus, for R ≥ R 0 , (4.81) holds, and the Brouwer degree with
                                                       n
                           respect to the set D R ={x ∈ R :||x|| <R} and 0 of the map (t,u)  à
                           u − H(t,u) is well defined for all t ∈[0,1]. Set R 1 = P   (−q)  and let
                           R> max{R 0 ,R 1 }. Using the homotopy invariance property and the normalized
                           property of Brouwer degree, we obtain
                              deg(id R − P   (id R − (F(.) + q),D R ,0) = deg(id R − H(1,.),D R ,0)
                                              n
                                    n
                                                                         n
   115   116   117   118   119   120   121   122   123   124   125