Page 56 - Complementarity and Variational Inequalities in Electronics
P. 56

A Variational Inequality Theory Chapter | 4 47


                           have
                                                α            α
                                                  λ n d n + (1 −  )x 0 ∈ K,
                                                λ n          λ n
                           provided that n is large enough to have  α  ≤ 1. We obtain
                                                           λ n
                                                   α
                                         αd n + (1 −  )x 0 → αd + x 0 ∈ K as n →∞
                                                  λ n
                           since K is closed. The parameter α> 0 has been chosen arbitrarily, and thus

                                                 (∀ α> 0) : αd + x 0 ∈ K,

                           which means that d ∈ K ∞ .
                              Let us now show that K ∞ is a closed convex cone.

                                                 n
                           Proposition 5. Let K ⊂ R be a nonempty closed convex set. Then K ∞ is a
                           closed convex cone.
                           Proof. 1) The set K ∞ is closed as the intersection of a family of closed sets.
                           It is also convex as the intersection of a family of convex sets. 2) Let λ> 0. If
                                          1                               1
                           x 0 ∈ K, then 0 ∈ (K − x 0 ), which results in 0 ∈  (K − x 0 ).3)If z ∈
                                          λ                           λ>0 λ
                           αK ∞ (α > 0), then z = αd for some d ∈ K ∞ , and thus there exists λ n →+∞,
                                 z                              λ n
                           d n →  such that λ n d n ∈ K.Ifweset μ n =  and z n = αd n then we see that
                                 α                              α
                           μ n z n ∈ K,μ n →+∞, and z n → z. This yields z ∈ K ∞ .
                           Remark 13. If K is a nonempty closed convex cone, then K ∞ = K.
                                                                                n
                           Proposition 6. Let K be a nonempty closed convex subset of R . Then (i) If
                           x ∈ K and e ∈ K ∞ , then x + e ∈ K; (ii) If K is bounded, then K ∞ ={0}; and
                                      n
                           (iii) (∀z 0 ∈ R ) : (K − z 0 ) ∞ = K ∞ .
                           Proof. i) This property is a consequence of the relation

                                                       1
                                              K ∞ =      (K − x) ⊂ K − x.
                                                       λ
                                                    λ>0
                              ii) Suppose by contradiction that d 
= 0 ∈ K ∞ . Then there exist sequences
                           λ n →+∞ and d n → d such that λ n d n ∈ K. However, λ n  d n  →+∞, which
                           is a contradiction since K is bounded.
                              iii) If z ∈ (K − z 0 ) ∞ , then there exist λ n →+∞ and z n → z such that
                                                    z 0
                           λ n z n ∈ K −z 0 . Set w n = z n +  .Wehave λ n w n = λ n z n +z 0 ∈ K and w n → z.
                                                    λ n
                           This means that z ∈ K ∞ . Conversely, if z ∈ K ∞ , then there exist λ n →+∞ and
                                                                     z 0
                           z n → z such that λ n z n ∈ K.Ifweset w n = z n −  , then we have λ n w n =
                                                                     λ n
                           λ n z n − z 0 ∈ K − z 0 and w n → z, so that z ∈ (K − z 0 ) ∞ .
   51   52   53   54   55   56   57   58   59   60   61