Page 102 -
P. 102

86     2. Boundary Integral Equations

                           first row of C Ω which leads to the following system for the interior Dirichlet
                           problem,
                                                                     1
                                      &        '&    '   &                   '&   '
                                       V 23  V 24  σ 1      −D 21    2  I − K 22  ϕ 1
                               Vσ :=                   =                            =: f .
                                                           1                            i
                                       V 13  V 14  σ 2      I + K 11  −V 12     ϕ 2
                                                           2
                                                                                       (2.4.26)
                           The solution of the interior Dirichlet problem has associated Cauchy data σ
                           which satisfy the three compatibility conditions:

                                              (σ 1 n + σ 2 x)ds x = 0 , − σ 2 ds =0 .  (2.4.27)
                                            Γ                      Γ
                              As we shall see in Chapter 10 , V is known as a strongly elliptic operator
                           for which the classical Fredholm alternative holds. Hence uniqueness will
                                                                                 α
                           imply the existence of exactly one solution σ ∈ C 1+α (Γ) × C (Γ).
                              For the exterior Dirichlet problem, by using C Ω c and the representation
                           (2.4.13) we obtain the system with integral equations of the first kind,

                                                              1
                                                   &                  '&    '
                                                      D 21    2  I + K 22  ϕ 1
                                     Vσ + Rω = −                              =: f ,
                                                                                  e
                                                    1
                                                    2  I − K 11  V 12     ϕ 2

                                                                                       (2.4.28)
                                              (σ 1 n + σ 2 x)= A 1 , − σ 2 ds = A 0
                                             Γ                    Γ
                           where

                                              0  n 1  n 2                          3

                                    R(x)= −                and ω =(ω 0 ,ω 1 ,ω 2 ) ∈ IR .
                                              1  x 1  x 2
                           Lemma 2.4.3. The homogeneous system corresponding to (2.4.28) has only
                                                                   3
                                                          α
                           the trivial solution in C  1+α (Γ) × C (Γ) × IR .
                           Proof: Let σ 0 , ω 0 be any solution of
                                                        ∂v

                               Vσ 0 + Rω 0 =0 on Γ,        ,v σ 0 ds = 0 for all v ∈   (2.4.29)
                                                         ∂n
                                                      Γ
                           and consider the solution of (2.4.1),
                                                                 ◦   ◦      ◦              c
                             u 0 (x):= V σ 0 (x)+ p 0 (x) with p 0 (x)= ω 0 + ω 1 x 1 + ω 2 x 2 for x ∈ Ω .
                           Then A 0 =0 , A 1 = 0 because of (2.4.29) and (2.4.18), hence u 0 = O(|x|)at
                                                                                     c
                           infinity due to (2.4.13) which implies u 0 (x)= p 0 (x) for all x ∈ Ω ∪ Γ.On
                           the other hand, u 0 (x) is also a solution of (2.4.1) in Ω and is continuous
                           across Γ where u 0 | Γ = 0. Hence, due to Theorem 2.4.1, u 0 (x) = 0 for all x
                                                                ±
                                                 ±
                           in Ω. Consequently, Mu | Γ =0 and Nu | Γ = 0. Then the jump relations
                                                 0
                                                                0
                           corresponding to C Ω c −C Ω imply σ 0 =([Mu]| Γ , [Nu]| Γ )    = 0 on Γ and
                                       +
                                                                           ◦
                                −
                           0= u | Γ = u | Γ = p 0 implies p 0 (x) = 0 for all x, i.e., ω =0.
                                       0
                                0
   97   98   99   100   101   102   103   104   105   106   107