Page 120 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 120

Relaxation theory                                                 107

                           ν
                                                         ν
                   (ii) Let u (x)= x/ (|x| +1/ν). Show that u  u in W 1,p ,for any 1 ≤ p<
                2.
                   (iii) Let δ (0,0) be the Dirac mass at (0, 0), which means
                                  ­       ®
                                   δ (0,0) ; ϕ = ϕ (0, 0) , ∀ϕ ∈ C  ∞  (Ω) .
                                                            0
                Prove that
                                           ν
                                      det ∇u  πδ (0,0) in D (Ω) .
                                                          0
                3.6    Relaxation theory

                Recall that the problem under consideration is
                            ½       Z                                      ¾
                                                                      1,p
                    (P)  inf I (u)=    f (x, u (x) , ∇u (x)) dx : u ∈ u 0 + W  (Ω)  = m
                                                                     0
                                     Ω
                where
                          n
                   - Ω ⊂ R is a bounded open set with Lipschitz boundary;
                                 n
                   - f : Ω × R × R −→ R, f = f (x, u, ξ), is continuous, uniformly in u with
                respect to ξ;
                   - u 0 ∈ W 1,p  (Ω) with I (u 0 ) < ∞.

                   Before stating the main theorem, let us recall some facts from Section 1.5.
                Remark 3.27 The convex envelope of f, with respect to the variable ξ,willbe
                           ∗∗
                denoted by f . It is the largest convex function (with respect to the variable ξ)
                which is smaller than f.In other words
                       g (x, u, ξ) ≤ f  ∗∗  (x, u, ξ) ≤ f (x, u, ξ) , ∀ (x, u, ξ) ∈ Ω × R × R n

                for every convex function g (ξ → g (x, u, ξ) is convex), g ≤ f. We have two ways
                of computing this function.
                   (i) From the duality theorem (Theorem 1.54) we have, for every (x, u, ξ) ∈
                         n
                Ω × R × R ,
                                  ∗      ∗            ∗
                                 f (x, u, ξ )= sup {hξ; ξ i − f (x, u, ξ)}
                                              ξ∈R n
                                                                  ∗
                                                     ∗
                               f  ∗∗  (x, u, ξ)= sup {hξ; ξ i − f (x, u, ξ )} .
                                                           ∗
                                            ξ ∈R n
                                             ∗
                   (ii) From Carathéodory theorem (Theorem 1.55) we have, for every (x, u, ξ) ∈
                         n
                Ω × R × R ,
                                (                                                )
                                  n+1                n+1                n+1
                                  X                  X                  X
                 f  ∗∗  (x, u, ξ)= inf  λ i f (x, u, ξ ): ξ =  λ i ξ ,λ i ≥ 0 and  λ i =1 .
                                              i
                                                           i
                                  i=1                i=1                i=1
   115   116   117   118   119   120   121   122   123   124   125