Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

An Integral Question

  1. Nov 25, 2008 #1
    I'm having a hard time understanding the proof of the following: Let a < b < c and let f: [a,c] -> R be Riemann integrable on [a,b], [b,c] and [a,c]. Then

    [tex]\int_a^c f(x) \, dx = \int_a^b f(x) \, dx + \int_b^c f(x) \, dx[/tex].

    Proof. Let C and C' be the characteristic functions of [a,b] and [b,c] respectively, defined on [a,c]. Then f = C f + C' f and the addition formula above follows from the linearity of the integral.

    This is such a facile proof. Sigh. I'm trying to fill in the missing details: I know that since f = C f + C' f, then

    [tex]\int_a^c f(x) \, dx = \int_a^c C(x) f(x) \, dx + \int_a^c C'(x) f(x) \, dx[/tex]

    where I've used the fact that characteristic functions are Riemann integrable and products of Riemann integrable functions are Riemann integrable. Now how would I show, without much fuss, that

    [tex]\int_a^c C(x) f(x) \, dx = \int_a^b f(x) \, dx [/tex]

    for example?
  2. jcsd
  3. Nov 25, 2008 #2


    User Avatar
    Science Advisor

    Just use the fact that C(x)=0 in the interval [b,c], so that C(x)f(x)=0 in this interval, and its integral=0.
  4. Nov 25, 2008 #3
    Where do I use that fact exactly?
  5. Nov 25, 2008 #4


    User Avatar

    Adding what mathman said to the end of the proof would be sufficient, but a rigorous proof could be done using upper and lower sums.
  6. Nov 25, 2008 #5
    Yeah. I decided to ditch that proof. I found a more elementary but longer proof using the definition of Riemann integration, which I found satisfactory.
  7. Nov 25, 2008 #6


    User Avatar
    Science Advisor
    Homework Helper

    Why don't you find the proof in the OP satisfactory?
  8. Nov 26, 2008 #7
    As I wrote, I don't know why

    [tex]\int_a^c C(x) f(x) \, dx = \int_a^b f(x) \, dx [/tex]

    is true.
  9. Nov 26, 2008 #8


    User Avatar
    Science Advisor
    Homework Helper

    But you were given two justifications as to why that is true. The Riemann sums of C(x)f(x) will always be zero on [b,c].
  10. Nov 26, 2008 #9
    Those two justifications did nothing for me. I was hoping someone would give a simple argument that would not use Riemann sums. Oh well. Thanks anyways.
  11. Nov 27, 2008 #10


    User Avatar
    Science Advisor

    The integral from a to c is the sum of the integral from a to b and the integral from b to c. From a to b, C(x)=1, so the integral of Cf = integral of f. From b to c, C(x)=0, so Cf=0 and the integral = 0.
  12. Nov 27, 2008 #11
    That's a circular argument: You're using the very fact I want to prove.
  13. Nov 27, 2008 #12
    I don't know whether you will find the following proof useful, if you haven't already seen it before, but here it is an elementary proof of

    [tex]\int_a^c f(x) \, dx = \int_a^b f(x) \, dx + \int_b^c f(x) \, dx[/tex]


    Let a<c<b. Since the reiman sums do not depend on the way we may partition the interval [a,b] we can partition this interval first into two subintervals, let them

    [a,c] and [c,b] where c is the same point we are using in the integral

    Then we may do the following partitition to both intervals

    [tex]a=x_0<x_1<x_2<...<x_k=c[/tex] and [tex]c=x_k<x_{k+1}<...<x_n=b[/tex]

    So, we can form the following integral sums for both intervals:

    [tex]\sum_{i=1}^kf(\delta_i)\triangle x_i[/tex]

    [tex]\sum_{i=k}^nf(\delta_i)\triangle x_i[/tex]

    where [tex]\delta_i \in [x_{i-1},x_i][/tex] and

    [tex]\triangle x_i=x_i-x_{i-1}[/tex]

    Now, since we are dealing here with finite sums, then we have the following relation

    [tex]\sum_{i=1}^nf(\delta_i)\triangle x_i=\sum_{i=1}^kf(\delta_i)\triangle x_i + \sum_{i=k}^nf(\delta_i)\triangle x_i[/tex]

    Now, if [tex]max \triangle x_i->0[/tex] then taking the limit on both sides,we get our desired result.

    A simmilar argument follows in two other cases when c is not between a and b.
  14. Nov 28, 2008 #13


    User Avatar
    Science Advisor

    It is NOT circular. I don't see what is needed to prove, unless you need a proof that multiplication by 1 leaves the function unchange and multiplication by 0 results in 0.
  15. Nov 28, 2008 #14
    I appologize to just throw stuff here, but i think that e(ho0n3's real question is what is a characteristic function. I think he is dealing with this problem, and hence it is not clear to him why C(x)=1 in [a,b] and C(x)=0 on [b,c]

    I think that a characteristic function of an interval say


    is as follows

    [tex]C(x)=\left\{\begin{array}{cc}1,&\mbox{ if }
    x\in[x_1,x2]\\0, & \mbox{ if } x_2<x<x_1\end{array}\right.[/tex]

    So, this automatically would mean that if C(x) is your characteristic function on the interval [a,b] then C(x)=1 when x is in [a,b] and C(x)=0 when x is in [b,c] the same for C'(x)=1 if x is in [b,c] and C'(x)=0 when x is in [a,b]

    P.s. I once more appologize for i am only a freshman, and my advice shall probably turn out to be helpless and inappropriate.
    Last edited: Nov 28, 2008
  16. Nov 28, 2008 #15
    Well, these two would also follow from the axioms of a field, right?

    Unless one wanted not to rely on algebra at all, but rather on dedekind cuts etc.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook