Page 181 - Complementarity and Variational Inequalities in Electronics
P. 181

172  Complementarity and Variational Inequalities in Electronics


                           We check that both assumptions of Corollary 12 are satisfied. Indeed,

                                                                       1
                                                     2
                                         (∀ (x 1 ,x 2 ) ∈ R ) :∇V(x 1 ,x 2 ) = (x 1 , x 2 ).
                                                                       k
                           If x ∈ K, then
                                                            k − 1
                                             x −@ V(x) = (0,    x 2 ) ∈ K,
                                                             k
                           and thus
                                          ∀x ∈ K : ψ K (x) − ψ K (x −@ V(x)) = 0.

                           We obtain
                                                                         h  2
                                    
Ax,∇V(x) + ψ K (x) − ψ K (x −@ V(x)) =  x ≥ 0.
                                                                            2
                                                                         k
                           Here
                                                         h  2
                                     E(F,ψ K ,V ) ={x ∈ K :  x = 0}={(x 1 ,0);x 1 ≥ 0}.
                                                           2
                                                         k
                           Let z = (z 1 ,0) ∈ E(F,ψ K ,V ). We claim that if γ(z) ⊂ E(F,ψ K ,V ), then
                           necessarily z 1 = 0. Indeed, suppose that γ(z) ⊂ E(F,ψ K ,V ) and set x(·) =
                                          dx
                           x(·,t 0 ,x 0 ), x (·) =  (·,t 0 ,x 0 ). From the dynamics in E(F,ψ K ,V ) we have
                                           dt

                               x (t)(v 1 − x 1 (t)) + kx 1 (t)v 2 ≥ 0, ∀v 1 ≥ 0,v 2 ∈ R, a.e. t ≥ t 0 .  (5.55)
                                1
                           Setting v 1 = x 1 (t) in (5.55), we get

                                             kx 1 (t)v 2 ≥ 0,∀v 2 ∈ R, a.e. t ≥ t 0 .  (5.56)

                           Setting v 2 = 0in (5.55),wealsoget


                                             0 ≤ x 1 (t) ⊥ x (t) ≥ 0, a.e. t ≥ t 0 .  (5.57)
                                                        1
                                                        2
                           From relation (5.57) we obtain  d  x (t) = 0, a.e. t ≥ t 0 , from which it easily
                                                     dt 1
                           follows that x 1 (t) = z 1 ,a.e. t ≥ t 0 . Relation (5.56) then gives z 1 = 0. Therefore,
                           M ={0} is the largest invariant subset of E(F,ψ K ,V ), and for any x 0 ∈ K,we
                           have
                                                   lim x(t;t 0 ,x 0 ) = 0.
                                                  t→+∞
                           Example 68. Let h> 0 and k> 0. Suppose that
                                                    2
                                             (∀x ∈ R ) : F(x) = Ax +∇ (x)
   176   177   178   179   180   181   182   183   184   185   186