# Proving a Function is Rieman Integrable

1. Jan 21, 2008

### SNOOTCHIEBOOCHEE

1. The problem statement, all variables and given/known data

Prove that
f(x) = 1 x (element) E = {1/n : n (element) N}
0 x (element) [0,1]\E
is integrable on [0, 1] by using the definition of integrability

2. Relevant equations

Definition of Integrability: Let f be a bounded function. for each epsilon greater than 0 there exists a partition P such that. U(f,P)- L(f,P)< epsilon

3. The attempt at a solution

Fix $$\epsilon$$ >0

I think that for any partition P, L(f,P) is gunna be zero. Since mk will always be 0 for all k.

so all we need to do is find a partition P such that U(f,P) < $$\epsilon$$

i could be completely wrong up to this step, and even if im not, i don't know how to go about choosing P. HELP MEH PLZ!!!

2. Jan 21, 2008

### Mathdope

A partition of [0,1] may have the form {0,1/k,t_1,...,t_s = 1}. On [0,1/k] sup(f)=1. On [1/k,1] there are now only a finite number of elements for which f = 1, namely, x = 1,1/2,1/3,...,1/k. If you choose k large enough so that 1/k < $\epsilon/2$ and you have a finite number of elements remaining for which f = 1, you need to make sure that the subintervals surrounding 1,...1/(k-1) can also have a total summed length < $\epsilon/2$. I leave the rest to you.

3. Jan 21, 2008

### SNOOTCHIEBOOCHEE

Ok im still having trouble with this.

from what i can tell the picture looks like this: http://i31.photobucket.com/albums/c373/SNOOTCHIEBOOCHEE/Graph.jpg

Sorry for the crude drawing.

Anyways. so 1/k <epsilon/2

that means that the whole area from [0,1/k] must be < epsilon/2. I understand this and can write it out using sigma notation...

Now choosing the the intervals around the other parts are proving to be difficult for me. I know they have to sum up to be less than Epsilon/2.

So what i am thinking is that if we make the with of the partitions epsilon/k, that the maximum value of the area of the graph will also be epsilon/2. Is this correct?

4. Jan 21, 2008

### Mathdope

Close. I think you want to make sure that the subintervals have length less than $\epsilon/2k$.

5. Jan 21, 2008

### SNOOTCHIEBOOCHEE

ya thats what i meant. Thank you sir.