Page 87 - Vector Analysis
P. 87

§3.4 The Fubini theorem                                                                                                     83

                ż

Therefore, 1B(x) dx = 0, so there exists a partition P1 of R such that

                        R

                       ÿ                    =  U (1B, P1)        ă   [   sup  f  A  (R)   ε                ]  .
                                 ν(∆)                               2                    ´ inf f A(R)  +1

                  ∆PP1,∆XB‰H

                  ()
Let U ” int              ď ∆ . Then B Ď U . Since the discontinuity of f A is a subset of

                           ∆PP1,∆XB‰H

B, f A : R X U A Ñ R is continuous. Since R X U A is closed and bounded, f A is uniformly

continuous; thus there exists δ ą 0 such that

              ˇˇf A(x1)    ´  f A(x2)ˇˇ  ă     ε               if x1, x2 P R X U A and }x1 ´ x2} ă δ .
                                            2ν(R)

Let P be a refinement of P1 such that }P} ă δ, and define two classes C1, C2 of

rectangles    in  P  by    C1  =  ␣∆1    P  P  ˇ  ∆1  Ę  ∆     for  all  ∆  P  P1   satisfying   ∆XB       ‰  H(  and   C2  =
                                               ˇ

␣∆1  P  P  ˇ  ∆1  R  C1(.  Then   if  ∆1    P  C1,  then    ∆1   Ď    RzU A;     thus
           ˇ

        U (f A, P) ´ L(f A, P) = ÿ             []
                                      ∆(1PP                A1R)(x)                  A1R)(x)
                                                  sup(f                  ´  inf (f               ν(∆1)

                                                  xP∆1                      xP∆1

                                        ÿÿ                     )[ ]
                                  =+                                sup f A(x) ´ inf f A(x) ν(∆1)
                                                                                       xP∆1
                                            ∆1PC1     ∆1PC2         xP∆1
                                                                                                        ]
                                             ε                          [                                  ÿ
                                                      ÿ ν(∆1) + sup f A(R) ´ inf f A(R)
                                  ď      2ν(R) ∆1PC1                                                             ν(∆1)
                                            ε
                                  =              ν(R) +        [                 ´             ]           ∆1PC2
                                         2ν(R)                  sup f A(R)          inf f A(R)
                                                                                                        ÿ
                                                                                                                 ν(∆)

                                                   [                                    ]         ∆PP1 ,∆XB‰H
                                                                                    (R) ε
                                         ε            sup  f  A  (R)  ´  inf  f  A

                                  ă      2  +   [   sup  f  A  (R)       inf f A(R)    +   ]  ă  ε,
                                               2                                          1
                                                                    ´

and we conclude that f is Riemann integrable over A by Riemann’s condition.                                                 ˝

3.4 The Fubini theorem

If f : [a, b] Ñ R is continuous, the fundamental theorem of Calculus can be applied to
computed the integral of f over [a, b]. In the following two sections, we focus on how the
integral of f over A Ď Rn, where n ě 2, can be computed if the integral exists.

Definition 3.24. Let A Ď Rn and B Ď Rm be bounded sets, S = A ˆ B be a product set in
Rn+m, and f : S Ñ R be bounded. For each fixed x P A, the lower integral of the function
   82   83   84   85   86   87   88   89   90   91   92