Page 102 - Complementarity and Variational Inequalities in Electronics
P. 102

A Variational Inequality Theory Chapter | 4 93


                           Remark 24. The famous result on complementarity problems with P-matrices
                           (see e.g. [37]) can be easily deduced from Theorem 7. Indeed, if   =   R , then
                                                                                     n
                                                                                     +
                                                                      (1 ≤ i ≤ n), and if M is a
                             can be written as in (4.48)–(4.50) with   i =   R +
                                                                                     n
                           P-matrix, then Theorem 7 can be applied to ensure that for each q ∈ R , there
                           exists a unique u ≥ 0 such that Mu + q ≥ 0 and  u,Mu + q = 0.
                           Example 47. Let
                                                     ⎛             ⎞
                                                        1  −2   −8
                                                 M = ⎝ 0    1   2 ⎠
                                                                   ⎟
                                                     ⎜
                                                        0   0   2
                           and
                                                     3
                                              (∀x ∈ R ) :  (x) =|x 1 |+|x 3 |.
                                                                 3
                           The matrix M is a P-matrix, and   ∈ D (R ;R ∪{+∞}). Thus, for each
                                3
                           q ∈ R , problem V I(M,q, ) has a unique solution.
                           4.9 SEMICOERCIVITY AND SOLVABILITY CONDITIONS
                           The results discussed in the previous section require that B(M, ) ={0}.If
                           B(M, ) 
= {0}, then in assuming some semicoercivity condition on the ma-
                           trix M, we may determine conditions on q ensuring the solvability of problem
                           VI(M,q, ).

                                             n
                           Theorem 8. Let   : R → R∪{+∞} be a proper convex lower semicontinuous
                           function with closed domain, and let M ∈ R n×n . Suppose in addition that

                                           (∀x ∈ D( ) ∪ D( ) ∞ ) : Mx,x ≥ 0.

                              If there exists x 0 ∈ D( ) such that
                                                             T
                                  (∀v ∈ B(M, ), v 
= 0) : q − M x 0 ,v +   ∞ (v) > 0,  (4.60)

                           then problem VI(M,q, ) has at least one solution.
                                         n
                           Proof. Let q ∈ R .Here (M, ) ∈ PD0 n , and using Proposition 20 and Theo-
                                                                         n
                           rem 4, we see that for all i ∈ N,i 
= 0, there exists u i ∈ R such that
                                   1                                          n
                                  ( I + M)u i + q,v − u i  +  (v) −  (u i ) ≥ 0, ∀v ∈ R .  (4.61)
                                   i
                           We claim that the sequence {u i }≡{u i ;i ∈ N\{0}} is bounded. Suppose on the
                           contrary that ||u i || → +∞ as i →+∞. Then, for i large enough, ||u i || 
= 0,
   97   98   99   100   101   102   103   104   105   106   107