Borel measure on the real line

275 Views Asked by At

This problem is about the construction of Borel measure on the real line.(in Folland chapter 1,section 5.), more specifically, construction of premeasure on h interval.

To prove that

$\mu_{0}\left(\bigcup_{1}^{n}\left(a_{j},b_{j}\right]\right)=\sum_{1}^{n}\left[F\left(b_{j}\right)-F\left(a_{j}\right)\right]$ is a premeasure for finite disjoint h interval .

The idea is first approximate those h interval ($a_i$,$b_i$] that form (a,b] by $[a_i +\delta,b_i]$ and since compact set has finite open cover,then we can use "size" of finite open cover to approixmate the "size" of (a,b].

My question is why finite open cover is necessary in the proof,why can't we just use infinite open cover $(a_i,b_i+\delta)$ for all component of (a,b] to approixmate "size" of (a,b],More clearly :

$\mu_0(I) \le F(b)- F(a+\delta) +\epsilon \le F(b+\delta)-F(a_1) +\epsilon \le \sum_{i\to \infty}(F(b_i+\delta)-F(a_i)) +\epsilon \le \sum_{i\to \infty} \mu_0(I_j) + 2 \epsilon$?

The third ineuqality is due to the incresing of function F.

Do I state the problem clear?