Page 155 - Complementarity and Variational Inequalities in Electronics
P. 155

146  Complementarity and Variational Inequalities in Electronics


                           and

                                                   y L (t) ∈ ∂ (y(t)).                (5.3)
                                                                         p
                              Note that if (∀t ≥ 0) : u(t) = u for some given u ∈ R , then the stationary
                           solutions of (5.1)–(5.3) are given by the solutions of problem NRM(A,B,C,
                           D,u, ) discussed in the previous chapter.
                                               1
                                                                        n
                                                          p
                              We suppose that u ∈ L (0,+∞;R ) and for x 0 ∈ R , we consider problem
                                               loc
                                                                           n
                           P(x 0 ,A,B,C,D,u, ): Find a function x :[0,+∞[ → R ; t  → x(t), and a
                                                 m
                           function y L :[0,+∞[ → R ; t  → y L (t) such that:
                                                      0          n
                                                 x ∈ C ([0,+∞[;R ),                   (5.4)
                                                       1          n
                                                By L ∈ L (0,+∞;R ),                   (5.5)
                                                       loc
                                                 dx    1          n
                                                    ∈ L (0,+∞;R ),                    (5.6)
                                                       loc
                                                 dt
                                                      x(0) = x 0 ,                    (5.7)
                                         dx
                                           (t) = Ax(t) − By L (t) + Du(t), a.e. t ≥ 0  (5.8)
                                         dt
                                                 y(t) = Cx(t), ∀ t ≥ 0,               (5.9)
                           and
                                               y L (t) ∈ ∂ (y(t)), a.e. t ≥ 0.       (5.10)

                              Let us now make the following two assumptions.

                           Assumption (G1): There exists a symmetric and invertible matrix R ∈ R n×n
                           such that
                                                          T
                                                     R −2 C = B.
                                                          n
                           Assumption (G2): There exists z 0 ∈ R such that   is finite and continuous at
                           y 0 = CR −1 z 0 .

                                               ) .Using (5.8), (5.9), and (5.10), we may consider the
                              Note that R −2  = (R −1 2
                           differential inclusion
                                               dx
                                                  ∈ Ax − B∂ (Cx) + Du.
                                               dt
                           Setting z = Rx, we remark that

                                               dx
                                                  ∈ Ax − B∂ (Cx) + Du
                                               dt
                                                         ⇔
   150   151   152   153   154   155   156   157   158   159   160