Page 90 - INTRODUCTION TO THE CALCULUS OF VARIATIONS
P. 90

Fields theories                                                    77

                Since ξ → f (x, u, ξ) is convex, then the function E is always non negative.
                   Note also that since u (x)= Φ (x, u 0 (x)) then
                                      0
                                      0
                                                    0
                             E (x, u 0 (x) , Φ (x, u 0 (x)) ,u (x)) = 0, ∀x ∈ [a, b] .
                                                    0
                Now using the definition of exact field we get that, for every u ∈ X,
                                                                   0
                      f (x, u (x) , Φ (x, u (x))) + f ξ (x, u (x) , Φ (x, u (x))) (u (x) − Φ (x, u (x)))
                                                       d
                   = f (x, u, Φ) − pΦ + pu = S x + S u u =  [S (x, u (x))] .
                                                   0
                                        0
                                                       dx
                Combining these facts we have obtained that
                             Z  b
                    I (u)=       f (x, u (x) ,u (x)) dx
                                           0
                               a
                             Z  b  ½                            d           ¾
                                                        0
                          =       E (x, u (x) , Φ (x, u (x)) ,u (x)) +  [S (x, u (x))]  dx
                               a                               dx
                             Z
                               b
                                 d
                          ≥         [S (x, u (x))] dx = S (b, u (b)) − S (a, u (a)) .
                               a  dx
                Since E (x, u 0 , Φ (x, u 0 ) ,u )= 0 we have that
                                      0
                                      0
                                   I (u 0 )= S (b, u 0 (b)) − S (a, u 0 (a)) .
                Moreover since u 0 (a)= u (a), u 0 (b)= u (b) we deduce that I (u) ≥ I (u 0 ) for
                every u ∈ X. Thisachievesthe proofofthe theorem.
                   The quantity
                                         Z  b
                                            d
                                               [S (x, u (x))] dx
                                            dx
                                          a
                is called invariant Hilbert integral.

                2.6.1   Exercises
                                                                                   N
                Exercise 2.6.1 Generalize Theorem2.21tothe case where u :[a, b] → R ,
                N ≥ 1.
                Exercise 2.6.2 Generalize Hilbert Theorem (Theorem 2.27) to the case where
                           N
                u :[a, b] → R , N ≥ 1.
                Exercise 2.6.3 (The present exercise establishes the connection between exact
                field and Hamilton-Jacobi equation). Let f = f (x, u, ξ) and = H (x, u, v) be as
                in Theorem 2.10.
   85   86   87   88   89   90   91   92   93   94   95