Page 163 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 163

150                                                   Minimal surfaces


                          We therefore only need to show that if ku 0 k C 2,α ≤  , we have indeed (5.20).
                       Note that for every ν ≥ 2,wehave
                                    ⎧
                                    ⎨ ∆ (u ν+1 − u ν )= N (u ν ) − N (u ν−1 )  in Ω
                                    ⎩
                                                u ν+1 − u ν =0          on ∂Ω .
                       From Step 1 and Step 2, we find that, for every ν ≥ 2,

                                     ku ν+1 − u ν k  2,α ≤ C kN (u ν ) − N (u ν−1 )k  0,α
                                                C                         C
                                                                                       (5.22)
                                        2                     2
                                    ≤ C (ku ν k C 2,α + ku ν−1 k C 2,α) ku ν − u ν−1 k C 2,α .
                       Similarly for ν =1,wehave

                                                                     2    3
                                     ku 2 − u 1 k C 2,α ≤ C kN (u 1 )k C 0,α ≤ C ku 1 k C 2,α .  (5.23)
                       From now on, since all norms will be C 2,α  norms, we will denote them simply
                       by k·k. From (5.22), we deduce that it is enough to show

                                                            2
                                             2
                                           C (ku ν k + ku ν−1 k) ≤ K, ν ≥ 2            (5.24)
                       to obtain (5.20) and thus the theorem. To prove (5.24), it is sufficient to show
                       that
                                                     µ      4 2  ¶
                                                          C
                                           ku ν k ≤ C  1+        ,ν ≥ 2 .              (5.25)
                                                          1 − K
                       The inequality (5.22) will then follow from the choice of   in (5.21). We will
                       prove (5.25) by induction. Observe that by Step 1 and from (5.18), we have
                                                        µ     µ       4 2  ¶¶
                                                                    C
                                      ku 1 k ≤ C ku 0 k ≤ C  ≤ C  1+         .         (5.26)
                                                                   1 − K
                       We now show (5.25) for ν =2. We have, from (5.23) and from (5.26), that

                                                          ³           ´
                                                                2    2       ¡     4 2  ¢
                             ku 2 k ≤ ku 1 k + ku 2 − u 1 k ≤ ku 1 k 1+ C ku 1 k  ≤ C  1+ C
                       and hence since K ∈ (0, 1), we deduce the inequality (5.25). Suppose now, from
                       the hypothesis of induction, that (5.25) is valid up to the order ν and let us
                       prove thatitholds true for ν +1. We trivially have that
                                                           ν
                                                          X
                                           ku ν+1 k ≤ ku 1 k +  ku j+1 − u j k .       (5.27)
                                                          j=1
   158   159   160   161   162   163   164   165   166   167   168