Page 225 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 225

212                                            Solutions to the Exercises

                               2
                       has no C solution satisfying w (0) = w (1) = α and hence (P α )alsohas no C 2
                       minimizer).
                       Exercise 5.2.4. By hypothesis there exist a bounded smooth domain Ω ⊂ R 2
                                       ¡
                                            ¢
                                                                                        ¡ ¢
                       and a map v ∈ C 2  Ω; R 3  (v = v (x, y),with v x ×v y 6=0 in Ω)sothat Σ 0 = v Ω .
                       Let e 3 =(v x × v y ) / |v x × v y |.We then let for   ∈ R and ϕ ∈ C 0 ∞  (Ω)

                                            v (x, y)= v (x, y)+  ϕ (x, y) e 3 .

                                         Ω .Since Σ 0 is of minimal area and ∂Σ = ∂Σ 0 ,weshould
                       Finally let Σ = v    ¡ ¢
                       have            ZZ                 ZZ
                                                              ¯      ¯
                                                               v × v
                                           |v x × v y | dxdy ≤  ¯      ¯  dxdy .       (7.23)
                                                                x   y
                                          Ω                  Ω
                                                      ¯ ¯
                                               ®
                                   2
                                         ­
                                                                                           2
                                                                   2





                       Let E = |v | , F = v ; v , G = v y  2 , E = |v x | , F = hv x ; v y i and G = |v y | .
                                                      ¯   ¯
                                              y
                                           x
                                 x
                       We therefore get
                                                                                      ¡ ¢
                                                   2
                         E    = |v x +  ϕe 3x +  ϕ e 3 | = E +2  [ϕ hv x ; e 3 i + ϕ hv x ; e 3x i]+ O   2
                                               x              x
                                                                                 ¤
                                      £
                                                                                      ¡ ¢
                         F     = F +   ϕ hv y ; e 3 i + ϕ hv x ; e 3 i + ϕ hv y ; e 3x i + ϕ hv x ; e 3y i + O   2
                                        x          y
                                                            ¤
                                       £
                                                                 ¡ ¢
                         G    = G +2  ϕ hv y ; e 3 i + ϕ hv y ; e 3y i + O   2
                                         y
                       where O (t) stands for a function f so that |f (t) /t| is bounded in a neighborhood
                       of t =0.Appealing to the definition of L, M, N, Exercise 5.2.1 and to the fact
                       that hv x ; e 3 i = hv y ; e 3 i =0,weobtain
                                                                                    ¡ ¢
                                                                               2
                                       2


                           E G − (F )    =(E − 2 Lϕ)(G − 2 ϕN) − (F − 2 ϕM) + O       2
                                                                                ¡ ¢
                                                    2
                                         = EG − F − 2 ϕ [EN − 2FM + GL]+ O        2
                                             ¡       2 ¢             ¡ ¢
                                                                      2
                                         =    EG − F   [1 − 4 ϕH]+ O     .
                       We therefore conclude that
                                        ¯      ¯                        ¡ ¢
                                        ¯       ¯  = |v x × v y | (1 − 2 ϕH)+ O
                                                                          2
                                        v × v
                                         x    y
                       and hence
                                                     ZZ
                                                                             ¡ ¢

                             Area (Σ )= Area (Σ 0 ) − 2   ϕH |v x × v y | dxdy + O   2  .  (7.24)
                                                       Ω
                       Using (7.23) and (7.24) (i.e., we perform the derivative with respect to  )weget
                                       ZZ
                                           ϕH |v x × v y | dxdy =0, ∀ϕ ∈ C  ∞  (Ω) .
                                                                      0
                                         Ω
                       Since |v x × v y | > 0 (due to the fact that Σ 0 is a regular surface), we deduce
                       from the fundamental lemma of the calculus of variations (Theorem 1.24) that
                       H =0.
   220   221   222   223   224   225   226   227   228   229   230