Only Transcendentals Known

Calculus Level 4

Consider all functions f : R R f: \mathbb{R} \to \mathbb{R} satisfying that

  • f ( x ) = 0 f(x) = 0 for all transcendental numbers x x ,
  • the Lebesgue-integral f ( t ) d t \displaystyle \int_{-\infty}^{\infty} f(t) \, dt exists.

Let S = { f ( t ) d t } S = \left\{ \displaystyle \int_{-\infty}^{\infty} f(t) \, dt \right\} . Which of the following describes S S ?

S S is empty S S is a singleton set S S is a finite set with at least two elements S S is a countably infinite set S S is a uncountablly infinite set

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Calvin Lin Staff
Oct 20, 2016

R T \mathbb {R} - \mathbb{T} are the algebraic numbers, which are roots of polynomials with integer coefficients. We can show that there are only a Countable number of these, hence the algebraic numbers have a measure of 0.

Since we're working with lebesgue integral , the value of the integral is independent of the value on a set of measure 0. As such, the integral can be calculated just over T \mathbb{T} , and we obtain that the value is 0.

Hence, for all functions f f that satisfy f ( t ) = 0 f(t) = 0 for transcedental t t (and have no restriction otherwise), we have f = 0 \int f = 0 .

Note: All functions that satisfy f ( t ) = 0 f(t) = 0 for transcedental t t are Lebesgue integrable.


If the question used the improper riemann integral instead, then the proof is slightly more complicated. We have to work from first principles, using the fact that the integral exists.

First, for the improper integral \int_{-\infty} ^ \infty which is interpreted as lim ( a , b ) ( , ) a b \lim_{(a,b) \rightarrow ( - \infty, \infty) } \int_a ^ b , we get that f : [ a , b ] R f : [a,b] \rightarrow \mathbb{R} is Riemann-integrable.

Second, we will show that for fixed a , b a,b , we have a b f ( x ) d x = 0 \int_a^b f(x)\, dx = 0 . This follows from the definition of the Riemann integral:

For all ϵ > 0 \epsilon > 0 , there exists a δ \delta such that for any tagged partition x 0 , , x n x_0, \ldots , x_n and t 0 , , t n 1 t_0, \ldots , t_{n-1} whose mesh is less than δ \delta , we have
i = 0 n 1 f ( t i ) ( x i + 1 x i ) S < ϵ | \sum_{i=0}^{n-1} f(t_i) ( x_{i+1} - x_i) - S | < \epsilon

For any ϵ > 0 \epsilon > 0 , there exists a δ ϵ \delta_ \epsilon such that for all tagged partitions with mesh less than δ ϵ \delta_\epsilon , i = 0 n 1 f ( t i ) ( x i + 1 x i ) S < ϵ | \sum_{i=0}^{n-1} f(t_i) ( x_{i+1} - x_i) - S | < \epsilon . Fixing the partition to use internal interval endpoints of the form n δ ϵ n \delta_\epsilon and picking transedentals t i t_i as our tags, we see that i = 0 n 1 f ( t i ) ( x i + 1 x i ) = 0 \sum_{i=0}^{n-1} f(t_i) ( x_{i+1} - x_i) = 0 . This gives us S < ϵ |S | < \epsilon . Now, taking limits as ϵ 0 \epsilon \rightarrow 0 , we conclude that S = 0 S = 0 .

Finally, = lim ( a , b ) ( , ) a b = lim 0 = 0 \int_{-\infty} ^ \infty = \lim_{(a,b) \rightarrow ( - \infty, \infty) } \int_a ^ b = \lim 0 = 0 .

Note: In contrast to the Lebesgue integral, there are a lot of functions which satisfy f ( t ) = 0 f(t) = 0 for transcendentals t t , but are not Riemann integrable. This includes functions which are unbounded on a compact interval.

Note: An alternative is to use the Riemann-Lebesgue criterion in the first half to conclude that we can perform the Lebesgue integral which yields a value of 0.

@Guillermo Templado @Deeparaj Bhat I've added how to deal with the Riemann integral version of the problem. We have to be careful with each stage of the formulation to get at the actual result.

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

Wait, didn't you assume that f f is bounded?

A Former Brilliant Member - 4 years, 7 months ago

Log in to reply

No. Given that the integral exists, I found a tagged partition which makes the value 0, and so a b < ϵ |\int_a^b | < \epsilon . Then, by taking limits, we can conclude that a b = 0 \int_a ^b = 0 . Let me edit this step in completely.

Since it is a property that "any Riemann integrable function on a compact interval is bounded", it follows that the function is bounded on a compact interval. However, the function over the reals need not be bounded. For example, we could have f ( n ) = n f(n) = n for integers and 0 otherwise. This is why I had a note of "functions which are unbounded on a compact interval and have f ( t ) = 0 f(t) = 0 are not Riemann integrable".


To recall what I said in the previous discussion, my main point of contention was "Be clear with which integration approach you are using, the question is trickier for Riemann integrals" (The other issue I had was with the definition of "completely determined", which we cleared up).

Calvin Lin Staff - 4 years, 7 months ago

Log in to reply

@Calvin Lin Right. I didn't notice that you'd said that f is Riemann integrable initially. Also, like you said, once we have f is bounded on any compact interval, we can pass to the Lebesgue integral by the Riemann-Lebesgue criterion (If we had that f is continuous on T \mathbb{T} ).

A Former Brilliant Member - 4 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...