Page 132 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 132

The model case: Dirichlet integral                                119

                and observe that v ∈ C 0 ∞  (B R (x)). An easy computation gives, for r = |x − y|,

                                         n − 1         1−n  d £       ¤
                            ∆v = ϕ (r)+       ϕ (r)= r       r  n−1 0
                                                                 ϕ (r) .
                                                0
                                   00
                                           r              dr
                From now on we assume that n ≥ 2,the case n =1 is elementary and is discussed
                in Exercise 4.3.1. We next let
                                               d £         ¤
                                        ψ (r)=    r n−1 0                        (4.9)
                                                      ϕ (r) .
                                               dr
                Note that ψ ∈ C  ∞  ( , R) and
                              0
                                           Z  R
                                              ψ (r) dr =0 .                     (4.10)

                Remark that the converse is also true, namely that given ψ ∈ C 0 ∞  ( , R) satisfying
                (4.10) we can find ϕ ∈ C 0 ∞  ( , R) verifying (4.9).
                   Step 2. Let ψ ∈ C 0 ∞  ( , R), satisfying (4.10), be arbitrary. Define then ϕ and
                v as above and use such a v in (4.7). We get, since v ≡ 0 on Ω \ B R (x),
                            Z          Z              Z  R         Z
                     0=       u∆vdy =        u∆vdy =      ψ (r) r 1−n    udσ dr
                             Ω          B R (x)                     ∂B r (x)
                              R
                            Z
                        =      ψ (r) w (r) dr

                wherewehaveset
                                                  Z
                                       w (r)= r 1−n      udσ .
                                                   ∂B r (x)
                We can use Corollary 1.25 to deduce that

                                    w (r)= constant, a.e. r ∈ ( , R) .

                We denote this constant by σ n−1 u (x) andweuse thefact that   is arbitrary to
                write
                                   w (r)= σ n−1 u (x) ,a.e. r ∈ (0,R) .         (4.11)
                   Step 3. We now conclude with the proof of the theorem. We first observe
                that (4.11) is nothing else than (4.8) with the fact that u is independent of R.
                We next integrate (4.8) and get

                                                    Z
                                              1
                                  u (x)=                   u (z) dz .           (4.12)
                                         meas B R (x)
                                                      B R (x)
   127   128   129   130   131   132   133   134   135   136   137