Page 34 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 34

p
                L spaces                                                           21

                see, for example, Theorem 2.1.5 in [31]. We will also assume, without loss of
                generality, that
                                               1
                                             Z
                                         u =    u (x) dx =0 .
                                              0
                   Step 1. Observe that if 1 ≤ p< ∞,then

                                         Z  1            Z  1
                                  p                p               p
                              ku ν k  =     |u ν (x)| dx =  |u (νx)| dx
                                  L p
                                          0               0
                                           Z  ν           Z  1
                                          1         p             p
                                      =       |u (y)| dy =   |u (y)| dy .
                                         ν  0              0
                The last identity being a consequence of the 1-periodicity of u. We therefore
                find that
                                           ku ν k L p = kuk L p .                (1.1)
                The result is trivially true if p = ∞.
                   Step 2. (For a slightly different proof of this step see Exercise 1.3.5). We
                                       p
                thereforehavethat u ν ∈ L and, since u =0,wehavetoshowthat
                                   Z  1
                                                              p
                                                               0
                               lim    u ν (x) ϕ (x) dx =0, ∀ϕ ∈ L (0, 1) .       (1.2)
                               ν→∞
                                    0
                                                  0
                                                 p
                Let  > 0 be arbitrary. Since ϕ ∈ L (0, 1) and 1 <p ≤∞, which implies
                1 ≤ p < ∞ (i.e., p 6= ∞), we have from Theorem 1.13 that there exists h astep
                                0
                    0
                function so that
                                           kϕ − hk  p 0 ≤  .                     (1.3)
                                                  L
                Since h is a step function, we can find a 0 =0 <a 1 < ... < a I =1 and α i ∈ R
                such that
                                 h (x)= α i if x ∈ (a i−1 ; a i ) , 1 ≤ i ≤ I.
                We now compute
                    Z                  Z                         Z
                      1                  1                         1
                       u ν (x) ϕ (x) dx =  u ν (x)[ϕ (x) − h (x)] dx +  u ν (x) h (x) dx
                     0                  0                         0
                and get that

                  ¯Z  1           ¯  Z  1                        ¯Z  1           ¯
                  ¯               ¯                              ¯               ¯
                  ¯               ¯      |u ν (x)||ϕ (x) − h (x)| dx +  ¯  u ν (x) h (x) dx .
                                                                                 ¯
                      u ν (x) ϕ (x) dx ≤
                  ¯               ¯                              ¯               ¯
                   0                   0                          0
   29   30   31   32   33   34   35   36   37   38   39