Page 145 - Complementarity and Variational Inequalities in Electronics
P. 145

136  Complementarity and Variational Inequalities in Electronics


                           now deduce that necessarily z 2 ≥ 0. The second relation in (4.123) then gives

                                                  + (R 2 z 2 − α I R 2 z 3 ) ≤ 0.

                           We have

                                       (R 2 z 2 − α I R 2 z 3 ) ≤   + (R 2 z 2 − α I R 2 z 3 ) ≤ 0.

                           However, R 2 z 2 ≥ 0 and −α I R 2 z 3 ≥ 0, so that

                                                 (R 2 z 2 − α I R 2 z 3 ) = 0.

                           Let us now suppose that b) z 1 ≥ 0. Then necessarily   ≤ 0, and the third relation
                           in (4.123) entails also that z 3 ≥ 0. From (4.124) we now deduce that necessarily
                           z 2 ≤ 0. We have z 2   ≥ 0, and the second relation in (4.123) then gives

                                     z 2 (R 2 z 2 − α I R 2 z 3 ) ≤ z 2   + z 2 (R 2 z 2 − α I R 2 z 3 ) ≤ 0,

                           and since z 2 ≤ 0, we get

                                                  R 2 z 2 − α I R 2 z 3 ≥ 0.
                           However, z 2 ≤ 0 and z 3 ≥ 0 yields R 2 z 2 − α I R 2 z 3 ≤ 0, so that

                                                  R 2 z 2 − α I R 2 z 3 = 0.

                           So, in both cases a) and b), we have proved that R 2 z 2 − α I R 2 z 3 = 0, and thus

                                                 1
                                                   (R 2 z 2 − α I R 2 z 3 ) = 0.
                                                 K
                           This property implies

                                         1                 1
                                          (R 2 x 2 − α I R 2 x 3 ) =  (R 2 X 2 − α I R 2 X 3 ),
                                        K                  K
                           that is, the uniqueness of the output signal

                                                          R 2

                                               V o = R 2 I E =  (I − α I I ).
                                                           K
                           For the last two examples, we get the existence of at least one solution x of
                           VI(M, ,Fu). Using now Proposition 15, we see that if x is another solution,
                           then
                                                                 T
                                                 x − x ∈ ker{M + M },               (4.125)
   140   141   142   143   144   145   146   147   148   149   150